Music Contents Strange * stromic Music: Systems, Techniques, and Controls Front Mater Foreword Preface ry Statements About the Subject Mattor 2. Considerations ofthe Basic Parameters of Sound Text 4. Basic Signal Processi Text 5. Concepts of Voltage Contral Text 6. Control Voltage Sources Text 1. Sub-Audio Modulation Text 8. Audio Rate Modulation Text 8. Equalization and Filtering Text 10. Magnetic Tape Recording Text 11, Audio Mixing Text " 1 18 5 Fy 6 36 o 10 101 16 16 6 46 m9 1912, Reverberation, Echo and Feedback 194 Text 18 13.Panning and Sound Location Control 26 Text 26 15,Performance Electronics za Text 29 16,Scores for Analysis and Performance 20 Text 208 hk Mattor os Afterword, 267 Annotated Bibliography 268 Index aForeword by Gordon Mumma When in 1972 the first edition of Allen Strange’s Electronic Music: Systems, Tech- niques and Controls was published, the magenta, blue and white covered book rap- idly became ubiquitous. It was the first comprehensive and useful guide to the sub- ject, and was relatively easy to obtain. It had occasional errors of detail, and was, involved in the technological tumult before general standards were agreed upon, 50 that some of the illustrative graphic symbols became relics. Nonetheless, that edition Proved quite robust. At least two factors explain the first edition’s survival for nearly a decade. Firstly, Allen Strange organized the relatively new and complex material so that it evolved ‘with pedigogical sensibility. Secondly, his explanations of conceptual matters were Jucid. This lucidity may be due to a balance in the author's own world. He is an experienced performer of electronic and acoustical instruments, a versatile com- poser, # historian-theorist of diverse cultural background, and a very effective teach- ex In the ensuing years several other books on the subject appeared. Some contrib- ‘uted updated material, and others devoted more attention to certain areas, though often at the expense of others. In spite of the example set by Allen Strange's book, none of the others seems to have achieved his balanced presentation. And few matched his marvelous attitude towards the subject. an attitude which induced the reader to be alert to the many possibilities of a rapidly developing creative me ium. This second edition, as the author notes in his preface. is in many respects a new book. But it repeats that most important achievement of the earlier edition: it is a comprehensive, detailed, and clearly organized guide to working with the instra- ‘ments and technical procedures of electronic music. Besides the expected updating ‘which includes many devices and procedures developed during the 1970s, the author continues his method of explaining details within the contest of general operating Principles. This makes the book applicable to virtually any analog electronic music apparatus. In its relatively short history—a bit more than half a century-music made with electronic and electro-acoustic means is-well on its way to becoming as pluralistic as that produced during many centuries with purely acoustic resources. It already has both “cultivated” and “vernacular” traditions, which are widely disseminated by broadcasting and recording throughout every part of the world. A major part of re- cent popular and commercial music would not exist without synthesizers and the creative use of multi-track recording. The recording studio, whether the relatively simple home-variety or 2 multi-million dollar commercial facility, has itself become musical instrument—in Brian Eno's words, “a compositional tool.” Electronic mu- sic has even developed “folkloric” aspects. Electronic sensors originally designed to detonate anti-personnel weapons are now used as components of public-access elec- ‘ronic-music environments in shopping centers and galleries. This is certainly ana- ogous to the use of cast-off oil drums in the making of steel-band music. ‘As with Allen Strange’s earlier book, this new edition will continue to be an im- portant text for schools and universities. But perhaps more important, in a time of declining support for arts innovation in educational institutions, this book will be vi- tal to creative people who develop their work independently.Preface ‘The original edition of this text was completed over ten years ago. Compared to this new edition the orig- inal writing was a very simple task. In 1970 there were only three or four commercially available instruments today the number has increased to over thirty. What ‘was in 1970 a basic instrument format has expanded in many directions as there are available instruments. Each manufacturer has a different design format and different implications in terms of the application of the instrament. To cover such 2 subject area called for a ‘complete rewriting of the text. Due to the incredible growth in the field of electronic music instrumenta- tion the subject matter covered in this present text has nearly doubled that of the first edition. In a sense it is inaccurate to refer to this as a second edition; it is really a second book created out of the ever ex- panding field. At the same time this new version makes no claim to cover every possible resource and tech- nique available to today’s musician. This text does, however, deal with all of the generally accepted de- signs and techniques common to today’s electronic in- struments. The instructor and reader will find that it provides a firm and basic foundation of understand- ing which allows the user to develop the techniques and processes specific to an individual instrument or studio situation. Regarding pedagogic technique, I know of no two people who approach this extensive subject in exactly the same manner. The organizational approach to teaching electronic music is greatly dependent on the resources of the instrument and the resources of the studio. Instrument X may be very keyboard oriented, implying one approach, while instrument Y is based fon pre-programming and suggests a completely dif- ferent orientation and approach. Some studios are de- signed around multitrack recording facilties, while others exist strictly as real-time performance spaces with little or no recording equipment. Some electronic music programs are compositionally oriented, while other programs deal only with techniques and prob- lems of instrument performance. This book is orga- nized in a manner which is adaptable to any approach. Each chapter is a progressive overview of the what and how of electronic instrumentation. Beginning with a discussion of what “electronic music” means in this decade, each chapter progresses from the basic con- siderations of electronic sound through basic tech- niques of control to advanced process of instrument patching and sound modification. The reader should not consider this to be 2 text on electronic music composition. Composition texts can do more than explain either general or specific tech- niques of organization and structural manipulation. ‘When it comes to the “composing,” this must be dealt ‘with on 2 personal, one-to-one relationship between student, instructor, and the specific composition at hand. This book is simply 2 text on “technique”: how to operate the instruments with various insights on developing 2 consistent working method. Specific com- position assignments or “etudes” have to be designed ‘by the instructor, 2s appropriate to the given resources of a studio. There are certainly compositional impli- cations of many of the patches and techniques explored in this book. A particular mode of control or routing of the signal patch through various modules deter- mines the variables of a musical situation. The per- former's manipulation of these variables is up to the individual, and he/she should be able to explore the possibilities outside of any aesthetic I might be prone to dictate. I have my own set of assignments specific to my studio's resources, and I would expect that every other teacher has his or her own set of specifics. ‘While every instrument has its own unique char- acteristics, the general operational principles remain the seme. With every instrument the performer must deal with basic problems of routing signals through various shaping and modification devices, establish- ing operational norms and configuring ongoing con- ‘trols to produce the desired event. The signal flow through any instrument begins with the sound source, goes through selected modification circuits, and each circuit is assigned some form of control, be it manual manipulation or different kinds of pre-programming. ‘This text is organized in the same way the instrument 4s organized. The initial chapters deal with the basic sonic resources available to the musician. After the introductory considerations each chapter is profusely ‘lustrated with patch diagrams and examples of ex- tant equipment. Several of the unique features avail- able on certain instruments are described in 2 general manner, with the explanation that in many instances these features can be replicated on other instruments. ‘Chapters 4 through 13 are dedicated to various tech- niques of control and sound processing. In these chap- ters the initial “basic patches” are logically expanded and notated in a unified format. At the end of the appropriate chapters there are projects and exercisesintended to enhance the reader's understanding of various techniques. These projects and exercises have been suggested by colleagues, and range from concep- ‘ual problem solving to the realization of compositions by different composers. The last third of the text deals, with techniques and devices usually extemal to the actual performance instrument. Basic techniques of mixing, tape recording, stereo and quadraphonic con- siderations, as well as performance electronics are cov- ered in sufficient detail to give the reader a substantial point of departure for advanced development. This book is not meant to be a substitute for an instrument's operation manual or 2 general handbook for a given studio; rather, it is a basic organizational guide for learning. While many detailed techniques are discussed, the specific patching format and specific techniques have to be left to the individual instructor and the manufactuerer’s own documentation. Every studio has a different design which is applicable to itself, and this text makes no attempt to describe specific studio format. Such information must be left to the instructor, and most of the initial class discus- sions deal with the layout and general operation of 2 given studio. Because we present as many of the re- sources as are generally available, we realize that the reader may find some of the material redundant or too technical, depending on the goals of the class. The in- structor should design the reading assigments in con- sideration of this fact. Certain material may be irrele- vant to some situations and can easily be excluded. Each chapter does, however, begin with the bare es- sentials of a subject area or technique and will be applicable to every instrument. The teacher will also find it necessary to augment certain portions of this text, discussing devices and techniques specific to a studio's own resources. This present text contains over 400 illustrations de- signed to clarify, to demonstrate, and to expand the reader's understanding of the techniques of electronic music. At the same time it is recommended that the reader keep 2 notebook of patches developed in class or on his/her own studio time. Such patch libraries prove invaluable as references while developing new skills and techniques. At the end of each course T have found it useful to make 2 class project out of com- piling all of the most useful student-generated patches into single document for future reference. Another activiity I have found very useful to students is to hhave certain patches and techniques documented on tape. Any patch can potentially generate a wide range of sounds, depending on the fine adjustment of the various controls, and sonic models can be very useful as guides for the beginning student. Each student may also wish to keep a personal library of recorded patches. ‘Various patch assignments can be recorded on a cas- sette tape by the students and indexed to a specific diagram in 2 patchbook. The readers who are first beginning the study of electronic music outside 2 controlled class situation are encouraged to deal with the subjects as they appear in sequence in this text. This will insure continuity of subject matter and avoid encountering undefined terms. The class instructor undoubtedly has his or her own approach to which my approach may be per- fectly applicable. Or the instructor may wish to re- arrange the sequence of the middle chapters to ac- commodate his own methods. This is certainly to be ‘encouraged, as each chapter attempts to be as inde- pendent as possible. The problem of undefined terms due to re-sequencing is solved by cross-references and an extensive index. Certain definitions and clarifice- tions can also be presented by in-class lectures. During recent years low cost technology and the development of the micro-processor have generated much interest in digital applications to the analog electronic music system. At the present time there are several operational systems which make computer con- ‘tol of electronic instruments very attractive. It would ‘be redundant to attempt to rewrite the documenta- tion of these systems for this text since such infor- mation is available from more direct sources. The reader, however, will be introduced to some of these approaches, and certain instruments specifically de- signed for local and/or computer control will be cov- ered. Digital control, at this point at leas, is providing, for more accessible and accurate performance tech- niques in terms of establishing and recalling complex situations. It is also providing for the expansion of a particular instrument's capabilities in terms of more complex functions. This means that the computer is increasing the efficiency of what these instruments ‘can usually do, The concepts of voltage control and parametric design are the same with or without com- ‘puter aid, and at the present time a separate chapter dealing with digital control is not warranted. The number of digitally controlled studios is limited and their operational procedures are usually unique. Until some common ground for digital control has been e5- tablished in commercially available instruments, in- depth documentation is best left to the manufacturer. At the present time certain companies offer instruments specifically intended for digital control. These devices will be dealt with in a context established by their musical function. For example, a digitally controlled amplifier is still an amplifier and serves the same para- ‘metric function irrelevant of manner of control. ‘Modern digital technology has provided the mu- sician with some exciting new handles on musical structure, especially in the area of time manipulation. ‘These instruments, with their basic operational meth- ‘ods and musical applications, have opened up somepossibilities for the musician which ten years ago were considered impossible. In spite of these instruments’ dependence on digital sound reconstruction, their op- exational modes are compatible with modem analog equipment and will be discussed with basically the same vocabulary. On the other side of the coin this book does not concem itself exclusively with ‘hi-tech’ electronic mu- sical instruments. Today's state-of-the-art had its be- ginnings with non-musical devices forced into a music- ‘making chore. This is a healthy attitude for the arts and is not to be discouraged: Techniques involving extent equipment other than commercially available instruments will be dealt with if those techniques pro- vide a workable musicmaking situation, and any sup- portative literature for these applications will be docu- mented. In certain situations 2 mailing tube and 2 microphone may make a very suitable oscillator (see age 207), and bits of editing tape can be used to articalate complex rhythmic sequences. Some may find these types of jury-rigged techniques far from so- phisicated but if they work, then they work! After al, there must be some reason that the muscians’ activi- ties are often called ‘playing. The last chapter of this book is concerned with the final task of the musician—the making of music. Sev- eral composers have been kind enough to make their scores available specifically for the purpose of analysis ‘and performance. These works appear in their original format as designed by the composers. Each score is given in its complete form and is intended to be real- ized by those users with the proper resources. Even if performance is not practical a detailed analysis of each ‘work will be time well spent. Since a score can never completely represent the musical event, four of the five works are available on Dolby cassettes specifically as performance models. For information concerning these tapes please write to Ocean Records, #4 Euclid Avenue. Los Gatos, Calif, 95030. Acknowledgments To acknowledge every person who contributed to the development of this book would be impossible. This 4s a standard first line for acknowledgment statement ‘and it is unfortunately true. Perhaps the greatest help ‘came from my students. Their feedback and never end- ing questions helped me clarify in my own mind just what needs to be said about the subject of electronic ‘music. Of no less importance was the help of Robin and Pat Strange who tolerated my hours of isolation in the studio preparing this manuscript. Special thanks go to the many instrument designers who willingly provided me with documentation, photographs and hours of telephone conversation answering my ques- tions. Of these, special mention must be made of Har- ald Bode of Bode Sound Company, Scott Wedge, Ed Rudnig and Marco Alpert at Eu Instruments, Robert Moog of Moog Music, and Donald Buchla of Buchla and Associates. Last but certainly not least special thanks go to all of the composers and performers who suggested techniques and supplied scores. Without these artists there would be very little need for the instruments and certainly no need at all for this book. Allen Strange1 Preliminary Statements about the Subject Matter ‘The term “electronic music,” 2 common post-1950 term, has been the source of a certain amount of misunder- standing among musicians and audiences. Prior to the 1960s it usually referred to a type of music character- ized by presentation on pre-recorded tapes. The phrase itself often called to mind a music based on discordant sounds and angular structures. I certainly do not wish to imply any negative value judgments about early electronic music but am attempting to identify some generalized characteristics in order to explain a gen- eralized aesthetic comment. To the non-practitioner, there was little or no distinction betwee the various schools of musique concréte and ‘pure’ electronic mu- sic. While pre-1960 electronic music resulted in many significant compositions and was the catalyst of the research that Jed to present day electronic instraments, there is some basis to this generalized electronic music aesthetic that still lingers in the minds of many people. ‘Even today one might attend a concert of new ‘acoustic’ music concerned with new timbres and un- ‘usual structural relationships and hear the comment that it “sounded like electronic music’! What is the reason for this association? The early history of elec- tronic media in music was a period of experimentation with very little prior history to provide direction. Schaeffer and the musique concréte group made music ‘with whatever sound they could capture on dise, wire or tape. Eimert and Stockhausen, representing German elektronische musik, heralded the use of the electric oscillator as the source of ‘pure’ electronic music; the ‘Columbia-Princeton school was coaxing music out of $250,000 computer and capturing it on tape. What expressions in that decade had in common has had a significant role in formation of the layman's concept of electronic music today. Looking back on the years between 1950 and 1960 ‘one's first observation is that the practitioners were making music with devices designed for other pur- poses. The tape recorder, the oscillator and early com- ‘puters were designed as tools of science and not tools of art. Prior to that time our musical traditions told us that musical instruments were plucked, bowed, blown into, or struck, and these actions enabled us to reproduce a history of musical thought. When the computer, oscillator, and tape secorder were given the role of @ musical instrument, pluck, bow, blow and strike were joined by a plug in, punch, tum and splice! ‘The musician was busy making music on instromenta- tion not designed for conventional musical use. Since the generation and control of sound on these devices ‘were generally foreign to our traditions af that time, it ‘was only logical that the resulting music would be pro- portionately removed from the norm. The results ‘brought to light some forgotten basic ideas about mu- sic, specifically about musical instruments. ‘To discuss musical instruments in a manner rele- vant to electronic media calls for a re-orientation rela- tive to what we may assume to be a simple subject. We all know what musical instruments are—they are things we play and make music with! This straight- forward definition implies some rather far reaching concepts which may appear to be simultaneously quite simple and somewhat technical. However, I think that the patient reader will benefit by considering the fol- lowing ideas. Any musical instrument requires at Jeast three things: a method of playing (imput or perbaps stimulus), structural organization of the object being played, and a resulting sound (output or response). A violin is bowed or plucked, it is made of wood with strings attached and tuned in a prescribed manner and it produces a sound that we identify with that input and structure. A laboratory oscillator has a dial which 4s tumed (input), it is made of a specific collection and organization of electronic parts (structure) and it ‘Dehaves like a laboratory test oscillator. A violin pos- sesses the structure end inputting capabilities which make the production of virtually any specific pitch within a four octave range a simple matter for the trained player. A laboratory test oscillator, while cap- able of a far greater range, camnot readily produce the specific pitch patterns found in pre-1950 musical lit- ‘erature. One can imagine trying to play a Bach “Par- tite” by manually tuming the dial on the front of an oscillator. At the same time, however, it would be 1very difficult to coax a smooth ten octave glissando with a consistent timbre out of a violin. Thus the na- ture of @ musical instrument.—its input, structure and output —defines the musical characteristics of that in- strument. What really makes an instrument musical is that a musician decides to make use of it. ‘This may all seem rather obvious, but may serve to answer some basic questions and provide us with guidelines to the functions of electronic media in the sonic arts. During the decade between 1950 and 1960 composers and performers were making music on in- struments not designed for conventional performance. The task of inputting or playing was difficalt, and per- formance of any kind of pre-existing musical literature was almost impossible. Even simple sequential event structures were difficult to achieve. In this situation fone can readily understand that one reason for the seemingly radicel sounds of early electronic music was partly due to the fact that the musicians could gen- erate such sounds with relative ease. Music exists in continuum of time and time will not wait for one to find the next pitch on an oscillator or program 2 new set of instructions for 2 computer. There were two obvious solutions to the problem. One was to build a new kind of music to accommodate the type of instruments used. If precise pitch at specific times was difficult to attain, then make a music that took advantage of the things these instruments could do— such as extreme glides, non-flumuating timbres, ex- panded dynamic ranges and so on. Hence—our first models for the “electronic music” The altemate so- Tution to the problem was, in essence, to stop time in order for the performer or composer to rearrange his collection of instruments, making ready for the next event. In the field of electronic music the tape re- corder is an instrument which provides for this need. To produce a precise sequence of pitches on a labo- ratory test oscillator one finds the starting note, re- cords it, turns off the tape recorder; finds the second note, records that, tums off the tape recorder; finds the third note and repeats this process until the se- quence of desired notes are on the tape. He then cuts the tape into lengths which produce the desired rhyth- mic pattem and puts it all back together again with splicing tape. The tape recorder then reproduces five second stream of pitches the composer needed thirty minutes to construct. The reason for this out- ‘of-time performance was that the generating instrv- ment, in this case an oscillator, did not have the putting capabilities needed to produce the desired ‘event in real-time. The two obvious solutions to this problem were to evolve the literature to suit the in- struments or to develop the instruments to suit the literature. The evolution of technology and conscious- ness provided both. The development of new musical 2 instruments has continued to remind us that we can make music out of anything from which we choose to make music. There is more to the art of sound than twelve pitches, bowing, blowing, plucking and strik- ing. Many of what seemed like definitely unmusical events in 1950 are now quite acceptable, even to the conservative ear. The revolutionary music from the Columbia-Princeton studios during the early and mid- le 1950s is very tame compared to today’s orchestral music by composers such as Iannis Xenakis. Along with the evolution of our aesthetic, the instrumentation itself underwent significant development. Alterations in oscillators were made, which allowed the performer to control precise pitch change in real-time. We learned that workable performance modes could be developed for the musician. Oscillators and other artifacts once designed for the communication studio were rede- signed to accommodate a wide variety of musical thought. By the early 1960s electronic devices ded cated exclusively to the production of music were available. The composer in the electronic medium no longer had to work within narrowly defined limits, nor was it necessary to manipulate the flow of time ‘with a tape recorder. Contemporary electronic musical instruments are devices capable of a wide range of inputs, structures and resulting sounds—all of which are decided and implemented by the composer/per- former. A trip to the local record store provides substan- tial insight into the current state-of-the-art. In the bin marked “Electronic Music” one can find works by John Cage, Milton Babbitt, reorchestrations of mas- terpieces ranging from Monteverdi to Stravinsky, a variety of popular artists ranging from Herbie Han- cock to Klaus Schulze, contemporary masterworks by ‘Morton Subotnick, the Sonic Arts Union, and Karlheinz Stockhausen; and I am sure this bin has grown sig- nificantly since this text went to the printer. If such a wide variety of music is listed as “electronic music,” then the term cannot possibly refer to any single aes- thetic production. What am I talking about when I say this is a book about electronic music? What is it ‘that all of thase records have in common? The answer I provide and one which is the basic premise of this text is instrumentation and orchestration. The purpose of this book i to guide one in the technique of playing electronic instruments. Beware of the simplification of that statement. The subject matter is a bit more complicated than might be imag- ined at this point, and the complication is due to the nature of these instruments. A woodwind, string, brass or percussion instrument is structurally constant. Per- formance modes have been established which accom- modate their structure and their sounds can be pre- cisely predicted. The reason a clarinet, trumpet, viola, whatever, exists as a workable musical tool is becauseit has a fixed structure which is based on a given set of performance modes to provide a consistent response. ‘Whenever the composer, through whatever suitable notation, instructs a performer to produce a certain pitch by bowing a string, he is calling for a known event to be produced. If he tells the performer to change the nature of his input, perhaps to pluck the string, he is still operating within a set of known per- formance modes and resulting sounds. This is pos- sible because the instrument is a fixed instrument—its structure being the result of its technical evolution tempered by the demands of its evolving literature. Most electronic instruments, specifically the ones to be dealt with in this book, are not fixed instru- ments. The contemporary electronic music system has no pre-defined structure, but is initially a collection of possibilities—a set of musical variables or parameters such as pitch, loudness, space, timbre, etc., that exist in an undedicated state. The contemporary electronic rusic system offers several potential means for pitch control; it has capabilities of shaping articulation and loudness; it provides many methods for the control of timbre and density, and so on. All of these pos- sible sources and controls are components of a vet to bbe produced musical event. The art of electronic mu- sic involves organizing these parameters into desired structural relationships, coming to some decision about performance modes suited to the situation, and final dealing with the task of actually playing the instru ‘ment to produce the musical experience. For example. 4 composer or performer may require a two octave ascending glissando from an instrument that has @ very harsh, metallic sound. In one situation it may be convenient for the musician to produce that event by tuming a knob, In another situation it may be more convenient to produce the same event by touching a key, illamination of a light, or even by slamming a door! Or perhaps he wishes for that door slam to set off a 16 note sequence of pitches, while playing a Keyboard causes the sound to spin around the room in various patterns. In another situation the harshness of the sound may be determined by some internal biological function such as the performer's brainwave activity or perhaps by an external natural function such as the air temperature in the room. The real task at hand is the conception of the musical structure and event and then deciding what sort of stimulus oF ac- tion will call forth and/or alter those events. To the reader not familiar with current electronic music lit- erature or instrumentation these examples may appear to be far fetched. As unlikely as they may seem, they sre viable methods of controlling electronic sound and serve to illustrate the range of performance modes open to the musician today. In the earlier edition of this book I reacted against the use of the word “synthesizer” as applied to con- temporary electronic musical instruments. Too often that term has led to such phrases as ‘synthetic music’ and ‘synthetic sound.” Although I hold with my initial viewpoint, the term is meaningful but in a different sense than was originally implied. If synthesis refers to building up from component parts, what is being synthesized are musical instruments, ( Perhaps the in- struments should be called “synthetics” and the players could then be called the “synthesizers!”) There are some final observations which should be made in defining the scope of the subject matter in this book. This is not a text on music theory. com- position, or detailed recording techniques. Thus it is assumed that the reader possesses the basic academic skills such as note reading. This work is written for the musician from a practical musical viewpoint. It explains how to approach, structure and play very {idiomatic instruments, An experienced composer knows that itis difficult to compose for idiomatic instruments unless he has some working knowledge about them. The guitar and harp are good examples of this. If one wishes to compose for an instrument he should have a basic knowledge about what that instrument is cap- able of doing and how it is done. Dealing with the concepts of parametric organiza- tion, voltage control, and tracing patch connections is certainly enough for the beginner, without having to bbe concerned with the graces of compositional ges- ture, at least in the initial stages. This is not meant to remove composition from the realm of this study. but rather to point out that dealing successfully with composing for an idiomatic situation requires ade- quate working knowledge of that idiom. The approach to be used here is to deal primarily with the task of Teaming the sonic vocabulary of the electronic me- dium and developing the intellectual and performance skills needed to bring those sounds to life. The act of composition involves primary decisions about ‘when’ and ‘how long’ an event is to occur. The ‘when’ and ‘how long’ are aesthetic decisions bome out of compo- sitional attitudes. This book can instruct one about how to make sure things happen ‘when’ they are sup- posed to and offers techniques for controlling “how long” Although certain compositionally based etudes and tutorially valuable scores will be suggested as a means for exploring a specific technique available on the instruments. large scale compositional assignments are best left to the discretion of the instructor where they can be tailored to meet individual situations. This is also not to say that electronic instruments have not contributed new ideas to the discipline of composi Since the art of electronic music involves structural design it brings with it some new methods and slants on traditional manners of organization, Electronic in- strumentation hus also given us anew handle on the 3parameters of time and space, actually offering us some new things to compose with. These areas will be covered, and their suggested applications hope- folly will lead to some new ideas for the composer. But it should be recalled that the purpose of this book is to teach the user to play the instrument. Not every- one involved with electronic music is especially in- terested in composing. And I am quite sure that the more competent performers of electronic instruments there are, the happier the composers of electronic music will be. Parametric Design A teacher of any subject matter has a two-fold respon- sibility; to be proficient and knowledgeable about the subject matter; and to be familiar with the pedagogy of the subject. As of this writing I know of no generally accepted pedagogic method in the area of electronic music. How does one go about learning and teaching in this field? Perhaps the very nature of the subject as it has been expounded in the previous pages makes a single teaching approach impossible. A consistent approach is only possible when one has become aware of the consistencies in the subject matter. In the previous paragraphs I have gone to some length to illustrate what T believe to be a consistent concept in this area. This is the instruments’ ability to be structured according to an immediate musical need. A successful practitioner in the field of electronic music is one who can: 1) envision a musical event (either of his own or others’ invention); 2) demon- strate the musical knowledge and technical skills to set up the instrument to produce the required events; and 3) finally possess the artistic sensitivities to bring the event to life—independent of whether he is in a compositional studio environment or a real-time per- formance situation. This book is designed to deal with ‘two of these three areas. The reader will be exposed to certain operational principles which are cormmon to all current electronic musical instrumentation. These principles are based ‘on two concepts: parametric design and voltage con- trol. The study of brass instruments teaches one to deal with the production and control of sound in terms ‘of resonating tubes. Likewise, the early chapters will show the reader how to view a musical event as @ set of electrical analogs. The pitch, loudness and tim- bre of a vibrating string can be described in terms of that string’s activity. In the same manner, the pro- duced pitch of an electronic instrument is described as electricity behaving in a certain manner. Electronic instruments provide a collection of circuits (in some cases called “modules”) which provide control over ‘one or more elements of a musical situation. Struc- turing a musical event is a process of isolating those 4 elements needed for the particular situation at hand, and then prescribing a set of controls that will enable those elements to work together in the most efficient manner possible. Parametric design is an analytical process in which ‘one envisions the individual and corporeal influences of all of the parameters of an existing or imagined sound, One will leam that parametric activities are not isolated from one another. Pitch influences loud- ness, loudness influences timbre, timbre, in turn, in- fluences pitch, and around the circle goes. Thus, in deciding how to obtain the desired results from an electronic instrument one must be familiar with the roles and limitations of each part of the instrument, as well as possess a basic understanding of the psy- cho-acoustic nature of what he is going after. As these clectrical analogs of sound are explained, they will be accompanied by whatever psycho-acoustic information is needed to make the behavior of the instrument meaningful in a musically perceived situation. Once one is able to describe an event parametrically that description bas to be transferred to the instrument. This is a physical or kinesthetic process of intercon- necting various parts of the instrument through switches, patcheords or whatever, and tuming knobs. ‘The mathematics used here will not go beyond ele- mentary multiplication; and discussion of internal cir- cuitry will be almost non-existent. Those interested in the more detailed and technical aspects of this field are referred to the annotated bibliography at the end of this book. ‘Voltage Control involves the physical assignment of various types of electronic activities to specific parts ‘of the instrument, and the production and routing of those controls comprises the actual performance techniques. Such controls are in the form of steady and fluctuating voltages. Some voltages are pre-pro- grammed to come into play by the touch of a switch, others are more efficiently handled by manual means in real-time. Voltage control techniques are given ‘prime attention in this course of study, since this is the basic operational principle behind current elec- tronic music instruments. The subject matter of this book ends at this point. tis within the scope of the text to prescribe meth- ‘ods of electronic sound production, to provide ‘lus- trative models by means of exploratory etudes, and to suggest supportive literature in the form of re- cordings and scores to stimulate creative thinking. It is not within the scope of this text to dictate musician- ship or aesthetic direction. Musicality is something the student or teacher must bring into the situation on hhis own. Aesthetic applications are so numerous that I would not attempt to represent any others but my own, and since my own are those only I am expected to agree with I prefer to avoid the subject as much as possible. Any references to literature will be interms of successful models of specific techniques; and the student is urged to expand on these models. The student is urged to not be limited by such references, this text included, or by the front panel graphics of an instrument. Txy everything! Never wonder “should T do this?” Instead an attitude of “I wonder what will happen when I do this?” may lead to an expansion or development of a unique technique. A designer may manufacture a device to satisfy 2 specific musical need, and an innovative musician may discover that the de- vvice suits many other needs which never entered the mind of the designer. For the most part current elec- tronic instruments are “student proof,” so that any- thing one has access to through the front panel can- not cause any damage. In the case of a blown circuit fone can be comforted by the fact that electr struments are usually easier to repair than acoustic cones. Please don't interpret these remarks as though the instruments are indestructible. Realize that they are musical instruments and require the same care as, any fine instrument, But stil, if you have an idea— give it a try. After the basic concepts described above have been dealt with, the subsequent chapters will be individu- ally dedicated to the exploration of techniques and processes relevant to specific parameters’ and situa- tions. At this point there arises an organizational situa- tion which only the user can solve for himself. A com- posite musical event treats all of the parameters collectively. Pitch, loudness, tone color, space, etc. all happen at the same time. But a successful approach to the electronic media demands individual attention to these parameters, at Jeast in the initial stages. We have to consider pitch, loudness, ete. as independent variables in the leaming stage, but we all know that music does not occur as a set of individual variables. If I were trying to describe how to imitate the music of 2 particular composer it would be a relatively easy task to approach and organize all the parameters in ‘2 manner that would articulate 2 particular style. The same can be said of a situation in which one was also concerned with teaching the basics of music through clectronic instruments. Nevertheless, in dealing with the “how to” of highly idiomatic instruments the teach- ing reference should be as variable or “modular” as the instruments it represents. To some situations it may be more useful to deal with the chapter on tape recording as an introduction to the subject area. Those interested in live performance may find that tape tech- niques have litle to do with their anticipated appli- cations. Those with limited instrument facilities may find the chapters on modulation techniques go into much more detail than the instrument can accommo- date, while persons with access to larger instruments will find these chapters essential. ‘Whatever the case, if you continue to develop your ‘techniques and instruments, all of the information here will be applicable at some point. therefore, don't sell the book when you finish it! Patching and Notation Since the synthesizer is not a “fixed” instrument the musician working with electronic instruments has a two-fold notational responsibilty: 1) the notation of the musical events and 2) the notation of how the instrament is configured. In spite of many national ‘and international attempts to codify a notational sys- tem each artist still continues to use whatever method suits his or her own needs (this actually depends on the templates and press-on letters the artist has avail able at the time!). This text will be as consistent as possible, but it should not be taken as an attempt at universal codification. The evolution of music itself resulted in and is still causing changes to the notational system. As new instruments and new processes are developed the various patching notations will change. At the present time, however, the system used in this text is clear and easily understood. The prime con- cern of this system is that the notation be relevant to practically any electronic instrament. Each instru- ‘ment provides sound sources and means of modi ing the sounds. Each system requires the performer to establish certain norms—to what pitch or frequency are the oscillators tuned? Where in the spectrum is the filter set? What is the initial setting of the amplifier? All of these variables are referred to as offsets. Such indications will be notated, when needed, directly next to the module to which the notation refers. As you will learn later in the text, all of the variables on an instrument are usually controlled in some manner by different functions on the instrument. A given con- trol may affect a great change in a parameter (a four ‘octave jump in pitch), or a minute change in a para meter (2 X tone change in pitch). The magnitude of the change is determined by how much a control is “attenuated,” meaning literally to lower or reduce. In other words a given control may be “attenuated” or “processed” before it reaches the part of the instru- ‘ment it will actually control. Thus the various “attenu- ation” or “processing” indications have to be notated, along with the offsets.” Where attenuation specifica- tions are critical the levels will be indicated next to the affected module. If the attenuation is variable the standard electronic attenuator symbol will be used as illustrated in fig. 11. AA Figure 1.1, Attenuator aaa C— BS Oo eee eee ZI seen Figure 1.2 illustrates the symbology for the basic de- vices found on electronic instruments. Although some of these terms may mean absolutely nothing at this point they are illustrated here because you will en- counter them all very soon. Any specialized or umusual device encountered in a patch will be clearly iden- tified in the diagram. Throughout this text there are several patch charts which are designed by the manu- facturer for specific commercial instruments. Where needed a patch diagram of the chart indications is ‘given so configurations may be transferred from in- strument to instrument.the Basic of Sound ‘The purpose of this chapter is to give the reader a basic understanding of the technical side of the sub- ject matter and to provide a point of departure for those who wish to rearrange the chapter order to suit their particular situation. Basie musical parameters are understood by all musicians; we relate to the basic variables of pitch, rhythm, loudness, timbre, et, in- asmuch as they are constantly under our control in conventional performance practices on acoustic instru- ments. Simultaneous control of relevant parameters is 2 skill developed through dedicated practice, and such control is a leamed tactile response to some stimulus— bbe it a notated phrase, 2 conductor's direction, or one's ‘own internal musical sense. Terms such as “more le- gato,” “more articulate,” “richer.” “funkier,” “straigh- ter,” ete. are adjectives or adverbs that call for alter- ations in almost all the ongoing parameters. To play “dirtier” or “cleaner” calls for alterations in pitch, loud- ness, articulation, and so on. This sort of response is 2 skill learned as a single performance variable through long term training that deals with simoltaneous at- tention to many individual variables. We are tradi- tionally taught to alter our playing response (ic, style, techniques . . . whatever word is applicable) in re- action to a single stimulus. If the playing is not cor- rect it usually means that many variables have to be altered in various degrees. When adjusting to a “ba- Toque” manner of playing, 2 trained performer almost instinctively makes the proper adjustments without really thinking about all of the variables he is deal- ing with; his prime stimulus is the total musical sound. ‘As pointed out in the previous chapter, the one function that really characterizes electronic media for the musician is parametric thinking We are consci- ously and separately attending to many different para- meters. One must individually, yet simultaneously, be aware of and in control of pitch, (an oscillator), dy- namics and rhythm (an amplifier), timbre (usually a filter), and a host of other variables still to be intro- duced. Taking care of all of these things at one time is a sizable chore, both mentally and physically. One aid in this task is the ability to view all the variables Considerations of Parameters 1 a single concept. For example, a certain situation may call for a fuller, louder sound. Loudness is a pe- rameter which has several contributing factors (as do all the parameters), but in terms of the physics in- volved it means ending up with more acoustic energy reaching the ear. In one case an increase in loudness may require tuming up an amplifier. In another case it may involve decreasing the amount of mechanical reverberation, and in still another situation it may re- ‘quire a change in the timbral quality of the sound. All of these operations in turn involve @ change in how the instrament is behaving, the end result being ‘an increase in the energy that reaches the ear. Thus ‘the musician is ultimately concerned with different mamners of vibration, Vibrations and Musical Sound ‘The production of any kind of sound, musical or other- ‘wise, is due to rapid vibration of some object. With acoustic instruments 2 string, reed, membrane, etc. is forced into vibration and this causes its immediate surrounding air space to similarly vibrate. When these vibrations ultimately reach our ear they are perceived 1s sound. The characteristics of a particular sound are largely, but not exclusively, dependent on the man- ner in which the vibrating object is behaving. The object's physical structure, how fast it vibrates, how forcefully it vibrates, and several other factors, deter- tine the final sonic characteristics that object is re- sponsible for. Musical performance techniques can be thought of as actions applied to an instrument (in- puts) which affect how that instrument vibrates. A ‘trumpet player may press a valve to change the vibra- tion rate of air in a metal tube; the same player blows harder to increase the force or amplitude of those vi- brations. A string player changes finger placement on a string to change that string’s vibration rate and si- multaneously may press harder or softer with the bow to affect the amplitude of that vibrating string -(its loudness). A mute placed across the bridge of a violin or in the bell of a trumpet is another performance 7{input that changes the way the instrument responds, and this has various effects on tone quality. Many mu- sical parameters involve the generation and alteration of vibrations. Anything that vibrates within a certain frequency range and with enough force has the potential of be- ing 2 sonic event. If you happen to be sitting in 2 room with fluorescent lighting at this moment you will probably be able to hear a faint pitch around the area of Bp. Florescent lights actually turn on and off at a rate of 120 times each second; this vibration rate is fast and strong enough to be perceived as a defi- nite pitch. The device which helps produce the image on a home television screen scans back and forth on the picture tube at 2 rate of 15,750 times each sec- ond, producing a very high pitch. This fast frequency can be very annoying to persons with sensitive hearing. Musical pitches we refer to with letter names such as Bb, F, Ci, ete. actually describe different rates of vibration. The standard tuning reference for instru- ments in America assigns to middle“A” a vibration rate of 440 times each second. Every musical pitch refers to a specific rate of a vibrating string, tube, mem- brane, electronic circuit, etc. Sound is multidimensional. One canndt perceive pitch without perceiving 2 sensation of loudness, tone ‘quality, duration, and apparent source. Pitch is fairly easy to deal with because we have some well defined references; consequently it has been the most acces- sible musical parameter for the composer. Loudness is less clear, since there are not as many well defined limits and references. Loudness is perceived primarily as the result of how much or how forcefully something is vibrating, Stated a little differently, loudness is the result of how much air is displaced by the vibrating object. Blowing softly into a clarinet will produce 2 soft sound. Blowing harder into the instrument causes the reed to vibrate with more energy, resulting in a ‘greater amount of air being displaced in the immediate environment, and hence 2 louder sound. Loudness does have 2 specific unit of measurement called the decibel; this will be discussed in chapter 9. Timbre, or tone quality, is partly determined by the pattern of vibration. A string, when bowed, vi- brates in a particular way producing the sound of a bowed string. A plucked string has a markedly dif- ferent sound. It may be vibrating at the same rate and with the same energy as the bowed string, but its ‘manner or pattern of vibration is different. Pitch, loud- ess, and timbre are terms we assign to different as- pects of a vibrating object. Pitch refers to rate, loud- ness to perceived energy, and timbre, in part, refers to the pattern of vibration. All of the foregoing analogies have been in terms of familiar acoustic instruments, and transferring these ideas to electronic instruments is 2 simple matter. 8 Electricity is a source of energy that can be specified and controlled. Through various types of circuit de- signs and controls, electricity can cause objects, ust- ally speaker cones, to vibrate in specified ways. One ‘can design a circuit to produce energy fluctuations at certain rates, amplitudes, and pattems. When these fluctuations are transmitted to a speaker cone the speaker transfers these vibrations into the air, and from that point the sound takes essentially the same path to our ear as any other sound Once the elec- trically generated signals have become translated into airborne vibrations their behavior is independent of the sound source. Electronic sound is only “electronic” in terms of generation and control. The generation and control of sound on electronic and acoustic in- struments have conceptual similarities—a sensitive cel- list is continually concerned with how fast, how bard, ‘and in what pattem the strings are vibrating. The mu- ssican relates to electronic instruments in precisely the same manner; he is concerned with telling the elec- tronic circuits how fast to produce energy changes, the amount of energy to be transmitted, as well as the various shapes and pattems of energy changes pro- duced by the electronic circuitry. Musical Structure and ‘Temporal Measurements All masical processes can ultimately be defined as tem- ‘poral pressure variations perceived by the ear. The mind's ear is continually making measurements and comparisons of information on multi-dimensional lev- els. On one level we may observe the length of 2 composition, movements, or phrases. Such long term measurements are usually spoken of as form. On an- other level we measure the durations of and intervals between individual notes and call that rhythm. On still another level we measure the number of air fronts moving past our ear in order to establish the identity ‘of a single pitch or composite sound. On another di mension a stronger or more forceful vibration will uso- ally be perceived as 2 louder sound than a weaker vibration. The parameter of timbre is indeed enig- matic and eludes precise definition. As mentioned be- fore, timbre is related to the manner in which an ‘object vibrates, but this is only one of several con- ‘tributing factors in timbral identities. Such compli- cations are subsequently elaborated in cited refer- ences For the present we may accept the statement that timbre is @ dynamic parameter subject to an in- finitude of changes or variations in time. 1 For further reading in the area of musical timbre refer to Robert Erica's Sound Seructure in Music, University of Cali fornia Press, 1975.Although perhaps premature at the time in terms of available technolosy. the theories of Karlheinz Stockhausen during the middle 1950s, merit consid eration by the practitioners of electronic music, if not for all media. Stockhausen’s writings detail the pre- viously mentioned idea that long durations of time (macro-time) contain musical forms; shorter temporal variations contain phrases or motifs, while still smaller ‘temporal divisions enter the realm of rhythm. A thythm, if increased to about 18 times a second, be- gins to be perceived as a pitch (micro-time). Continu- ing with this same manner of thinking, a higher pitch (aster rhythm?) enters perceptual domain provid- ing some, but not all, information about timbre. Stock- hhausen's theories imply that form, phrase, rhythm, pitch, and timbre are all the workings of a single sys- ‘tem-vibrations or variations, ~The musical organization is carted into the vibra- tional structures of the sound phenomena. The sound Phenomena of a composition are an integral part of this organization and are derived from the laws of structure: namely, that the texture of the material and the struc- ture of the work should form a unity; and the microtonal ‘and macrotonal form of the work have to be brought into a conformity that accords with the basic formal idea ‘which every single composition has.”= The amount of energy or force behind any of the vibrating domains influences loudness. It is notable that this energy affects each of these perceptual do- ‘mains in the same manner. We will leam that in- ‘creased energy in the higher frequencies will alter the parameter of timbre; increased relative energy between 18 and about 17,000 times a second results in louder pitches; energy variations below 10 or 15 times a sec- ‘ond affect the dynamics of rhythm (accents), phrases, and formal contrasts. All musical instruments are dedicated to the pro- duction of variations in air pressure at various rates, with forces, and in differing patterns. A vibrating string, reed, lip, or membrane ultimately results in @ sensation we hope will be of musical interest. Assum- ing that the listener is perceiving sound in an acoustic space we know that the air is conveying the vibra- ons to our ears from some vibrating source like a string, reed, lip, or membrane (such 2s a speaker cone as is the case with most electronic instruments). 2, Quoted by Serpo Heiknhcino in The Elecronic Music of Korlheins Stockhausen (Studies on the Esthetical end Formal Pobloms of te Fist Phas). (Bryo Mave, Penna Ata Ms feologia Fennica, ‘Theodore Presser Co,, 1972) p. 15 from Teme at eletrntchen und isramenalon Mest (Volume 1) Non-Linear Perception ‘Most human sensory perception ix won-linear. A phys iological response such ax hearing or sight is not di rectly proportional to the stimulus (sound or light) causing that response. Non-lincarity means that equal changes in some stimulus (vibration rate, intensity. etc.) do not result in equal changes in perception (pitch, loudness, ete.). Pitch is the result of a given rate of change or vi- bration of some physical object—in this context a speaker cone. A vibration of 65.4 vibrations per sec~ ond, or Hertz (abbreviated H=) registers the pitch sensation of low C. By doubling that frequency to 150.8 Hz the C an octave above will be heard. The next octave C, middle C is 2616 Hz, or four times the original low C. Note that between C, and C, there is a difference of 654 Hz (1308 — 65.4 = 65.4), but between C and middle C there is a difference of 130.8 ‘Hz (2616 — 1308 = 190.8). Continuing up in octaves ‘one may observe in figure 2.1 that each octave is twice as large as its lower neighboring octave when mea- sured in Hertz. Equal changes in perception (octaves) require uneven changes in stimulus (each octave be- ing twice as large as the previous). This type of non- linearity is called an exponential progression because pitch sounded in subsequent octaves increases by fac- tors of 2 Pitch perception is consistently based on such an exponential progression. Note that in figure 22 the lower perfect 4th, C3 to Fs, is a difference of 43.8 Hz, while the same musical interval three octaves higher is 3504 Hz. The unit quantity of the stimulus, Hertz, increases by a constant factor of 2 every oc- tave. To calculate what the Hertz measurement of @ perfect 4th from C to F (43.8 Hz) would be four oc- taves higher, one would multiply that number not by four, but by two to the fourth power (2) or 16. Hence, scientific measurement of intervals can be carried outin Hertz, but in terms of our ears” response, higher {intervals contain more Hertz than identical musical in- tervals in lower octaves. ‘Changes in loudness are due to perceptions of the change of physical strength or amplitude of vibration. The more energy contained in the air fronts moving ppast our ears, the louder the perceived sound. Like pitch, the perception of loudness is also non-linear. Loudness is measured in units called decibels (abbre- viated db); this is the smallest unit of noticeable loud- ness difference the ear can detect. The decibel is usu- ally used as a measurement of relative loudness be- tween two events. If 1 db is assumed to be the softest possible sound, then 60 db would represent the loud- ness level of a normal conversation at a distance of about three feet. However, a db level of twice that figure, 190 ab, is not twice that loud, but 1,000 times as loud! For the mathematically minded the decibel equals 20 logis P1/P2. Pl and P2 are the two differ- ence levels being compared. At this point it is only necessary to realize that the decibel is a non-linear unit of loudness measurement: perceived equal changes in loudness taking more energy at louder levels ‘Subjective and Objective Measurement ‘The perception of vibrations may be dealt with either subjectively or objectively. An objective measurement would be the observation of such vibrations against a precisely calibrated measuring device, and under ‘every condition that same rate of vibration would al- to keep in a ort) pe: the reader oreo to tas aie or layman's study of the decibel 10 ‘ways measure the same. For example, speaking of pitch in terms of Hertz is an objective measurement. A = 440 Hz is an objective statement because the reference, a period of 1 second, is not variable. When, ‘however, these vibrations are forced through a variety of media (around comers, through walls) under a variety of conditions (different loudnesses, timbres, etc.) the subjective measurement, what we actually perceive and register, may not agree with the objec- ‘ive measurement. It is not the intent of this book to dealve into a detailed study of psycho-acoustics, but it is important for the musician to realize that there is a difference between objective and subjective mea- surements! Objective appraisement involves measure- ments against a consistant norm: subjective appraise ment is @ perceptual measurement which can be in- fluenced by many variables. Frequency is an objec- tive measurement, but pitch is a subjective measure- ment. Frequency is objectively measured in Hertz, and pitch is subjectively measured in musical intervals such as thirds, fifths, octaves, etc, or in specific pitch references such as By and Ct. Amplitude is an ob- jective measurement of the subjective phenomenon we call loudness. Amplitude may be messured as voltage levels, and loudness may be measured in terms of decibels or traditional musical dynamics such as piano and forte. In some cases decibels may be objectively measured with various types of meters, but speaking practically, the db is a measurement of what we hear. ‘The various conditions that alter our subjective per- ceptions will be discussed in situations where those variables can be put under some sort of control. ‘5, For further reading in the area of psychoacoustic musical Shsnemeos te readers planed to Jun G. fosderers late ‘duction to the Physics end Paychophasice of Music (New York, Springer-Verlag), 1973.3 Electronic Sound Sources and Their Characteristics Logically, the production of 2 musical event begins with the generation of an initial sound. The performer provides certain information to the instrument, and ‘the instrument responds by producing a characteristic and often musically raw sound. This can be demon- strated by listening to a beginning plaver on any in- strument. The basic sound is modified by the phys- ical properties of instrument and also by additional performance nuances supplied by the player. This process applies with equal accuracy to both electronic and acoustic instruments. Although any sound can be modified and disguised almost beyond recognition, its initial characteristics will suggest applicable modifications in terms of the desired result. If an orchestrator or arranger wants @ delicate, shimmering effect more than likely he will not begin with a tuba! A major skill required of elec- tronic instrument composers and performers is orches- tration. Obtaining the desired result is dependent on beginning with the most effective source. Professional electronic instrument programmers? must be aware of the physical and sonic characteristics of the basic lectronic sounds before-beginning to add the subtle ‘muances and processing called for by the producer or performer. Voltage and Sound As the fiddler’s bow hairs force the violin string and sounding board to vibrate and cause the air pressure in that string’s environment to fluctuate, electrical volt- age causes the paper or metal cone of a loudspeaker to vibrate—again resulting in variations in air pressure. Electric current is 2 measurement of the flow of elec- tric energy. Voltage is the force that causes the cur- rent to flow through a wire. Although voltage and current are measurements of two different electric ac- tivities they are mutually related. At this point it is 1._A programmer is 2 musician responsible for “making the patch” on the instrument for a performer. People such as Mike Boddicker and Jan Underwood are highly respected progam ser and are success due to their ability Yo work qucty and efficiently in the studio not really necessary to distinguish between the two. If a speaker cone is connected to an altemating cur- rent, certain physical changes take place. When no voltage is applied to the speaker, the speaker cone is in a neutral position (fig. 3A). When a “positive” voltage is applied to the speaker, the cone is pushed ‘outward (fig. 9.1B) and then, as the positive voltage decreases, returns toward its original position. As a “negative” voltage is applied to the speaker, the cone is pulled back to a point opposite the positive voltage Position of the speaker (fig. 31C). Every time the speaker cone is moved by the alter- nating current-AC (positive, passing through a neu- ‘tal position, negative, and again passing through a neutral position) masses of air or pressure waves are moved past our ear, producing the sensation of pitch. If the speaker cone is moved back and forth 440 times in one second, for example, we hear a sound which 4s commonly referred to as “tuning A” If the cone moves back and forth at a rate of 261.6 times a sec- cond, we perceive a pitch of “middle C” Each back- ané-forth movement caused by the application of @ positive voltage followed by a negative voltage is re- ferred to as a “cycle,” or a “Hertz” Therefore, 440 Hertz (Hz) produce “tuning A” ‘The volume, loudness, or amplitude of a sound is determined basically by how far the speaker cone is moved back and forth on its neutral axis, which is 2 result of the voltage level of the AC. If an alternat- ing current hypothetically displaces a speaker cone 1/4 inch in each direction from its neutral position n261.6 times a second, the ear will perceive “middle C” at a certain loudness level. If the speaker is displaced 1/2 (2/4) inch from its neutral position at the same rate of frequency, the ear will perceive the same pitch but at a louder volume or, in objective terms, at a ggeater amplitude (see fig. 32). It should be noted that amplitude does have certain effects on pitch per- ception. For frequencies about 1k H2* there is a small but perceptible correlation between pitch and loud- ness. Careful listening will show that above 1k Hz the pitch will rise slightly as the loudness of that sme frequency increases? And the inverse is true in that the pitches in differing registers at the same ampli- tude (an objective measurement!) will be perceived at different loudnesses. For example, high C* Z= (1046.4 Hz) at an amplitade perceived as “piano,” when transposed down four octaves to C; == (654 Hz) at the same amplitude is barely audible, Besides be- ing @ good example of the difference between sub- jective and objective measurement, this also illus- trates the concept of dependence of various musical parameters on each other.‘ ‘Characteristics of frequency and amplitude are represented graphically in figure 33. The line of zero voltage represents the speaker cone in the neutral po- sition, or no movement of air. The horizontal direction of the line represents the passage of time. The plotted carve on either side of the zero-voltage line represents the back-and-forth movement of the speaker cone, or positive and negative voltage, which in relation to the time each cycle takes represents frequency or pitch. ‘The height of each cycle gives an indication of amp- litude or loudness. Figure 3.34 is the same frequency but greater amplitude than figure 338; figure 33C is the same amplitude but lower frequency than figure 338. 2, “Kis an abbreviation of the term “Ilo,” meaning. 1,000. Therefore Ik Hz means 1,000 Hz. 3. For information oa this subject the reader is referred to the study by SS. Stevens, “The Attributes of Tone,” Proced- tres of the US. Academy of Science, 20:54, 1904. 4. Based on information given by Roederer, op. ct. 2 The Basic Oscillator ‘When the musician wishes to produce a specified pitch he sets an “oscillator” to generate a voltage oscillating at the desired frequency. Oscillators are often cali- ‘rated with a dial or lever that corresponds to the desired frequency. Often there is no calibration and tuning is left to the ear. The dial is usually referred to as a “pot,” which is an abbreviation for “potentio- meter,” a resistance device that in this instance con- trols the frequency of the oscillator or rate of vibra- tion. There may be a second pot to control the amplitude of the signal, but in many cases such an amplitude control is extermal to the oscillator. This ‘will be discussed later. At this point, it will suffice to know that an oscillator is a frequency-producing device that has the capability of producing any de- sired single frequency. ‘Our range of hearing perceives only those frequen- cies between 18 Hz and 22k Hz There are many dif- fering opinions about the actual audio range. These different statements of audio perception range from 2 low of 16 to 30 Hz and a high of 18k Hz to 30k Hz.) ‘As we shall see later, however, frequencies far below and above the audio range are necessary to the pro- duction of many types of sound. Oscillators that spe- cialize im frequencies below our hearing range are mown as “subaudio oscillators” or “modulation oscil- lator” or “low frequency oscillators” (LFO's) (see page 13) and generate frequencies as low as one eyele every minute and lower. Oscillators specializing in frequencies immediately above the audio range are referred to as “ultrasonic ascilltors.” The three types of oscillators generally overlap in frequency range. The fdeal oscillator for use in electronic music is one that will cover all three frequency ranges with the same degree of accuracy.Figure 3.44 is the front panel of the Moog 921 05- illator with rotary pots and 248 is the Buchla 208 os- cillator with linear or slide pots. Some performers pre- fer the slide pots as they give a more graphic indi- cation of the produced pitches—the higher the knob the higher the pitch; rotary pots have the advantage of higher “setability.” With the first and second generation electronic mu- sic oscillators there was often a trade-off between sweep range and stability. An oscillator with a single continuous range from 5 to 20k Hz might suffer from lack of accuracy and drift (uncontrollable pitch varia- tions). To minimize this problem some oscillators are designed to have their total range divided into octaves by a calibrated switch, such as the Moog 921 in figure ‘4A. In this case, the octaves are indicated in “feet,” taking a cue from standard organ terminology. The exact pitch within each octave can then be man- vally determined by the frequency control. This par- ticular oscillator has the additional capability of serv- es “2 KEYBOARD ry) fine sign tuning ing as a sub-audio oscillator by setting the “coarse range” switch to “sub-audio.” “Coarse-range” refers to 1 very broad operational range. In the case of the Moog, 921 the two course ranges available are 01 Hz (one cycle every 100 seconds) to 400 Hz and 1 Hz to 40k Hz. Other oscillators may have ranges calibrated in Hertz. The Electro-Comp EML-200 has six switchable ranges: .O Hz, .1 Hz, 1 Hz, 10 Hz, 100 Hz and Ik Hz. The Buchla 258 oscillator is a single sweep oscil- lator with @ range from about 2 Hz to 20k Hz (this ange is extended through the application of external controls). Fine tuning may be accomplished by means of a small pot known as a trimmer. The Synthi VCS-3 provides a single sweep oscillator with a range from 1 He to 10k Hz, calibrated with numbers from 1 to 30. In this case fine tuning is done by means of a ten-tum rotary pot. This means that the pot can be completely rotated 10 times—one complete rotation covering only 1/10 of its total range. This is advan- tageous in the sense that it provides more manual pre- cision when trying to pin-point exect pitches. One possible disadvantage is that the ten-turn format takes slightly longer to offset manually to e different reg- ister. While on the subject of pitch precision 2 word about oscillator drift may be in order. In the “Ice Age” of electronic music systems (the early 1960s) there ‘5. Offset is a term refering to the establishment of the initial f 2 particular Fate of reference ‘parameter. This will be cov- fered in deeall im chapter 5. 13was a great deal of criticism of voltage controlled os- cillators for not holding a set pitch accurately for ex- tended periods of time. Formerly musicians either used 2 different oscillator or composed around the problem. In newer instruments drift has been minimized to an almost insignificant degree. It is my own opinion that electronic: musical instruments are still musical instru- ments, and any musical instrument requires tuning. One should not really expect an oscillator to hold its reference pitch in a performing situation to any greater degree than one would expect a violin, late. or tram- pet to hold a pitch in the same situation (although many of the available electronic instruments do have great stability and will stay exactly in tune for long periods of time). At the same time we realize that with any instrument. including electronic ones, pitch will not be stable until it is sufficiently warmed up. Leave ample time for the instrument to be tumed on and left in the environment in which it will be used. Some studios make a practice of leaving the instruments on continually. Unless there is tube circuitry in the in- strument this should not cause any problems with over-heating and the practice will add to the stability of the system. Basic Waveshape and Spectra Before exploring some specific oscillator formats it is necessary to Jeam the relationship between various patterns of generated audio voltages or waveshapes, sound spectrum, and the aural experience. Mamufac- turers have come to some agreement as to the basic “orchestra” of commercial electronic music instruments, and the-waveshapes to be discussed here are common to most instruments. If we think of waveshape as the graphic representation of the rise and fall of voltage from zero to a masimuin positive and/or negative and back again, it is possible to identify basic electronic- ally generated sounds by their shape. There is a direct relationship between the visual and sonic quality of a sound. Nearly all sounds eshibit a spectrum or 2 collection of many individual frequencies which com- bine to make a single aural event. In striking a gong it is possible, by careful listening, to isolate aurally the multitude of pitches which make up the total com- plex sound of this instrument. Most instruments pro- duce sounds which consist of many combined fre- quencies called overtones or partials. Gongs, bells and other percussive instruments dis- play unusual spectra where apparently there is no con- sistent relationship between the frequency components or partials. Most strings. brasses, reeds, and certain electronically. generated sounds, however, display predictable spectrum based on a rather simple coo- cept. Any vibrating source is capable of the excitation of generation of additional frequencies. The standard “4 classroom method of observing this is to silently de- press middle C* on the piano and strike C® an octave elow. One will hear middle C ring out as it has been forced into vibration by the lower octave C*. This type of forced vibrtion is called sympathetic vibration and demonstrates that one vibrating system has the potential of generating other vibrations. Tt is an observed fact that nature accomplishes things in the easiest and most efficient manner. The easiest way we have of expressing simple relation- ships is in terms of integers, or whole mumbers (1, 2,3, etc) The most common type of excitation in vibrating systems also occurs at whole-number inter- vals. A low C: of 654 Hz is capable of generating ‘twice that frequency, 130.8 Hz (the octave C; or 2 x 654 Hz), the multiplier “2” being the firs integer above 1 (the unison). The same vibration of 654 Hz is also capable of generating 1982 Hz (3 x 65.4), ‘which is the pitch G®, a perfect 12th above the original frequency. The process of multiplying a basic fre- quency by whole numbers could continue ad infini- tum. Most sounds made by what we call “traditional” musical instruments exhibit a spectrum containing fre- quencies related by whole numbers. This type of spec- ‘rum is called the harmonic overtone series and is llus- trated in part in figure 35. The lowest frequency of this spectrum is called the fundamental and is the gen- erator of the series—each overtone being an integral or whole number multiples of the fundamental. The over- tones are, in this case, harmonic, referring to the fact that each is an integral multiple of 2 single fanda- mental frequency, forming a consistent system of re- lationships one with another. The term “harmonic,” for my purposes, does not refer to the obvious pro- gression of whole numbers, but rather to the fact that there is 2 consistent relationship between the num- ‘bers, irrespective of the nature of the consistency or simplicity. They are “in harmony” with each other due to a consistent relationship. By this definition there can be other harmonic relationships built on systems other than whole number multiples; this idea vill be further explored in chapter 8 (see page 114). In the meantime we will retain the tradition of “har- monic” referring to whole number relationships. There is often confusion as to the precise defin- tion of the terms overtone, portial and harmonic. For the purposes of this text overtone will refer to any frequency component in a spectrum above a given fundamental. If the overtone bears an integral rela- tionship to the fundamental it will be called a har- ‘monic overtone, or just a harmonic. There is also occa- sional confusion about the term harmonic, and the con- fusion arises from differentiation between harmonic and harmonic number. If one uses or assumes the term harmonic number, the series begins with the fun- damental as harmonic #1, the octave then being har- monic #2 and so on. On the other hand, the term IstFigure 3.5. Hamonic series harmonic usually refers to the first integral multiple above the fundamental or the octave. The 2nd har- ‘monic is the second integral multiple (the perfect 12th) above the fundamental, etc. This text will use the former practice of harmonic numbers, the fundamental being 1, the octave being harmonic #2, the 12th being harmonic #3, and so on. This is perhaps the more logical, since the harmonic number then agrees with its integral multiple (harmonic #1 is fundamental times 1; harmonic #2 is the fundamental x 2 har- monic 343 is the findamental x 3, etc.). The term partial refers to specific spectral content irrespective of harmonic or non-harmonic relationships. And as will be seen, certain sounds contain only selected over- tones. A clarinet, for example, has a spectrum contain- ing only odd-numbered harmonics (the fundamental, ‘rd, 5th, 7th, etc.). In this case, the fundamental is the first partial, and due to the lack of a scoend harmonic, the Srd harmonic is actually the second partial, the ‘5th is the third partial and so on. This garble of terms is clarified by figure 36. To 2 certain extent the overtone content of a sound, bbe it harmonic or non-harmonic, provides us with its aural signature or timbre. There are other factors which contribute significantly to timbral perception. ‘The loudness of 2 sound, the mamner in which the sound is activated and the manner in which it stops vibrating, the relative amplitude of the overtones, etc., all play an important part in timbral recognition and these factors will be explained as their control tech- niques are introduced. Perceptually, or pethaps even aesthetically the most ‘oncomplex type of sound is the sine wave. This par- ticular waveshape contains no overtones. The closest ‘sound to a pure sine wave in a symphony orchestra is that of a flute. As is shown in figure 37 the voltage is in a particular state of motion. Starting at 2er0, it ‘gradually increases to maximum positive, then decays through zero to maximum negative, then retums to the ‘A sine wave exhibits this same pattem indepen- dently of frequency or amplitude. Figure 37B shows the same frequency as figure 37A but with an in- creased amplitude. The device that electronically pro- duces sine waves is referred to asa sine-wave oscillator. Sine waves, like any other waveshapes, can exist in ‘any frequency range. Due to distortions caused by various components of the oscillator and/or distortions in the reproduction equipment, however, a precise sine AANA . U wave is very difficult to generate, and the composer usually has to settle for something less. It is the ob- servation of this writer that 2 pure sine wave used as an audio signal is not an especially monumental mu- sical sound and a bit of harmonic distortion (very small traces of additional harmonic content) may add to the incipient musicallity. Various manufacturers publish the amount of harmonic distortion of their sine wave oscillators along with other instrument spe- ‘fications. An imperfect wavesbape used as an audio signal can be accepted, and may often prove to be ‘more musically interesting than the textbook model Figure 38 is an oscilloscopic (graphic) represen tation of a sauctooth or ramp wave. In contrast to the pure sine wave, a sawtooth wave contains all har- ‘monic overtones of the fundamental frequency. These harmonic overtones have relative amplitades that de- crease exponentially as they exist higher up in the harmonic series. A sawtooth oscillator will produce this basic waveshape in any frequency range. Note that figure 38A is symmetrically inverse of 38B and is often referred to as an inverted sawtooth. Both are legitimate sawtooth waves and, in spite of the reverse ‘position of the leading edge, will sound the same. The sawtooth wave is very bright and piercing, somewhat like the sound of an oboe or violin. an Figure 2.8, Sewtooth waves. n Figure 3.7. Sine waves. 16A third basic waveshape is the triangle or delta wave (fig. 3.9). This waveshape consists of 2 funda- mental frequency and all of the odd-number har- monics, with amplitudes falling off in ratios of 1/9, 1/25, 1/49, ete. By using @ classic studio technique of “additive synthesis,” it is possible to construct @ tri angle wave (or any other waveshape) by using @ specific collection of different sine waves. Starting ‘with a fundamental of C (65.4 Hz) with a hypothetical amplitude of x, 2 second sine wave tuned to 65.4 Hz times 3 (for the second overtone) is added (196.2 Hz) with an ampiltude of 1/9 X. Then a third sine wave which is the fifth multiple of the fundamental (327 Hz) is added with an amplitude of 1/25 X. The com- poser continues this process until all of the necessary harmonics with the correct amplitudes are present. If the amplitudes of the harmonics are thought of in ratios, it is easy to understand how the perceived har- monic content of any given wave is dependent on the amplitude of the fundamental. ‘Perhaps the waveshape that is most commonly used by the composer is the pulse or rectangular wave, AS shown in figure 310, the positive and negative volt- ages of a pulse wave are never in a transient state. ‘They are instantaneously positive, then instantane- ously negative, whereas the sine, sawtooth and triangle ‘waves all exhibit various types of gradual rise and fall between positive and /or negative states. Ifa pulse gen- erator is programmed to oscillate anywhere below 7 to 10 Hz, the speaker cone can be heard snapping back and forth. Figure 3.10 shows a particular type of pulse wave known as a square wave. A square wave is related to a triangle wave in that it also contains odd-numbered harmonics, but with quite different amplitude relation ships. The amplitude relationships of the harmonics of a square wave are 1/3, 1/5, 1/7, 1/9, ete. A clarinet in the chalumeau register produces a sound that is very close to that of a square wave. Figure 3.11 ius trates in standard musical notation the harmonic con- tent of the four basic waveshapes. The relative amp- Iitudes are indicated by the size of the note. Note that there are no even numbered harmonics in any of the symmetrical waveforms. Due to the many uses of the square wave, it is thought of as a basic waveshape, even though it is a variety of a pulse wave. There are many other types ‘of pulse waves, and they are defined by what is known as their “duty cycle.” The duty cycle of a pulse wave is the positive or “on” portion of the entire cycle. Fig- ure 3.12 shows the duty cycle of two different pulse The duty cycle determines the harmonic content of the waveshape. In the case of the square wave, the uty cycle is one-half of the total wave, or a ratio of 1:2 (as in fig. 312A). Expressed as a fraction, the 16 See —————— mi Figure 2.11. The four basic waveshapes and ther harmonic ‘conten (ap tothe Sth matte) sip tt oa Ty Figure 3.12. Pulsewave duty cycles uty cycle of a square wave is one-half of the total ‘wave, and it is the denominator of this fraction that tells us its harmonic content. The denominator “2” in- icates that every second harmonic is absent from its harmonic overtone series, confirming the earlier state- ment that the square wave consists only of a funda- mental and the odd-numbered harmonics. A pulse wave with a duty cycle of 1:3 (as in fig. 312B) con- tains the fundamental and the first, second, fourth, fifth, seventh, eighth, tenth, etc, harmonics. In other words, the denominator is an indicator of what order harmonies are absent from the spectrum. Oscillators designed for use in electronic music systems may have controls or pots marked pulse width, duty cycle, sym- metry or waveform adjustment. Palse width and duty cycle refer specifically to pulse waves and may vary the “on” portion of the wave from 10% to 90% or more of the total eycle. ‘Waveform adjustment or symmetry are variations ‘which are possible with any waveshape. For example the Synthi VCS provides three oscillators with aKh Ww om NMI WA 3AM a V variety of waveshapes having symmetry control on cach, Figure 3.13 illustrates some of these waveshapes as they appear with the symmetry control in different positions. Each of these variations produces 2 slightly different timbre, except in the case of symmetrical in- version (as was the case with the inverted sawtooth wave in figure 38) Note that, with the sine wave, there is a symmetry position which places all of its voltage above the 0 volt reference in the shape of two crests. Waveshapers and wave multipliers can be used to transform triangle waves into other waveforms. The resulting waveform may be another classic waveform such as a sign or squarewave, or it may be a com- pletely new class of signal impossible to describe in terms of standard waveshapes. Such modules are really signal processing devices and will be discussed later in the text. Additive Synthesis ‘Composers such as Ravel and Hindemith, and more contemporary composers such as Ligeti and Kagel, have been very concerned with the mixing of various instruments to produce new orchestral timbres. One of the classic examples of this additive approach to orchestration is in Ravels Bolero. Beginning in mea- sure 149, Ravel combines a horn, celeste, and two Piccolos to produce a sound unlike any of the indi- ‘vidual instruments used. Examination of the score dis- closes that Ravel's apparent ti-tonality is actually 2 reinforcement of the harmonic series of each pitch in the melody. The horn plays the fundamental while the celeste plays the first and third harmonics and the piccolos provide the second and fourth harmonies. (See figure 3.14.) A more recent approach to this type ‘of composition is Maricio Kagel’s Music for Renais- ‘sance Instruments (Deutsche Grammophone Records, 10, 137 006), in which the composer is concerned with constructing various types of non-harmonic sounds. This same basic method is applied to electronic mu- sic in classical techniques of timbre construction. Sup- pose, for example, that a composer wished to synthe- size the spectrum of a square wave. Since a square wave's spectrum is virtually infinite, one hypothetic- ally would have to have an infinite number of sine wave oscillators, one for each frequency present in the square wave's spectrum. This is impractical. for the obvious reason that there is a limit on available oscillators. It is really an academic problem, since be- yond the third octave above the fundamental manual tuning becomes quite difficult. For sake of illustra- tion, we will limit this construction to the first five partials present in the square wave's spectrum. After arbitrarily deciding on a fundamental of low Fs (1746 Hz), a sine wave, one would then tune each remain- ing sine-wave oscillator to the other odd-numbered barmonies: C, A*, Ep’, G®. Since all of these frequency components must be perceived as a single sonic ob- ject, they must be combined into a single signal. This is accomplished by means of an audio mixer. Mixers and associated techniques are covered in detail in chapter 11. Here, it will suffice to state that their func- tion is to combine two or more frequencies or signals into a single signal in such a manner that minimal dis- tortion of the original signals occurs. A mixer also usually provides means for controlling the amplitude of each input signal by means of 2 volume control, pot, or attenuator. Therefore, the amplitude specifi- cations for the harmonics of 2 square wave can be accommodated (see page 18). Figure 3.15A illustrates the physical set-up or “patch” for this process; figure 15B shows the waveshape transformation from sine toward square as each frequency component is added in its correct proportion. This process is generally referred to as Additive Synthesis or Fourier Synthesis. Briefly, the Fourier Synthesis states that any sonic object can be created (more specifically re-created) by subdividing its in- stantaneous spectrum into individual sine wave fre- quency components, then combining the correct num- ber of sine waves with the proper amplitade relation- ships to recreate the original sound. Theoretically any ‘waveshape may be recreated through this technique. ‘Again, this seems to be a bit academic, since the basic waveforms are already available from various oscil- lators and additive waveform synthesis techniques may be put to more productive use in creating some new waveforms. From this viewpoint all that can be of- fered in terms of suggested technique is to be aware of and experiment with various amplitude relation- ships among the partials. If one wishes to experiment with harmonic frequency relationships read ahead in this chapter about sync techniques (page 20). Although additive techniques are powerful tools in the electronic media, precise Fourier Synthesis tech- niques are for the most part impractical on commer- 6, Refer to page 20 for syne techniques related to this type of tuning. 7cial instruments due to the numbered variables volved and the amount of instrumentation required Some writers have pointed out that the discovery of the Fourier series qualifies the meaning of the word “frequency” in reference to a complex wave shape. ‘When speaking of the frequency of a sawtooth wave being 440 Hz we are really referring only to the fun- damental frequency. Each of the harmonic overtones, if viewed as sine wave components, also has its own frequency. However, the sawtooth wave or any other harmonic waveform is perceived as a single identity Integrally related overtones reinforce the perceived fundamental, and it can be assumed that the term “ quency” refers to the perceived fundamental.* Oscillator Formats Although industry has reached some agreement as to what basic waveshapes are to be made available, the actual format of the oscillator varies a great deal. The simplest is a single oscillator with one available wave- shape. Most oscillators make use of waveshaping cir- ceuits which enable an initial waveshape (usually 2 sawtooth) to be converted to several other waveshapes simultaneously. Figure 316 shows the front panel of the Eu 2200 oscillator. The four basic waveshape are independently available unattenuated (at full ampli tude) at the outputs marked “Full Level Outputs.” At the same time each of the four waveshapes may be mixed in any amplitude proportion by means of the front panel pots, and the mixed output is available as indicated. The Buchla Series 258 oscillator takes a different approach (figure 3.17). First note that the chassis or module contains two oscillators, each with indepen- dent controls. Each oscillator has three outputs con- nected in parallel, meaning that the same signal ap- ‘pears at each of the three outputs. The top oscillator {in this dual package provides two basic waveshapes— sine and sawtooth. The wavesbape pot essentially es- tablishes a mix of these two waveshapes: a certain amount of sinewave oscillation mixed with a certain amount of sawtooth wave oscillation. This is not to be confused with the gradual introduction of the vari- ‘ous harmonics as in a Fourier build-up. It is merely 2 ‘mix that selects proportionately between two wave- shapes. At the 7:00 position (far left) the output is 3 @ Figure 3.17. Buchia Series 258 Oseltater a sinewave, and at the 5:00 position the output is a sawtooth wave." The waveshape pot may then be set anywhere between these two extremes. The bottom oscillator in this package has the same format except that the two waveshapes are sine and square. ‘The Buchle 208 oscillator (refer to fig, 34B) has three waveshaping controls. The rotary waveshape pot Provides for transitions or mixes between sine and either pulse’, square, or triangle as selected by the swaveshape switch. The linear timbre pot provides fur- ther waveshaping possibilities, generating a continum 9. It might be noted that this is not an academic sawtooth ‘waveshape, rather what Mr. Buchla calls an “auzmented saw tooth.” The harmonic coefficients are similar to those of a saw tooth, but of greater intensity; similarly for Buchle’s square wave, Hence, oscloscopic display of these waveforms will not show a true sawtooth oF square wave image. 10. Sometimes called a spike, this waveshape has very strong armonis and may be hurt ofa a pulse wave with vey arrow duty cycle. 19of harmonic intensity by gradually introducing and em- phasizing lower ordered harmonics as the pot is raised. This provides a rich set of timbres not associated with normal classic waveshapes. “Synchronization” The construction of timbres and orchestration of voices through additive techniques often requires very pre- cise tuning. Function generators provide simultaneous waveshapes which are referenced to the same fre- ‘quency. Other additive techniques often require that two or more oscillators maintain a perfect harmonic relationship and retain that relationship under a va- riety of performance conditions. Synchronization of two or more oscillators means thet they will maintain ‘2 consistent and pre-determined intervallic relation- ship under most control situations (the interval, of course, may be a unison). Although the most inter- ‘sting musical applications of synchronization, or sunc, cannot be explored until certain control processes are put into practice. the basic concept must be under- stood here as it has a direct affect on initial experi- ments with an oscillator. Let us imagine a situation in which two oscillators are producing frequencies of 4826 He SSE and 440 He BSE — a quartertone out-of-tune, By synching the first oscillator to the sec- ‘ond the lower frequency’ will be pulled up to exactly match the frequency of the first oscillator. This syn- chronous relationship may be established in several ways, according to the design of the oscillator. The populer ARP Odyssey provides the performer with ‘two audio oscillators, the second of which has a “sync switch” which allows it to be brought into perfect ‘tune’ with the first. In this case the first oscillator is called the master and the second is called the slave, The master provides the model frequency to which the slave must adjust. Note that in most cases the master oscillator is a higher frequency than the slave and the slave is pulled up to match the frequency of the master. In this case the sync function is “hard- wired” between the two oscillators and n0 external patching is required. The Eu 2200 oscillator has access to two synch con- nections which allows synchronization of any oscil- lators in the system. One set of oscillators comprising ‘a voice may be synched together on one bus and an- other voice (set of oscillators) may be synched on the second bus. The following excerpt from the Eu Oren- er's Manual explains this process and also provides some other relevant information on synchronization processes ‘The syne switch allows the VCO (escilater) to be con- nected to one of the two syne busses running throughout the machine. Oscillators which are connected to the seme ‘sync bus and all somewhat less than 2 semitone apart in pitch will be pulled, or ‘synched! to precisely the same frequency (equal to the highest frequency among them). In addition, an oscillator within semitone of any har- ‘monic of another oscillator will be pulled up if it fre- ‘quency is lower than the harmonic, of will pull up the ‘other if higher than the harmonic. Synching is essential when slow beating between oscillators is unwanted. If this so-called “choral etfec’ is desired the sync can be turned off, Synehing Is accomplished by the use of sync pulses; when the sawtooth of an oscillator falls, @ sync pulse is issued on any engaged sync bus. Thus, with all oscillators beth issuing and observing each other's sync ules, all are pulled to the rate of the fastest, This is known as ‘sof’ syne, since the slower oscillator will only discharge its sawtooth if it was already close to doing 0 anyway. ‘Hard’ sync, used for example in the ARP ‘Ocyssey, causes the sawiooth of a slave oscillator to discharge whenever the master oscillator does. This is ‘equivalent to soft sync when the slave is lower in fre- Quency than the master, but resuits in harmonic side- bands!? and consequent timbral changes when the slave is higher than the master. This may be either interest- jing or annoying. in any case, hard sync is incapable of synching to harmonics, as can be seen in the drawing below, which describes the operation of both types of ‘syne. 12. A sideband is an additional frequency produced by cer- ‘ip editions of a sgn The term i tal asoited with various types of (see page 112), 13. From the manuscript version of Eu Owner's Manual by Ken Provost, pages unnumbered. wal some good. If not, perhaps you might wish fo skip this sectio ‘unless you are just interested. Some specific performance tech niques dealing with sync are covered on page 119. 20It is not the purpose of this text to give a detailed description of every instrument. A glance at the front panel of any oscillator will usually reveal its format; and specific questions are usually answered in the own- e's manual or spec sheet. This brief exposure to o5- cillators and some common formats should provide the reader with enough information to approach any os- cillator with at least some meaningful questions, and may also serve as a guide for some applications of whatever format oscillator is available 2t the time. White Sound The last electronic sound (as differentiated from many other sounds available through transducers such as microphones, prerecorded tapes and discs, ete.) to be considered is white sound, or white noise (also some- ‘times referred to as Gaussian or thermal noise). Per- hhaps the most descriptive term is “white sound.” Ana- ogous to a color wheel that produces the color white as the wheel is rapidly rotated, white sound is 2 mix- ture of all the audible frequencies at random instan- ‘taneous amplitudes. Therefore, the term “white” is preferred over “Gaussian’—and the present writer pre- fers the use of “sound” because of the negative impli- cations of the term “noise.” White sound is heard as 1 hiss or as the sound of a jet engine. White sound is defined as having equal energy per unit frequency. ‘This means that there is the same amount of energy ‘between 500 Hz and 50] Hz as there is between 1500 Hz and 1501 Hz. Equal energy per unit frequency ‘means that the noise is spectrally flat. Another type ‘of sound common to most electronic music instruments is “pink sound.” Just as with white sound, pink sound contains ‘all the frequencies of the audio spectrum, ‘but its energy distribution or amplitude curve is dif- ferent. Pink sound contains equal energy per octave and is often represented as-}-and referred to as be- ing musically flat. It may be obvious to the reader that the term ‘pink sound’ is a logical color analogy as it refers to the lower end of the light spectrum; in terms of perception pink sound simply has more bass or rumble (see figure 3.19). Taking a cue from these ‘olor analogies one could easily construct color repre- sentations for any number of noise bands. ‘For ex- ample, boosting the higher end of the spectrum may produce what one could call blue or azure sound, while 2 spectrum with a boosted mid-range might even be ‘green! These terms are by no means representative of spectral standards but may serve as convenient termi- nology for describing different types or bands of sound. Electronic instruments provide noise coloration in. various ways. The Synthi VCS-3 instruments control noise spectra by means of a pot marked “colou:” with Figure 3.19. white and pin sound the far left and right positions appropriately marked “low” and “high,” and the middle straight up position assuming to be spectrally flat white sound. The Buchla Source of Uncertainty Model 265, among other things, provides three independent noise outputs. The output labeled “HI” is what has been defined as white sound. ‘The middle unmarked output provides pink sound and the output labeled “LOW” is a type of noise which is the musical inverse of the frequency distribution found in white sound. As explained, white sound has ‘equal energy per unit frequency, while pink sound has ‘equal energy per octave. If the white sound spectrum is measured in octaves one will see that there is ac- tually an increase of 3 db per octave. Buchla’s noise resources establish musically flat pink sound as the nom with “HI” having a 3 db increase per octave and “LOW” having 2 3 db decrease per octave. Some manufacturers refer to white sound, pink sound. etc., simply as random signal generators (such as the Moog ‘903A Random Signal Generator). The term ‘random’ is used because the amplitudes of the various fre- quencies at any given instant are random and must ‘be measured over a certain time period in order to calculate an average measurement. In other cases the term ‘random noise’ refers to extremely low random frequencies (around 15 Hz and lower); it is not aud~ ble but rather is used as @ control for random musical events (see page 83). Thus far, only the basic types of sound sources and some typical formats have been discussed. It is essen- tial that the reader understand how sound can be thought of as AC voltage, and that these voltages can bbe shown graphically on an oscilloscope (as they are represented in the various figures used in this book).. Since the scope of this writing does not pémnit the discussion of every type of sound available to the com- ‘poser, we have given the four basic waveshapes—sine, ‘sawtooth, triangle, and pulse—and the concept of white sound, An understanding of these is all that is needed for a basic understanding of electronic music system ‘operation. 24 Basic Signal Processing: Amplifiers and Filters It is now time to retum to music as set of dynamics— not merely dynamics in terms of loudness, but rather a more general view of simultaneous variability of co-existing parameters. Most musical events involve fluctuation, on many different levels of frequency: amplitude, and timbre. Other parameters such as space, rhythm, etc. can really be conceived of in terms of these basic three parameters and will be discussed as various forms of control are introduced. Compasi- tion and performance with electronic instruments de- mand that one make decisions and enforce controls (which may range from totally determinant or com- pletely random) on these parameters. Failure to do s0, at least in the initial stages of leaming, can lead to frustrations, the most serious of which is the dread of every electronic music novice, silence! As one will iscover, the production and control of electronic sound is usually an evolutionary process. One module ‘generates the sound, one module imposes controls on frequency. another controls amplitude, still another controls spectral content, ete. Failure to understand any one of these generation and control processes in the chain, will ultimately result ina discrepancy be- ‘tween the original musical concept and the final rmu- sical result. The shaping of musical events involves various processes dealing initially with frequency, amplitude, and timbre. The realm of frequency variation and waveform should be basically understood by this point. Before embarking on 2 study of precise control pro- ‘cedures one should understand the basic processes of amplitude and timbral variation in terms of voltage and waveform variation. An electronic circuit operates only in terms of voltages (shapes and amplitudes). The musician must be able to think in these same terms to communicate his musical needs to his instrument. Conceiving of 2 musical event or structure is only the beginning. The musician working in the electronic medium then must translate his concept to a language understood by his instrument. This chapter deals with the parameters of loudness control through amplifiers and various aspects of timbre through filtering, Again the emphasis is placed on what each circuit is doing to the electronic signal 22 Amplification and Gain With the existence of a multimillion dollar home hi- fi/stereo industry the term amplifier and its concept is familiar, at least in part, to almost everyone. An amplifier, of course, amplifies! Technically this means that it is a device which increases some externally generated voltage. Earlier, it was explained that voltage levels, or the degree of excursion that a waveshape exhibits, determines the loudness of a signal. This de- agree of excursion, or amplitude. is a musical variable that is usually spoken of as loudness (loudness being a subjective measurement of amplitude). The mu- sician working in the electronic media deals with amplification on several diffezent levels. Sounds—more specifically, signals produced by electronic instru- ‘ments—are quite low in magnitude. The exact strength of an audio signal from an oscillator varies with each manufacturer. Systems on the market today have out- ut voltages anywhere from ] to 10 volts. These volt- ages carry with them a variable called impedance, which is a measurement of electrical resistance, mea- sured in units called ohms (1). It would be unneces- sairly confusing to attempt 2 detailed discussion of impedance at this point. For those interested, the out- ut impedance in most electronic instruments and studios is usually between 600 and Ik obms. If this is meaningless to you, either forget about it for now or ask a technician. The point is that these signal levels coming out of most electronic instruments are not strong enough to drive a speaker cone. Hooking an oscillator directly to a speaker will produce little, if any, sound—certainly not enough to be of any real musical use. Therefore, one must raise the voltage level, or waveshape excursion, to such a magnitude that it can efficiently drive a speaker cone. And this ‘must be done without changing the original frequency ‘or waveshape. The device used to accomplish this is 8 power amplifier. The power amp is not an integral art of the electronic instrument but certainly is a major factor when it comes to the final sounds. The use 1L, This information is usually given on the instrument's spe- ification sheet. Because of voltage variations it is not advis- able to interconnect instruments made by different companies ‘without first checking the voltage and impedance levels.‘of power amps will be discussed in the consideration of live performances (chapter 15); but, for the present, it will suffice to say that their function is to take a Tine-lecel signal (about 1.4 volts) from an electronic instrument and boost it to a level thit has enough energy to efficiently move a speaker cone (see figure 41) Voltage Controlled Amplifiers It is certainly redundant to say that loudness is a dy- zamic variable for the musician. However, the effec- tive use of loudness variations requires 2 bit more detziled understanding of what is involved with amp- ification in electronic instruments. The amplifier of most concer to the missician is called a coltage con- trolled amplifier (VCA), sometimes referred to as a ate. At this point we need not be concemed with the “voltage controlled” aspect of this module, and there- fore this discussion will be limited to manual or hands- on front panel control. The VCA. provides gain. This is a word which has been used before and refers to the amount of voltage an amp provides. If a signal goes in to a circuit at 2 peak-to-peak reading of 1 volt (see figure 4.1B) and comes out at a peak-to-peak reading of 2 volts the amp is said to have a gain of 2 IE the signal comes ont at 0 volts (silence) the gain is 0. The amount of gain. under manual control, de- pends on the setting of the gain pot, gain offset or volume control. When you raise the gain offset the ‘output voltage is increased and you hear a stronger signal from the speaker. In electronic musical instru- ‘ments it is common for the VCAs to have 2 gain of 1 or unity gain. This tends to be a bit confusing because with unity gain there is no voltage increase. The sig- nal goes in. and if the gain offset is at maximum, the output signal has the same magnitude as the input signal. An amplifier with unity gain is really an at- tenuator, It can produce variations in gain or loud- ness not exceeding the strength of the original sig- nal before it was patched into the amp. The reason for this is explained by the basic design philosophy of electronic instruments. Each module, oscillator. amp- plifier, etc., is a parametric building block. In some situations certain building blocks may not be neces- sary for the generation of a musical event. so that the modules must be designed, in such a way that they can function with or without any other module. For example. there may be a situation where loudness is not 2 variable and the signal is needed at full gain directly from the oscillator. Therefore the amp may be by-passed, (see figure 4.2, oscillator A). This “voice” may be joined by another voice that does re- quire Ioudness variations (figure 42 oscillator B). Oscillator A has a constant loudness, while the loud- ness of oscillator B is controlled by a VCA. If the amp hhas more than unity gain, great care has to be taken to insure that it will not overpower oscillator A which hhas no gain increase. But since the amp has only unity gain, at full output (the gain offset all the way up) it will perfectly match the amplitude of oscillator A. Sometimes gain is specified in decibels. Unity gain is the same as a gain of O db. If en amp is used for setting final manual levels which do not require continuous variation, 2 mixer might perform the same function with a little more versatility. As explained previously, a mixer allows the combination of many signals down to a single line with- out any distortion of the original sounds. Adjusting the balances on a mixer is the same as adjusting the gain ‘on an amplifier or set of amplifiers. It is not unusual for mixers (and amps) to have gain, a common figure being 6 db or a gain of 2 On a mixer the amount of gain is initially established by the input pots. These pots are attenuators which determine the incoming strength of the various signals. Figure 43 illustrates @ mixing situation with two different input settings. Figure 4.2. Gain variationFigure 4.3. Goin contol in mizing Figure 4.3A shows two sine waves two octaves apart mixed down to a single signal with equal gain for both signals. Figure 438 illustrates a similar situation but the lower octave C* is attenuated. The final mix has the same frequency components but the lower octave is softer. ‘Amps used in electronic music instruments some times have mixing capabilities built into the circuits Such is the case with the Moog 902. This module, con- sidering only manual control capabilities, will accept ‘two incoming signals and provide an equal mix of both at the output. The gain of the mix is 0 with the fixed control voltage pot at 0 and unity gain with the same ‘pot at 10. There are also two outputs marked “+” and “—". These two outputs are 180° out-of-phase with each other. The Eu 2000 amplifier is of similar design except that it has three inputs; “full level,” *-" and “(see figure 44). All three inputs are mixed in- temally before the final amplification stage, which, in this case, has a gain of 2 (6db). The inputs marked 2. “76 is an electronic symbol for a variable resistor, or pot, ‘which & used to affect the attenuation. Figure 4.4. Es 2000 y permiasion) “+7 and “~" have input attenuators to allow various ix levels to be established before amplification. The *— input is an inverting input, as opposed to the Moog (902 inverting output. The real applications of these in- verting features will be described when dealing with voltage control (see page 37). Note that in figure 44 the amp has a switch marked near and exponential. In this case linear means that ‘the gain increases in exact proportion to the pot setting. Exponential indicates that the gain increases exponen- tially to the pot position; the higher pot positions give ‘greater gain variances. Expressed in another way, lin- ‘ear means equal changes in output voltage in response to the manual control (electrically smooth), and ex- ‘ponential means equal changes in output decibels in response to the control (perceptually smooth). Lin- ‘ear control provides a consistent response throughout the range of the amp. Exponential control means that the amp is very sensitive to gain increases at low pot settings, and the sensitivity decreases as the pot set- ting is higher. Therefore, if one needs very delicate balances at soft levels, he has better control in the ex- ponential mode. On some amps the linear/exponential switch will make a difference where the 0 db gain or silence occurs, in accordance with the pot position. JE no signal is passed through the amp in exponential ‘mode with the pot a certain setting, switching to linear mode will result in a slight signal gain at the output. ‘A comparison of exponential and linear response is, given in figure 45, The exponential curve is really a decibel response, reinforcing the statement that lin- ear control results in equal change in voltage, and ex- ponential control produces equal changes in decibels. Some instruments with linear/exponential options may attach this function only to incoming external control sources, and the manual control may be only linear. If there is any question, refer to the instrument spec- ification sheet or instruction manual. If your instra- ment does not have linear/exponential options don't feel that you have an inferior device. Many of these fedesign decisions are made to increase the compati- bility with other modules in the system and are not necessarily an indication of the instrument's quality or potential usefulness. If your studio or instrument facility is equipped with Buchla 100 Series instruments you wall find that the 110 Dual Voltage Controlled Gate (another term often used for VCAs) does not provide for manually applied front panel gain or offset. The initial ampli- fication is provided by external control voltages and the output gain is then attenuated to the desired level by the front panel pot. If this is the case the exer cises suggested in this chapter can be done using a Buchla 106 mixer (which has @ gain of 2). The Buchla 282 Quad Gates have the manual offset control (see page 28). Pre-Amplifiers Another type of amplifier of concem to the musician is a pre-amplifier or pre-amp. Many audio signals, usually those generated by mechanical means such as ‘2 phonograph or a microphone, produce signals too weak even for the power amp. These signals are so small that they are measured in millivolts. If one plugged 2 mike directly into 2 power amp the out- ut signal would be so weak that it would not be useful. Power amps are designed to amplify what is referred to as a high level or line level signal. Such signals are on the order of 1 volt at 600 2. To get a mike's signal up to that level an intermediate pre- amplifier is used. The low level signal is plugged into the pre-amp, the pre-amp boosts the low level signal up to line level, and at this point the electronic signal ‘can be treated as if it were the same as an oscillator signal subject to any kind of amplitude shaping by means of the VCA. The pre-amp will usually have an impedance or gain control on the front panel. Vari- ous types of mikes have different strengths of sig- nals referenced in terms of impedance. A low im- pedance, or low Z, mike has a rating of between 50 to 250 O. A high Z mike will have a typical rating of 25k 9. Check the impedance specs for your available mikes and see that the impedance switch on the pre- amp is in the correct position. If the sound is muddy or distorted this is an indication that the impedance match is not correct. Some preamps may have gain ‘switches. Such is the case with the Eu 2420 Dual Pre- Amp (see figure 46). The gain positions for this in- strament are marked 20 db, 40 db, and 60 db, and provide a gain of 10, 100, and 1000 respectively. After the signal has been boosted by one of these factors, the final signal is then attenuated by a level control. Readers familiar with rock amplifiers will know that they can plug a mike or guitar (which is also a low level signal) directly into the amp, apparently without the use of a preamp. This type of amp is Jmown as an integrated amplifier, and the pre-amp is built into the general circuit and wired directly to the power amp stage. The volume and tone controls are part of the pre-amp stage, and the power amp stage then amplifies the final signal. Most keyboard oriented electronic instruments such as the popular ARP Odyssey and Mini-Moog have two different out- put signals. The high level output or line level output is ready to be patched into a power amplifier. If this high level signal was fed to an integrated amp one would then be attempting to pre-amplify a signal that is already pre-amplified. In this case the low level ‘output of the instrument should be used, since it can be taken to the integrated amp and used with other low level signals such as electric guitars, mikes, etc. Much more will be said about pre-amps and mikes; the information given here is intended only to avoid some possible distorted sounds during the initial stages of exploring the instrument. Filtering: Subtractive Synthesis and Basic Filtering Concepts Eaulier in this chapter mention was made of additive synthesis whereby the musician builds up waveshapes and sound complexes by methods of mixing together ‘of less complex components. Another method which takes the opposite approach is called subtractice syn- ‘thesis. As the name implies, this method involves using, complex sound as the initial material.and removing those frequency components that are not desired. Just as additive synthesis can, at least hypothetically, build up any complex from its individual sine wave com- ponents, subtractive synthesis can, again hypotheti- cally, remove any number and combination of compo- rents from a complex structure. For example, it is possible to begin with a sawtooth wave and remove all of the overtones, leaving only the fundamental, 2 sine wave (figure 474). Similarly removing all of the even numbered harmonics would produce a rea- sonable square wave (figure 47B). Removing only 25SE rowear. ‘Stbiractive waveshaping the fundamental, leaving all of the overtones, would produce an unusual waveshape with transposed har- monics or “half harmonics,” thet is, all of the har- monic overtones are present but one octave lower than usual (figure 47C). If this is confusing, think of the Ist harmonic overtone, which is usually the ‘octave. being transposed down to the unison, the 12th is now a fifth above the new fundamental, and so on. This is an academic example and not as sonically strik- ing as it might seem. Low Pass Filtering The module that accomplishes these kinds of trans- formations is called a filter. Just as an amplifier pro- vides control in loudness in the amplitude domain, 30 also the filter provides for control in the spectral domain. In fact one may hear a filter referred to as a frequency selectice attenuator or frequency selectice amplifier. In other words. a filter is a device that pro- ‘vides gain for only selected portions of 2 composite signal or spectrum. The portion of the signal which is affected is determined by the function of the filter. A low pass filter, perhaps the most common of the filters built for electronic music instrumentation, vill amplify or “pass” only those frequencies below a spe- cified frequency setting or cut-off (f.).* Referring to figure 3.11 (page 16) the original spectrum is a saw- tooth wave with @ fundamental frequency of 654 Hz, © If this signal is patched into a low pass filter with a cut-off of somewhere between 654 Hz and 1308 Hz (the octave harmonic) the filter would suppress those frequencies above the cut-off, ideally leaving only the fundamental frequency. If the cut-off fre- quency were raised to slightly above 1308 Hz, the ‘3. Sometimes called the ‘break” frequency. filter output would be a structure with only the fun- damental and 2nd harmonic. Starting at the other end of the spectrum may’ be more in keeping with the idea of subtractive methods. If the 4. (cut-off frequency) of a low pass filter was set at maximum any wave- shape would theoretically pass unaltered. This is usu- ally not the case, since @ sawtooth wave in theory bas an infinity of harmonic overtones, and the cut-off of a filter should go infinitely high. The reality is that these two “infinities” may not be the same. A filter hhas a limit to its high end, and an oscillator wall not produce an unlimited number of overtones. This is ‘easy to test. Just listen to a signal straight from the cscillator and compare it with the same signal patched into a low pass filter with the fe at maximum, In addition to some slight high frequency loss there will ‘probably be some coloration. Coloration refers to slight attenuation or amplification of various parts of the spectrum caused by the circuit. Keeping this patch, gradually lower the {, and the upper harmonics will ‘be sequentially removed. Figure 48 illustrates this process. Also note that as f is lowered the edge of the waveshape begins to round off, eventually result- ing in a smooth sine wave. The lesson to be Jeamed here is that waveshapes with more crests and edges usually indicate the presence of more frequency com- ponents. Details about filter response such as roll-off and Q are best deferred to the chapter dealing specifi cally with filtering, but brief mention will be made here to introduce this concept. The below figure is a useful example, but things usually don't happen that efficiently. A filter cannot abruptly cut off at the spec- ified {.. The f. is a selected spot where the filter has ‘caused 2 3 db attenuation in the spectrum. The attenu- ation continues from that point at a given rate, slope ‘or roll-off. A common figure for roll-off in a low pass filter is 24 db per octave. This means that if our 645 Hz sawtooth wave was patched into 2 24 db/octave filter with 2 f of 196.2 Hz (654 x 3), the output would be a spectrum with the 3rd harmonic down 3 db, with the loudness decreasing as one goes higher in the overtone series. The octave above this cut-off,the fifth harmonic is down 24 db and the next octave, the 12 harmonic, would be down 48 db (see figure 49). Filter designs will vary with the manufacturer and such rolloff figures are usually given on the spec- ifications sheet. These figures will usually be mul- tiples of 6 (12, 18, 24, ete.). Q is a circuit modification that alters the effect of a filter. The mechanics, applications, and tricks made available with Q are covered in chapter 9. Here Q can be explained as an additional amplification proc- cess within the filter. If a low pass filter is set with cut-off of Ik Hz, any signal or spectral component just below cut-off frequency will be amplified as the Q is increased. Most low pass filters, such as the Moog 904A pictured in figure 4.10A have a Q pot on the front panel (in this case called regeneration). With the Q set high, manually sweeping through @ har- monically rich waveform will make the filtering pro- cess more obvious. Figure 4.10B illustrates what hap- pens to the cut-off frequency as the Q is raised. No- tice that as Q is raised the part of the spectrum just below f, becomes proportionately louder and stronger. At the same time the rest of the spectrum passed through the filter becomes slightly attenuated* The overall effect is the f, becomes more and more promi- nent. Much more will be said about Q in chapter 9, bbut for the present all we need to know is that Q emphasizes f, and has a numerical value related to the increase in gain at the cut-off: the higher the Q, the higher the gain. Gain can be increased to the point at which the filter begins to distort or “break up” the ‘cut-off frequency. Q can also be increased to the point where the filter will begin to “how!” or oscillate. This type of oscillation is a sine wave and can be useful for certain techniques (covered in chapter 9). This application is only mentioned here because itis an in- 4. This band-pass attenuation is not always the case and de- pends on individual cirait design. ro actos. Iitog 2048 Low Pass Fier Figure 4.108. 0 comparisons i low pass tering strument variable and can result in some phantom pitches and distortions if one is not aware of the proc- ss. Other names for Q in addition to “regeneration” are “resonance,” “response,” “emphasis” and “feed- back” 27Other Filter Functions Another common filter format is the High Pass Filter. As the name implies, it accomplishes the inverse of the low pass filter, attenuating all of the frequencies below f, and passing all of the spectral components above f. Other filter formats include band pass, band reject or notch, and the fixed filter banks, which is collection of band pass filters mixed to a single out- put. The functions of these filters are indicated by their descriptive names. Universal filters are single modules which combine the four basic filter functions, —high pass, low pass, band pass, and band reject,— into a single instrument. Another filter variation which may be available to the musician is what ARP In- struments call the Filtamp. It is a very common prac- tice simultaneously to control the amplitude and timbre of 2 signal by patching it through an ampli fier and then through a filter (see the “basic patch” ‘on page 29). ‘The Buchla 292 Quad Low: Pass Gate (figure 4.11) hhas similar functions. This module consists of four in- dependent gates, each of which has three possible ‘operational modes. The mode switch in the upper po- sition turns the gate into 2 low pass filter; the lower Position tums the gate into an amplifier, the middle Position, “combination,” combines the amp and low pass functions to provide “spectral gating”—a process similar to simultaneous spectral and amplitude con- trol. In this mode, as the amplitude is lowered there is a simultaneous loss of high frequencies. Both of these functions occur at the same time as they share the same offset control. The audible result of this process is a slight emphasis of the lower partials as the gain is lowered. This is most striking when accom- plished by fast external controls (see page 35). When doing the experiments at the end of this chapter, no- tice especially what mode the gate is in. The output of each gate is independently available, or an equal mix of all four is available from the output marked “all” The Buchla 200 Series instruments do not pro- vide a variable low pass filter as a separate module. ‘A common experience for the novice is to find that the filter is blocking the sound. A signal is present at the input but nothing comes out. If this happens at least one of three common errors is being made. First, if the filter bas mixing capabilities, as with the ARP 1006, the input attenuator may be turned down. Sec- cond, the f may be so high or so low that it is beyond any spectral content present in the sound. The third, and most common error is that one is trying to filter 2 sine wave. Remember that a sine wave has only one spectral component and therefore it is the only thing that can be removed. This is not so obvious as it may appear. If a filter is @ frequency selective attenuator all it can do with a sine wave is to attenuate the whole 28 Figure 4.11. Beenie Series CSE rae thas Low Pass Gate thing...In this case the filter is being used as an amp- lifier! This is a good thing to keep in mind when you ‘run out of amps, but for the present, avoid filtering sine waves! Just as with amps and oscillators, there are as many filter modules as there are manufacturers. And this vwaveshape,-the patterns an audio electrical voltage produces. The basic waveshapes introduced at the be- ginning of this chapter are by no means models of 2 “correct” or automatically desirable musical sound. They are only starting points which are usually fur- ther shaped by various processes. Pitch, loudness, and timbre are three of many vari- ables at work in all music, and this chapter has only skimmed the surface of the musician's concer with these obvious parameters. This chapter does not sup- ly definitive information about oscillators, amplifiers, and filters, but gives the reader enough information to proceed with his studies of electronic instruments. To aid in the task this chapter has explained pitch, Joudness, and timbre as a single concept-voltage fiuc- tuation. As one becomes involved with finer means of ccontrol and is introduced to new structural situations, hhe will find that he is still dealing with the same proc. cess.—the shaping of voltages. As musical events are imagined in the mind’s ear, the composer, and more often the performer, will have to adjust his battery of modules to accommodate input, structure, and out- put. If he can visualize what is happening to the elec- tical gestures at each point in the instrument be will be more effective and proficient at his job.The Basic Patch ‘An example of this process is illustrated with an in- strument configuration that has become known as the basic patch (figure 4.12). This is a three-parameter instrament consisting of an oscillator, a filter, and an amp connected in series. Let's assume that the oscil- lator only provides a sawtooth waveshape and has no output attenuation; the filter is a low pass with only ‘a moderate amount of Q, and the amplifier has unity gain. Looking at things quite graphically, le’s quickly review each function in this chain of parameters. The oseillator’s only variable is speed or frequency (as- suming a single waveshape). The low pass filter, as- suming it receives @ spectrally rich signal, has the ability generally to re-shape with original signal by removing overtones. Since the amplifier will probably have unity gain its only variable is to attenuate or compress the signal it receives from the filter. The nature of the final musical sound is dependent on the status of these three variables. Each may be static (non- varying) or dynamic, (undergoing change as the event is being perceived). The first set of exercises are sev- feral simple sonic events described verbally, then no- tated with the normal musical conventions and then renotated in terms of how each module affects the voltage. For this exercise timbre will be notated on a seale of 0 to 10, 0 being a sine wave with all of the overtones filtered out, and 10 being an ideally ot ee Figure 4.12. unaffected sawtooth wave with f at maximum, Thus timbral crescendi and diminuendi cen be notated as in figure 4.13. This notation is convenient, as the mum- bers can stand for analogous filter settings. Likewise the amplifier offset can be notated on a scale from 0 to 10, 0 being complete attenuation and 10 being unity gain, Pitch will be notated in Hertz for the os- cillator. The important point here is not the numbers but rather the nature of the sound as these param- eters are varied, Exercises and Projects 1. Study and discuss the examples in figure 4.13. Try them out on available instruments and observe whether the sonic result is as described. If you hhave access to an oscilloscope, invent some sim- ilar examples. Take these patches apart and lis- ten to how the signal sounds as it comes out of each module. A convenient way to do this is to ‘make the patch shown in figure 414. The unpro- cessed oscillator can be heard by turing up pot ‘A on the mixer; the signal from the filter before it reaches the amp can be heard by turning up pot B, and the whole series of processes is avail- able at pot C. 2. If possible set the oscillator to a sub-audio fre- quency (you may have to use a special sub-audio oscillator), using the same patch as in figure 4.12 ‘You will bear 2 series of clicks or pops that occur at the frequency of the oscillator. How does the £, of the filter affect these pops? 3. If you have the available instruments configure three or more of the “basic patches” with sub- = OR = = OF he wos Figure 4.13. Parametic patosaudio oscillators. Vary the frequency of each shythm and also vary its prominence with the filter and the amp. If there are not enough amps and filters substitute the patch shown in figure 43. ‘At what point in the above exercise does a rhythm become a pitch? Substitute a microphone for the oscillator in the ‘basic patch. Be sure to check the mike impedance and make the needed adjustments on the pre- amp. How do the filter and amp affect acoustic signals such as your voice. If possible observe this process on an oscilloscope. ‘Build some interesting drone chords. Then using the patch in figure 415 sweep the f. of a low pass filter around to highlight different spectral areas of the drones. Try “coloring” some timbres or chords with bits of filtered noise, as in figure 416. How much noise can be added until the sense of pitch is lost? This is called masking. Is there a relation- ship between the filter's f. (filtering the noise) ‘and the oscillator’s frequency when this masking threshold is reached? ~ | =7 Figure 4.16. ce Colring with noise 8. What do input, structure, and output have to do with these exercises? 9. Keeping in mind the idea of input, structure, and ‘output, combine two or more of these exercises and make some music. 10. Make up a set of shapes or gestures. Assign a player to control a single parameter each; one player for the oscillator, one player for the fil- ter, and one player for the amplifier. Each player is then given a series of gestures with indications of how long each gesture should take, and then ‘on cue those shapes are applied manually to each parameter. For example the “score for the o:- ‘llator player may be which would indicate «steady glissando followed ‘by some high frequency variations, a sudden drop toa medium register, and then a long down- ward glissando to the sub-audio register. Like- wise the filter and amp player should have their ‘own gestural score. If the oscillator has wave- form symmetry or waveform mixing functions, apply the same type of gestural scoring. If filter Q and linear/exponential amplification are pos- sible, include them as variables. Instead of invent- ing gestures look around your environment to see what shapes are already available (e.g, the hori- zon, temperature variations through the day, the ‘movement of a “dancer” in a given space). Com- poser Charles Dodge uses a similar technique for determining parametric design in his work Earth's ‘Magnetic Field (Nonesuch H-71250). Some scores by John Cage may be applicable for this type of etude. Generalize some gestures and attempt to apply them to the individual waveforms outputted from the filter or amp (micro-structure), as well as to the overall macro-structure. LL. With an ensemble of patches like in figure 415 make a realization of Frank McCarty’s score shown in figure 417. Use only sub-audio clicks ‘and pops, varying waveshape, £, Q and amplifi- cation. A recording of one realization of this work fs available on Ocean Records, Composer's Cas- settes, Series I, IssueFRANK L. MCCARTY FACIES TEMPUS o IN WHICH “SS ——anp a "SUCH AS “soieaiad 2 ignee ane airror Sone. COMBINE TO BROBUCE <-A Pulse-Train IN WHICH Rertormers,ingiviut ly following the BERSTTYINTENSITY CUE, wil make \isely-seperated PULSES watch graauaTiy (and accorateg 22 the lar foune belo) Teeredse"tn DENSITY AND INTASETY anther retary other original seace, THE DENSITY-INTENSITY CURVE . refsiry sogire TATERFOLATIONS aay be added co tbe baste process cadtcated by che CORTE shrough brief anticipation of of returs to ether sections of the CURE. ‘olen) shoal be varked throupost, Sech changes shows ane reser the FREE INTERPRETATIONS OF THIS PLAN ARE SINCERELY SOLICITED np OM. Extenty, Figure 4.17. Tactus-Tempus by Frank McCarty 315 Concepts of Voltage Control The preceding chapters have stressed the idea that clectronic instruments require the player to think para- metrically. A musical event consists of several param- eters shaped and combined to result in a final sonic event, however simple or complex. Chapter 1 intro- duced the notion that any musical instrument demands some manner of input, a determined physical struc- ture, and 2 “musical” response. Chapters 2 through 4 limited themselves to direct manual input or control, a basic parametric structure dealing with pitch, timbre, and loudness,—basically simple musical events. By this time it is certainly evident to the musician that more complex structures will require more involved para- ‘metric organization and more intricate methods of playing the instrument. At this point it is again neces- sary to examine the philosophy of what we understand by “playing” and apply that concept to electronic in- strumentation. The following lists categorize several ‘musical variables under the basic parameters of fre- quency, spectrum, and amplitude: Frequency Spectrum Amplitude pitch (discrete timbre Joudness and gliss) Joudness shythm vibrato vibrato/tremolo tremolo timbre ‘The appearance of a particular parameter under more than one category reminds one that perceptually these categories have definite effects on each other. This list is by no means all inclusive, but is sufficiently com- plete to make the point. Learning an acoustic instrument is, in the early stages, primarily the development of kinesthetic skills Depending on the instrument, one must master all of the dynamic parameters which the instrument pro- vides, and initially this is a matter of coordination. Controlling pitch on a violin is quite a different mat- ter from controlling pitch on a piano; the control of timbre on 2 trumpet requires a different manner of {input than controlling timbre on a clarinet. The mas- tery of these skills is made more difficult by the fact that on each instrament several parameters may be linked to or associated with one input. The fiddler's left hand is responsible for pitch, vibrato, note dura- tion, and varying degrees of articulation, while right hhand activity controls Joudness, rhythm (durations), ‘timbre, and articulations. On many instruments a single parameter is directly determined by two independent inputs: 2 trumpet derives its pitch from both the de- pression of valves and embouchure variation. This sug- gests that musical events involve two types of para- metric organization: the first is a matter of percep- tion, the second is a matter of performance mode. And just as with any other instrument, both are of concer to the performer of electronic instruments. Offsets: Fixed Control Voltages AAs stated earlier in this text, contemporary electronic, instruments are a collection of possibilities. Depend- ing on the desired events, the musician must phys- ically organize his available resources—the basic patch offered in chapter 4 being only one simple approach— and then decide on some means to initiate and con- trol the resulting sound(s). As integrated electronic music systems were being developed, it became evi- dent that the complexity of dealing with many dif- ferentiated and ongoing parameters could be simpli- fied by making each parameter controllable in a conceptually unified manner. The previous chapter ex- plained the basic dimensions of sound in terms of voltage changes. A change in some aspect of the sound ‘was accomplished by changing the way voltage was ‘behaving in a particular circuit Tuming up a pot on ‘an oscillator made the voltage fluctuations increase in speed, resulting in higher pitches; raising the pot on ‘an amplifier increased the strength of the voltage fluc- tuations, resulting in louder sounds. The behavior of an audio voltage in a circuit is controlled by the application of another structural level of voltage. These controls are usually lower in frequency as they impart information concerning ac- tual musical structure such as individual mote shapes, articulations, phrasing, and overall formal design. The reader might again consider the various frequency ranges of musical organization and response discussed in chapter 2, page 9. When one tums up the pot on an oscillator, that pot applies 2 fixed amount of pos!- tive voltage to the circuit, which causes it to oscillate faster. As the pot is turned down, the amount of fixed. control voltage decreases, causing the circuit to oscil-Figure 5.1. Parametc responses with manually ‘red control voltages late more slowly. By the same token, the direct man- xual control of a low-pass filter involves manually de- termining the amount of fixed control voltage, which hhas been referred to as offset. A low control voltage results in a spectrally low cutoff frequency; raising ‘the front panel pot increases the amount of control voltage applied to the filter control circuit, thereby raising the cut-off frequency. Figure 5.1 illustrates a version of the basic patch with two different sets of manually fixed control volt- age levels. Not all electronic instrumentation operates ‘within the same control voltage range. This is a hy- pothetical example with a control voltage range of 0 to +10 volts. Figure 5.1A has a fixed control voltage of 1 volt applied to the oscillator, which results in a frequency of 645 Hz (low C). In figure 5.1B this control volt- age has been raised to 2 volts, which raises the pitch ‘one octave. In figure 5.1A the fixed control voltage has set the filte's cutoff frequency at 512 Hz, at- ‘tenuating the spectrum above the 8th harmonic, in fig- ture S.IB the fixed control voltage was lowered one volt, lowering the cut-off one octave to 256 Hz. Since the pitch was raised an octave and the cut-off was lowered an octave, the resulting spectrum is hypo- thetically limited to just the fundamental. In figure 5.1A the amplifiers fixed control voltage wwas at maximum, allowing the signal to pass at unity gain. Assuming 2 linear mode of operation (see page 24), cutting the amount of fixed control voltage in half will cut the gain in half. The dynamic marking in this case is purely subjective, and “mp” is not neces- sarily half of “Hi” Offsets establish the nature of the other strata of voltage activity in electronic instruments, and the fluctuating audio voltages are eventually heard as sound. It has been this writer's experience that the newcomer to the techniques of playing electronic in- struments is often confused by these two levels of voltage activity. Whether the difference is philosoph- ical or electrical, it must be clear in the musician's mind at what point he is working with the source of the sound or with the control of that source. Terms that will be used consistently throughout this book are signal and control. A signal is a coltaze fluctact- ing at an audio rate that will eventually be heard os ‘asound. A control is a voltaze that causes some change in the signal. Dynamic Controls A control source may simply be a manually established offset. If one stops here the delineation is simple. How- ever, if one sweeps the frequency offset pot of an oscillator up and down, causing various rates of glis- sandi, isn’t that still voltage control? The only differ- fence is that, the control voltage itself is fluctuating. ‘A control may indeed be a fixed offset or it may be continually varied. A varying control is referred to as a dynamic voltage. This and succeeding chapters will cover a variety of control voltage sources which pro- uce static and preset controls, as well as control volt- ages that fluctuate at speeds from sub-audio to audio rates. It is still essential that the musician understand the difference between the signals and controls—signal is sound and control is structure. Both are voltage ac- tivities, and what is musician A’s sound may be mu- sician B's structure. For musicians A and B to perform efficiently, however, each must be very certain in his ‘own mind about the flow of these two kinds of in- formation. On some instruments this differentiation is ‘entirely user-determined, while on others the levels are physically separated by types of patch cords or 33switching formats. The purpose of this text is to ex: plain only the basic concepts and techniques of elec- tronic instrumentation in general, and to avoid dis ccussing the philosophical merits of various instrument designs. Parametric Response to Controls It should be obvious that direct offsets or manually manipulated dynamic controls are not the most effi- cient way to play electronic instruments. If one had to contend with just an oscillator it would still re main a complicated if not impossible task to play a precise musical scale. When this is combined with si- multaneous controls of filters and amplifiers, one can easily understand why the majority of early literature of electronic instruments was subjected to tape stor- age. Since the late 1950s and early 1960s, designers hhave continued to explore the area of performer in- put, making a variety of control voltage sources avail- able to the musician. A control voltage source is a cuit which produces 2 variety of changing control volt- age levels. Specific control sources will be considered in the next chapter. Voltage Controlled Oscillators A series of control voltage applications will be ex- plained in terms of an oscillator, a5 notated in figure ‘52. Inespective of the source of voltage, it may be fixed or fluctuating. If the control voltage is fixed, it is usually an internal reference which is offset by the front panel pot. This offset may then be added to or subtracted from by an extemal control voltage source. The external control voltage is internally mixed with the offset voltage, resulting in proportional changes of the behavior of the parameter for which the particular circuit is responsible. Figure 5.2A in- Gicates that an oscillator is to be manually offset to middle C (and presumably attached to a power amp and speakers). Instruments vary in voltage ranges and offset values, so that we will assume hypothetically that an offset of +3 volts will produce middle C. Since ‘we are exploring the process of control voltages, the oscillator will be referred to as a VCO (Voltage Con- trolled Oscillator). Figure 5.28 attaches an external control voltage of 0 volts to the VCO which is mixed with the internal offset. This obviously still equals 3 volts, which will result in no pitch change. Figure 52C adds an ex- ternal control of 1 volt, mixed with the offset to equal 4 volts. In this case the 4 volt sum causes the VCO to produce a pitch one octave higher, C 512. Many, ‘but not all, instruments have 2 one volt per octave response, but this standard will be assumed, in order to keep the numbers simple. Check your operation ‘manual in regard to your own instrament. We can 34 then postulate that each additional volt of control will raise the VCO frequency one octave. In some instru- mentation, a control may also be a negative voltage. ‘This poses no conceptual problem, since itis then only ‘2 matter of subtraction. A S-volt offset (middle C 256) plus ~1 volt, or 1 volt negative, is 2 volts, lowering the ‘VCO one octave to C 198 (see figure 52D). The following example, figure 53, illustrates the need for exponential response in the control of cer- ‘tain parameters (see page 9). Figure 5.34 illustrates hhow two oscillators which are offset to the same ref: ‘erence will sound 2 unison. If a single control voltage is applied to both VCOs, they should respond the same way, maintaining 2 constant unison. Exact track- ing of exponential oscillators can often indicate the quality of the instrument. Slight variations in re- sponse can be dealt with by syncing methods to be Aiscussed later. Figure 5.3B is a similar patch, but note that the VCOs are offset at the interval of an octave. VCO 1 has a 2 volt offset (C 128) and VCO 2 has a 3 volt offset (C 256). If both VCOs receive ‘2 1 volt control, they must respond exponentially to maintain the octave relationship. In other words, both ‘oscillators must double their frequency to keep tuned. ‘to an octave interval. VCO I changes to 256 (128 x 2) and VCO 2 changes to 512 (256 x 2). Both oscillators respond with the same intervallic change; but note that the change in terms of numerical values of Hertz was different. VCO 1 went up 198 Hz and VCO 2 LIF the orcilatrs in an instrument do not track, consult the ‘iaintenance manual for instructions on trimming.1 cr @@ ] t) | ft | Ie it) | € @) i 3 | == Figure 5.3. Votage conroted esciator = sacking ‘went up 256 Hz. Equal changes in musical intercal, Voltage Controlled Amplifiers regardless of offset values, require exponential re The control voltage examples thus far have been dis- ‘sponse. If the response were linear, for each volt of crete values (2 volts, 3 volts, 4 volts), all causing in- ‘control there would then be equal changes in the stant changes in the VCO's frequency. It is rather evi- terval. As in figure 5:3C, VCO 1 would perhaps change to 256, a numerical change of 128 Hz and en intervallic change of one octave; VCO 2, however, would only 0 up to G, an intervallic change of a fifth. To reit- erate an important principle, exponential response ‘ain musical responses at other levels of organization Tequire linear change, and this will be dealt with when the subject demands. If the composer/performer is in need of octave relationships and traditional orchestra- ‘ional devices, then exponential response is required.* 2% Some instruments, sich at the PALA, have linear ontors SE} Ratan exponential relationships by greeting exponen- dent that not all musical states change instantly or remain fixed for long periods of time. A crescendo, glissando, vibrato, and vowel shifts in a song are all examples of continuously variable parameters. To ac- commodate this operation the musician must have de~ vices that produce transitional or dynamic control volt- ages. Loudness is a parameter which usually requires some sort of evolutionary or continuously variable con- trol. Figure 5.4 illustrates a VCO offset to middle C patched to a Voltage Controlled Amplifier, or VCA. ‘The gain of the VCA is controlled by the sum of the intemal offset voltage and the applied external con- trol voltage. Figure 5.4A indicates that the VCA is offset to produce a loudness perceived as “mp"—per- haps 4 volts. Some dynamic extemal control voltage begins at 0 volts and increases over a two-second pe- riod to 10 volts, then decays once more to 0 volts. 35Figure 5.4. VGA response 10 a dynamic contr When applied to the VCA, the loudness will increase from its initial “mp” offset to a full “forte.” then fall back te “mp.” in proportion to the activity of the ex- temal control. As the external control voltage rises, it is added to the intemal offset. Note that at the highest external control value, the VCA is actually receiving a total sum of 14 volts, but the circuit may be designed for a 10 volt maximum. Usually if a mod- ule receives a control which exceeds its response capa- Dility, it will respond only to its peak value and the excess control (or saturation) will not harm the cir- cuit. This does not mean one should plug 110 volts into the control input of a module or even attempt some type of interaction between instruments. Check the operations manual before doing something that is not obvious on the instrament. If the VCA's offset is at maximum and one attempts to use an extemal positive control, nothing will hap- pen. The VCA is already at maximum gain and an- other control cannot make it any louder (usually!). In figure 54B the VCA is offset to 0, and no audio signal is passed. The signal will not be audible until ‘some external control raises the gain to 2 perceptible level. Application of the same control as in figure 54 will result in a crescendo from silence to“ and back to silence (the offset). It should be clear that negative controls cannot lower the gain past 0, but can lower any positive offset value. More will be said about this in chapter 7. Coexisting Controls Dealing with two ongoing controls is i complicated and is illustrated in figure 55A. A VCO, over a particular time period, receives 2 series of three discrete control voltage changes—1, 2 and 3 volts re- spectively. Assuming an offset of 198 Hz, the three is applied to 2 VCA with an offset of 0. The resulting sound would be three octave Cs, each beginning with silence and crescendoing to “ff.” If the controls were evened as a Signe EAD, the soe resst wonld be pward glssando at ‘three increasing loudnes Steps, Since the dynamic voltage goes from 0 to 10 lees locee ears oscillator does not go out of audible ange, the VCO sould be effet to some lower fre. that 2 10 octave range can be accommo- dated. It should be mentioned that these control vol- their production and control. The coexistence and re- lationship between a multitude of control voltages de- fines 50% of an instrument's structure. The other 50% Control Voltage Processing ‘Just as an initial signal from a VCO is not the final Sonic image in practical music making, most control voltages must be tailored and processed for certain applications. Once control has been generated, it may be processed in much the same manner as an audio signal, and a single control may be given @ ‘variety of identities or transformations. A control for- sat for an instrument may then consist of a numberof unrelated controls or a number of voltage trans- formations derived from a single source. A control may be transformed by means of attenuation, multi- plication, integration, interpolation, or incersison. De- tailed explanation of the processes will be covered as the various applications are discussed. The following definitions will provide sufficient information to make basic control voltage generation meaningful. Attenuation Just as an amplifier can attenuate a signal (see page 23), a control voltage may also be lessened in strength. For example, three one-volt steps usually re- sult in three one-octave skips when applied to an un- attenuated VCO. Most voltage controlled circuits have front panel control voltage attenuctors, attenuating in- put or processing inputs. By applying 2 voltage to an attenuating input, the incoming control can be compressed to any desired voltage range. In figure 36 the original three one-volt steps are attenuated by 50E to produce three volt steps, each in tum pro- ducing a skip of an augmented fourth. Note that the voltage ratios do not change, only their composite range. It should also be emphasized that input attenu- ation has no effect on the source of the control. The voltages are produced at full value, and. the attenua- tion takes place beyond this point of generation. ‘The majority of the EMS instruments, such as the Synthi AKS, are based on control voltage output at- tenuation. The magnitude of 2 control voltage is de- termined at the source of the voltage; and it is not possible to attenuate controls separately for every parameter. This imposes limits on instrument struc. turing but, as with any instrument, # the user is aware of the design, he can use it successfully. Some instrumentation has two kinds of control in- puts, “fixed” and “attenuated.” In the case of VCOs, the fixed imput is usually referenced to one volt per cetave and cannot easily be changed. This obviously is convenient for standard tuning. Other inputs may have variable front panel attenuators so that the con- trol voltage can be tailored to other needs. Some manu- facturers have designed instrumentation so that when the attenuator or processing pot is at maximum, the ‘VCO, filter or whatever will respond to one volt per octave. Control voltage attenuation can also be accom- lished by separate control voltage processors, and in some instruments by AC/DC coupled mixers or omp- Bjiers. Unlike a dedicated audio mixer, this type of mixer can accept and mix control voltages as well as audio signals. The Moog $84 Mixer is an example of such an instrument. A control may be taken into any available input, then freely attenuated and mixed ‘with other incoming controls. The final composite sig- nal then appears at the output. If only a single con- ——[E_ email: Figure 5.6. Conrot votage attenuation trol is taken, the module can be used just for attenua- tion as determined by the input pot. This is basically the same sort of circuit used within a module to mix manually the internal offset with an incoming com tol. Both the manual offset and the extemal control have to be mixed or summed before being applied to the controlled parameter. Coupled mixers such as the Moog 984 Mixer and the Eu 2010 Amplifier make this process possible by means of extemal modules. Un- less such a mixer provides some gain, ‘expansion of a control voltage will be impossible. A gain of 2 would mean that an incoming 2 volts would possibly be ex- panded to 4 volts, 3 volts to 6, and so on. ‘Putting these concepts into acoustic instrument an- alogies is a bit challenging, but it may be pedagogic- ally useful. Consider a guitar string in analogy to a ‘VCO. The tension of a string could be considered the offset, tuned to a specific pitch. The player's fingers, as they are placed on the strings, are analogous to control voltages. Beginning with the index finger, each successively higher finger produces a higher pitch. Changing the tuning of the string would constitute a change in offset, with the controls still having the same effect. Modification of the fret placement—for example, placing them closer together—would repre- sent the process of control attenuation, resulting in the production of smaller intervals. Inversion Investigating the process of inversion, one has to un- derstand the nature of the control voltages for each instrument. Some instrument designs use a control range operating on both sides of 0 volts—positive and negative control. Other instruments restrict the con- trol to the positive side of 0 only. In each case the process of inversion is slightly different. In the case of positive and negative control, inversison would be an inverse value around 0 (figure 5.7A). A positive voltage inverted is simply a similar negative value, In terms of application, suppose that a control raises the pitch of an oscillator one octave; if that voltage ‘were inverted it would then lower the pitch one oc- tave (figure 5.7B). 37If the instrument's control range is only positive voltage, inversion must be thought of in a different way. If the controls are only positive it would be im- possible to take a parameter below its offset value. ‘The lowest value a control could have is 0 volts, which will have no effect on the controlled parameter. How then does one invert a control that never goes below ©? Figure 58 illustrates inversion in relation to 0, as in the earlier example: +5 becomes —5, +7 be- ‘comes —7, and so on. Instead of locking at this as @ ~10 to +10 volt range, consider it to be simply @ total range of 20 volts. Now by moving the 0 volt reference line to the point of lowest voltage (figure ‘58B), we observe that what was +5 is still +5, +7 is +3, etc. What has happened is that 0 volts has ‘been established as the lowest value, and all controls are on the positive side of 0. Inversion now becomes 2 process of reciprocal values within the total control range. The inversion of a value is arrived at by sub- tracting that value from the total available range. Within a 10-volt operating range, the inversion of 5 fs still 5, the inversion of 7 is 3, the inversion of Gis 1, etc. Reciprocal inversion could possibly be ex- plained more simply; but showing its relationship to ositivenegative inversion as merely a re-establish- ‘ment of the lowest voltage value will be important toa conceptual understanding of all electronic instra- ments. Now comes the question—why invert? Why would someone wish to control something upside down? Con- sider 2 left-handed guitarist having to learn from scratch on a right-handed guitar. While there is no zeal problem with the coordination, all the books are upside down. Admittedly 2 simple example, but it makes the point about accommodating various inputs to an instrument. There is no acoustic law which states that highest notes on a keyboard have to be on the ight. Alter all, on an accordian the highest notes are down! This problem also relates to more complex in- strument structuring. Consider a case in which two oscillators are to produce mirror scales of each other. For one VCO you would need an ascending series of voltages and for the other VCO a descending series ‘of voltages. Instead of using two separate control sources, Why not use one source driving VCO 1 di- rectly and at the same time invert another leg of the same voltage to drive VCO 2 (see figure 5.9A)? As- suming that the controls regulate fixed voltage inputs, the only tuning required is the offset. The complexity of control is simplified and performance deals with ‘only one control source, not two. ‘Another example of the use of inverted control is illustrated in Figure 59B. This deals with more com- plex parametric correlation, but is certainly worth thinking about at this point. The task is to generate a specific pitch-timbre relationship in which all high pitches have simple timbres and lower pitches have proportionally richer timbres. Simply use the inverted form of the voltages driving the oscillator to control @ low-pass filter. A high voltage produces a high pitch, and simultaneously, its inverted form, a low voltage, es- tablishes a low filter cut-off frequency, removing the desired amount of harmonics. Integration Certain control devices such as keyboards and sequen- ‘cers (see pages 45 and 70) can produce either man- ually activated or pre-programmed voltage steps, as in figure 5.10. If applied to a VCO, the result is amom SS SS Figure 5.10. integration series of discrete pitches. The process of integration is comparable to a portamento or glissando between ‘and through the pitches. The time it takes to integrate ‘two voltage levels is called the “time constant.” A te of 2 means that it would take 2 seconds for a voltage to integrate or slope from its previous value to a new: value. The portamento control on an electronic key- board is a time constant pot which sets the porta- ‘mento rate. Some other control sources have integra- tion possibilities built into the instruments, while other instrumentation provides integration possibilities by means of a separate module. Other terms for integration are Tag and slew. These terms actually describe specialized integration func- tions. A “slew” can be either positive or negative. Serge System instruments provide modules dedicated to either positive or negative integration and are called the Dual Negative Slew and Dual Positive Slew (dual Im z mee meaning that it can process two independent voltages). The negative slew will integrate between descending ‘voltage levels only (see figure 5.11). ‘A Lag Processor, as on the ARP 2600, is usually associated with a device called an envelope detector (see page 53), which is used to convert an incoming audio signal into a control voltage. Acoustically gen- erated sounds have very complex waveshapes and are usually too transient, with very rapid fluctuations, to be used as controls. The lag is essentially a negative slew that slopes off the descending portions of the waveshape. More will be said about integration func tions associated with envelope detectors on page 53. Some integration slopes are linear and some are logarithmic,—linear meaning that the rate of voltage change is constant, and exponential meaning that the closer the voltage gets to the next value the slower 39it changes. If vour available instrument provides both possibilities, use your ear to determine your prefer- ence for each controlled parameter. If the instrument has only one slope, don't worry about it-the differ- ence is usually very subtle. It may, however, be use- ful to realize that with exponential curves the final value is never reached within the slope. At some point the slope is broken and the final value is latched onto, actually breaking the integration function as the slope approaches the final value. However, this latching is almost inaudible. I suppose someone could make a ‘case for linear integration, with the full transition heard since the rate change is constant, but . . Examples of acoustic counterparts of integration are evident, but some thought might be given to con- ventional music terms for various integrated param- eters: Integrated pitch slissando or portamento Integrated dynamics = crescendo and diminuendo Integrated tempo = _accelerando and ritardando Quantization (Quantization is the opposite of integration: it takes a continuously variable voltage and divides it into on- going discrete steps or values. Modules such’ as en- velope generators or random voltage sources (low fre- quency noise) produce continuously variable voltages. Figure 5.12 illustrates a voltage sloping from 0 to 10 volts. If one quantized this slope into five equidistant values, it would produce a sequence of 0, 25, 5, 75, and 10 volts. Slopes always quantize to the nearest pre-determined quantization value. If the 10 volt range is to be quantized into 5 equal steps, it would first produce 0 volts (step 1); then as the slope rises past 1.35 volts, halfway to the next quantized value, the ‘output would switch to 25 volts. The output would remain at 25 volts until the slope reached 3.75 volts (halfway between 25 and the next value, 5) and at that point switch to 5 volts, and so on. This process is not to be confused with sample/hold techniques (see page 80), since quantized values are usually fixed by the instrument or pre-determined by the per- former. string of a standard concert bass is capable of infinite pitch selection, and the performer can slide or glis- sando between any interval within the range of 2 single string, The standard electric bass guitar, how- fever, is quantized into equal tempered half-steps, so that all the performer must do is come close, making sure be places a finger bebind each fret or integration value. Harry Partch’s Adapted Guitar relocates the frets so that the guitar integrates at different values, with the player generating his own intonation require- ments. 40 Figure 5.12. Each interval quantization is often built into pro- grammable voltage sources such as the EML 400 Se- ‘quencer and the Buchla 248 Multiple Arbitrary Func- tion Generator. With the increasing use of digital control of analog instruments, user-determined quan- tization can be of immense value in the exploration ‘of non-tempered seales. Muttiptication This is not easy to describe, since it is difficult to find 2 simple acoustic analogy. Voltage multiplication is a process whereby one value (a voltage level) is de- termined by another value (2 second voltage level). If a series of voltages—0, 2, 4, 6, 8, 10—were applied to one input of « multiplier, the output voltage would bbe dependent on the voltage level at the other input. For example, assume that the multiplier is 5 volts. Since the total voltage range of the system is 10 volts, 5 represents a multiplicand of 5, one-half the total range. In this case the output series would be 0, 1, 2, 3, 4, 5 volts (figure 5.13A). If the multiplicand were 7 volts (figure 5.13B). Things get a bit more com- plex if the multiplicand is dynamic. For instance, sup- pose the multiplicand switched from 1 to 5 to 10 volts. This would represent values of .1, 5, and 1. If the ‘multiplicand changed each time the series began at 0, the output would be 0, 2, 4, 6, 8, 1,0, 1, 1,23, 4,5, 0, 2 4, 6, 8, 10 (figure 5.13C). ‘As mentioned earlier, such product functions are not so apparent in acoustic devices, especially where ‘both variables are performer determined. One such ex- ample might be the “Jaw’s Harp” The loudness is a product of the amplitude of the vibrating metal strip and the resonance of the performer's mouth. In the open air the instrument is barely audible, and a resonant space i useless without a vibrating source. ‘When combined, however, they produce loudness functions which can be controlled by either variable, ‘one determining the other.Ais wearsres Control voltage multiplication is also possible with an AC/DC coupled VCA,—one which can attenuate « control voltage by means of another incoming con- trol. With the offset at 0, the value of the VCA out- ppt voltage is the product of the input voltage (in this case 2 control) and the applied control voltage. ‘The Moog 902 Voltage Controlled Amplifier is cap- able of this process, with the added capability of in- verted or normal outputs (see figure 5.14). There are some other techniques which may be used to simo- late multiplication, and these will be pointed out in applicable situations. Now, what does a musician do with this plethora of information? It is through the processing of controls that the performer or composer determines the com- plexity of responses from his instrument. Consider a situation in which some traditional performance con- trols could be sensed and transformed into analogous control voltages. These controls could be processed in various ways, then redistributed to different param- ters of an electronic instrument. Suppose, for ex- ample, a string player producing the event notated in figure 5.154, a repetitive D major scale with a cres- cendo and dimimuendo; with suitable instrumentation the pitch and loudness activity could be turned into usable voltages. A Pitch-to-Voltage Converter (see age 55) would transform the strings’ frequencies into Figure 6.13. Matiptcation Figure 5.14. Moog 902 VCA aFigure 5.15. Parametric raster a series of voltages. Assuming a 1 volt/octave conver- sion with the lowest voltage being 0, the output would be a series as shown in figure 5.15B. The loudness variation could be sensed by an envelope detector (see page 53) and its output voltage would be as iMlustrated in figure 5.15C. Two obvious inputs have been extracted from the performer and transformed into independent controls. Now let's put the instrument ‘back together in a new way. We could simply reverse the controls: the rising and falling voltage originated by bow pressure would cause a rising and falling pitch, and the discrete voltage steps caused by finger placement would produce sudden changes in loudness (figure 5.15D). Now let's complicate things and add some control processing. In figure 5.16 the original pitch is multi- plied by 5 (5 volts, assuming 2 10 volt range) and the loudness contour is inverted. Redistributed to the instrument, the voltage sequence for the pitches would be one half the original values, producing 2 Snote quarter-tone scale, and the loudness contour would be a diminuendo followed by crescendo. Needless to say, one could incorporate quantizers, at- tenuators, more multipliers, and other elements to com- plicate the exemples. If you want to attempt it, try the exercises 1 and 3 at the end of chapter 6, and let your imagination run free. This is the heart of dealing with electronic instruments, and if you can come up with the ideas, they can probably be implemented, a2 Applications This chapter has been largely conceptual and men- tions a very limited mumber of specific instruments. Consequently, it would not be practical to suggest any hardware-oriented projects at this point. Nevertheless, ue to the importance of voltage control concepts to the electronic media itis strongly urged that the reader spend a day or two considering some ideas before go- ing on to chapter 6. The next chapter contains what are perhaps the most important pages in this text, and it & based on the reader's grasp on voltage control Figure 5.16. Parametric transfor and processingIt is suggested that one use the following situs. tion as a basis for some self-constructed exercises, con- siderations, and possible group or class discussions. ‘Theorom 1: Anything which can be measured can be translated into an analogous voltage. ‘Theorom 2: Voltages may be changed or processed by: Addition to another voltager Subtraction by another voltage Inversion Integration Quantization Multiplication by another voltage ‘Theorom 3: In an ideal situation within the electronic ‘media, any desired musical parameter may be de- termined and controlled by any voltage. ‘Theorom 4: Given theoroms 1, 2 and 3, you are free to create as you wish! MOO ee This is, of course, what every composer or per- former dreams of doing. However, used as 2 basis for various hypothetical constructions, this goes to the Iheart of the subject at hand: conceptualizing the re- lationships between control and resulting sound This exercise is to take any imaginable situation and isolate some variables within that situation. Once the variables have been defined, assign their units of measurement to 2 voltage range. It is suggested that you check your instrument manual and work within ‘thase voltage boundaries. In this way you can begin to think in terms of a specific instrument (the example given below will be based on a = 10 volt range). Since ‘measurements can be translated into precise voltages one can have any measured action control various ‘modules of an electronic instrument. The best place to begin is in terms of the “basic patch” explained in chapter 4 Example: A Sonic Weather Machine The measurable variables: Corresponding voltages: (subjective assignments) A-temperature: each 10 degrees Fahrenheit will equal 1/2 of a volt. B-speed: each 1 mile-per-hour measurement will equal .1 volt ‘C—humidity: every 5 humidity will equal 1 volt Voltage assignments: ‘Tt is here that one decides what voltage activity will control each part of the basic patch, Avair temperature controls the VCO. Each 10° F. equal 1/12 volt. Assuming that the VCO oper- ates on a 1 volt/octave range, each degree will change the VCO by ¥ step. B-air speed controls the VCF (Voltage Controlled Low Pass Filter). Assuming a 1 volt/octave range, each mph change will move the filter cut-off one octave. (C—humidity controls the VCA. Every 5% humidity will change the gain 108. Take special note of the offsets; they should be indi- cated in each patch The VCO is a sewtooth wave so there is a rich spec- trum to filter. The frequency is offset to 128 Hz (C) The VCF is offset to the fundamental pitch of the VCO. The VCA is offset to 2 gain of 5 (control of 50%) so the humidity variations can potentially create eres- cendi and diminuendi. ** GO BACK AND READ THIS EXERCISE AGAIN SO IT IS WELL UNDERSTOOD UP TO THIS POINT °* Conditions: It is 2 calm and stable day F Results: Pitch would be a perfect Sth higher (7 half-steps) Filter cut-off would be 2 octaves higher (512 Hz) Amplitude would be increased slightlyFigure 5.17. Sonic wind machine a a Figure 5.18. Sonic weather machine with applied conrais Problem 1: What happens is the temperature sud- denly drops 20° and, air speed increases 1 mph and, humidity increases 208? Problem 2: What sonic result would take place if the above changes took place steadily over a one minute period? Problem 3: How could you cause the pitch to rise in response to a felling temperature? Problem 4: What would be the result if the VCA and ‘VCO controls were interchanged? Problem 5: After the change described in problem 4, how would you get a smoothly changing humidity to result in discreet pitch changes? ‘Problem 6: Using only the basic patch, invent some new situations for controls and assign them to the instrument in various ways. Be sure to consider rela- tionships and offsets for each control. If this seems too easy, try processing the voltages and dealing with two or three basic patches at one time.6 Control Voltage Sources Control voltage sources are the basic organs of con- temporary instruments. The possibilities of control processing were introduced before discussing individ- xual control voltage generators, so that here it will be possible to discuss control sources in a more con- textual relationship to their applications. This chapter ‘will introduce a variety of control generators and make some suggestions concerning obvious and perhaps not so obvious applications by the musician In most instances of recorded electronie music, the composer fills @ dual role of creator and performer. In the older studios of Europe, it was a common prac- tice for the composer to realize his work through the aid of an engineer; he would seldom come in direct contact with the equipment. But due to simplified ‘methods and design, along with the contemporary com- poser’ increasing kmowledge of electronic methods and ‘an instinetive curiosity about the internal workings of his art, the electronic music composer today holds a tighter rein on the compositional processes. One of the major appeals of electronic music is that if offers the composer an opportunity to come into direct con- tact with the various parameters in which he is in- terested. Some composers appreciate this because of the unrestricted control it affords; others value it be- cause of the actual kinesthetic sensations involved, much as in the act of painting or sculpting. There is an endless number of kinds of voltage sources which the composer may use; and every studio may have its own particular version of a basic type that has been built or modified to meet a specific need. This chapter will discuss the basic resources of control sources and some of the more common re- finements and modifications found on much of the equipment which is commercially available. There will be no attempt to present any particular design in its cntirety, and all ofthe discussed refinements and modi- fications are usually not found on any one system. ‘There are so many individual variations in design that complete coverage would amount to a detailed de- scription of every electronic studio in the world. For purposes of lucid and coherent discussion, this overview will be divided into two sections. The first ‘will deal with kinesthetic sources, that is, those in which the generated voltage is the result of some di- eet and discrete action by the performer, such as with various types of keyboards, pitch-to-voltage converters, joysticks, and such. The second section will concen- trate on programmed controllers; in this case a voltage function is pre-set by the instrument or by the per- former and called into action as needed. In some in- stances this categorization is not clear because of a complexity of inputting possibilities. In these situe- tions the modules will be dealt with at a point that seems most logical. Kinesthetic Control Voltage Sources Keyboards The most popular, but not necessarily most useful, source of control voltage is the voltage keyboard or Keyboard controller. This is essentially a multiple volt- age source that will produce a different instantaneous voltage for each key activated. The keyboard may be similar to an organ keyboard, or it may be a series of metal plates (commonly known as capacitance plates or a “touch keyboard”) which are touched by the per- former. The similarity to an organ or piano keyboard is often confusing because the voltage keyboard need not necessarily change pitch. It produces voltages which may be used to control pitch as well as amp- litude, timbre, or any other desired parameter. Ideally, each individual key would have the capability of pro- ducing its own voltage independent of the other keys. ‘This may -be achieved by each key having its own voltage pot which could be manually set. Because it has individual controls, the keyboard is not limited to having any particular voltage sequence; going from left to right would not necessarily mean higher volt- ages. By depressing the individual keys, the performer would effect a change to the preset voltage of that particular key. Note that in figure 6.1 there is only one main out- pput for the voltage. The basic voltage keyboards are “monophonic,” since they will produce only one volt- age at a time, that is, if controlling an oscillator with them, the player would produce only one pitch at @ time. The output may be divided to control any num- ber of modules, but each module would receive the same voltage. Some keyboards will produce the pre- set voltage only while that particular key is depressed. 45This is often referred to as a “bold-no hold” function; certain designs of keyboards will function in either mode, selected either by a “bold-no hold” switch or by separate “hold-no hold” outputs. ‘Over the years there has been much discussion, arguments, and design variations relating to poly- phonic systems. My own opinion is that the synthe- sizer is, at least as of this writing, essentially not a polyphonic device. There are indeed many types of “polyphonic” keyboards and dedicated “polyphonic” electronic instruments. What the performer geins in polyphonic capabilities from a single control source, ‘one usually loses in user-implemented design possibili- ties. Ifind great difficulty in trying to remember which type of polyphonic keyboard has low or high note priority, which portions of the keyboard go to what ‘module, and become quite confused amid the “some- times, always, never, continuous, reset, reassign” op- erations. Manufacturers of polyphonic keyboards sup- ly the user with all necessary operation information, and therefore I will leave such descriptions to the user manuals. It would hopelessly confuse the reader if in a single text one attempted to cover all possible keyboard operations. I will deal only with the mono- phonic voltage keyboard; a thorough knowledge of this type of controller will facilitate the understanding of polyphonic systems, and all of the information will be readily transferable. Computer-aided polyphonic performance is really 2 different situation. In this case the performer can rogram a series of parallel functions. But now we are talking about computers and micro-processors, rather than manually operated keyboards. The state- of-the-art system is a multiphonic instrument capable of a myriad of interesting sounds and structures. Our ‘own physical resources of ten fingers would place a severe limitation on the control and implementation Of these structures if they were only accessible through dizect control of a keyboard. Perhaps this statement 48 Figure 6.1. Basic votage keyboard _——— *Q900Q000G Tin "OCSO5S66 — oe, mH may offend some manufacturers and a sufficient num- ‘ber of musicians. This, of course, is not my intent. I am only saying that the concept of polyphony can- not be limited to what is available from pianos and organs. An orchestra is also a polyphonic instrument, ‘we know how many hands it uses. A seasoned per- former with access to a large studio system still could not create that same level of structural complexity in real time if he were limited to a keyboard, no matter how polyphonic it may be. He can, however, pre- program functions and timbres to react to a keyboard, 2 computer or the eyeling of the moon; and I prefer this concept of polyphony to my limitations with a keyboard. Figure 62 is a type of “homoghonic” keyboard. ‘With this particular design, each key is able to pro- duce two independent preset voltages. The top row ‘of pots are preset and the voltage is taken from out- put A. The lower row of pots are also preset and their ‘output voltage is taken from output B. It is thus pos- sible to have any single key produce two independent voltages which may be used to control two indepen- dent modules. In certain instances it may be very useful to have 2 third independent output with its own set of con- trol pots. Each output would also have a set of cor- responding pots for the presetting of each key, and each separate voltage output could then be used to control any voltage-controlled module in the system. ‘Amore common keyboard design is the voltage di- vider format which divides a control voltage range into ‘equal steps, resulting in fixed equal division of pitches. This keyboard design achieves voltage regulation by28 = Trae. soe Leb monophonic keyboard using only two pots: “range” and “scale” (see figure 63). The range pot determines the total amount of voltage the keyboard will produce. If the kevboard is being used to control the frequency of an oscillator, and the range pot is at 2 maximum, the pitch differ- ‘ence between the two extreme ends of the keyboard may be as much as 15 octaves, depending on the par- ticular design. If the range pot is at a minimum set- ting, the total range of the keyboard could be as small as one-half step. If the keyboard had 40 keys, then there would be a possibility of having 40 different pitches between any two adjacent half-steps. The re- lationship between the individual voltages is con- trolled by the scale pot. This control allows for fine tuning of the voltages so that itis possible to achieve various types of scalar relationships of the controlled parameter. To facilitate performance of traditional 12- tone equal temperament music, the smaller perfor- mance systems such as the ARP Odyssey, the Cat, and others have the keyboard pre-tuned to standard intervals. Certain keyboard designs make use of a key- board programmer. This is a memory circuit which ‘can be instantaneously switched in to provide various pre-programmed tunings. It should be recalled that very fine and subtle control of all of the parameters {s just as important as the fine control of pitch, and that the Keyboard should not be approached as strictly a frequency controlling device. Since the typical monéphonic keyboard can only produce one output voltage at a time, what happens if two or more keys are depressed simultaneously? Most mechanical keyboards are of the voltage divided genre which have low-key priority. If a high key is held down, as soon as 2 lower key is depressed the voltage output will be that of the lower key, even if the higher Key is still held. This can be used to ar- ticulate wide interval shakes or tills by holding down a high key and rapidly depressing and releasing a low ‘key. Other keyboards incorporate last-key priority. It {5 virtually impossible to touch two or more keys at exactly the same time, and the voltage output is that of the last key, no matter how minute the time inter- val. Dual output keyboards function in various ways. With two voltage outputs, one output usually corres- Ponds to the highest key and the other to the lowest key. Such features as dual voltage output can be put to some interesting uses by using the main voltage to control a VCO and the second voltage to contro! another parameter such as a filter. In this way the top “voice” controls pitch and the lower “voice” con- trols timbre The voltage output from a keyboard may be hard- ‘wire connected to an oscillator (meaning that the user cannot disconnect it), connected and disconnected by means of 2 switch, or may call for user patching by means of cables, plugs, etc. When using the keyboard to control an oscillator, one still must determine the CO's offset. This means that when you first ap- roach the instrument, a C on the keyboard may not necessarily produce a C on the VCO. A logical tuning procedure is to touch the lowest key, meaning that the control voltage is at its lowest value, usually 0 volts Making sure that the keyboard is connected to the VCO, adjust the offset to any desired refer- fence pitch. If for some reason you are asked to play a By hom part (written in Bp), tune the VCO offset to produce Bp (in the correct octave) while the C key on the keyboard is activated. Some keyboards have the availabilty of preset references. Additional keys can add additional voltage to the keyboard output to transpase it by octaves, fifths, or any desired interval. Figure 6.4A illustrates this process. Three extra keys with associated pots may be used to set three voltage levels of 0, 1, and 1.5. By touching pre-set 1 the output voltage, 0 volts, is added to the ongoing keyboard voltages, perhaps producing a repetitive triad. Zero plus whatever the keys are generating has no effect on the final key board output. Activating pre-set 2, however, adds 1 volt to each key voltage, making the final keyboard output one octave higher. The result would be an oc- tave transposition of pitch. The third pre-set sums 15 volts with the keyboard voltages to make the trans- position an octave plus an augmented fourth. This same result could be accomplished by either playing 2 compound augmented fourth higher on the key board or by re-setting the offset on the oscillator (either manually or by means of a second control volt- age source). Preset keys or switches are useful for keyboards with a limited number of keys and place the transposition control in the immediate area of the generation of other pitch controls. ‘An alternate Keyboard format favored by some players is the matrix keyboard. Similar to a calculator keyboard, it is usually in a 4 x 4 matrix with each key user-tunable to any desired control setting. Ma- trix keyboards are available with either touch-sensi- tive or mechanical action keys. The EML Manual Con- troller (see figure 65) is a 16-key matrix providing two independently tuned voltages for each key. The “sampled voltage” output is like most other keyboard 1. Sometimes the zero volt reference i in the middle of the ‘keyboard, so that it generates positive and negative control. 47Figure 6.5. EML Manca! control voltage outputs in that each key depression is memorized by the circuit and the output voltage will remain at that level until 2 new key is depressed. The “voltage” output, however, is a no-hold process in which the control instantly drops to zero as soon as the key is released. Some applications of homo- phonic keyboards were given earlier, but this no-hold function provides additional possibilities. The “sam- pled” output could be used to control 2 VCO, and the “voltage” output then could be used to control the amplitude of a VCA. Since each key has separate tunable voltage outputs, each pitch could be pre- programmed to a specific loudness. A possible ad- vantage here is that as soon as the Key is released, the “voltage” retums to zero, thus instantly turning down the VCA. In effect this produces rather crude “envelopes” but still is useful when you run out of envelope generators (see page 64). The real advan- tage of the matrix keyboard is in the relationships one can establish between performance gestures and tun- ings. Tracking and Tuning, A format common to many in- struments is to have the keyboard patched to 2 VCO and simultaneously to 2 Voltage Controlled Filter (VCE). With this patch one can offset pitch and filter cut-offs, and both will track in parallel (see figure 48 6GA). Why do this? If, for example, a low pass filter fs offset to 512 Hz, the frequency component above that cut-off will be attenuated. That means as the VCO produces pitches closer and closer to that cut-off they will contain less and less spectral components and the higher notes will have proportionally simpler timbres, (figure 66B). If this is the desired effect, fine. If con- sistent timbres are desired, the filter cut-off must move higher as the pitch moves higher so that the filter at- tenuates at the same point relative to any given pitch fundamental. Figure 6.6C shows the VCO and VCF offset to where the filter cut-off is at the fourth har- monic. As a keyboard voltage is applied the pitch and filter cut-off move up in “unison” so that the cut-off is still at the fourth harmonic of any pitch. More Likely than not the fiter’s control voltage attenuator will have to be adjusted in order to track with the oscillator. If you have tried this (and it works), don’t stop with the obvious. Try different attenuation levels and see what you get. If you have the resources, invert or multiply the contrel to the filter. Controlling two ar more VCO's from the keyboard {s also possible. The oscillators can be offset to any de- sired interval and, if correctly tuned in terms of the control, they will track at that interval. With fixed volt- age inputs or hardwired keyboard connections the VCO tracking is no problem. If the control voltage attenua-Figore 6.6. Tracking an oscilator and Sher tors are continuously variable, tuning the correct re- sponse with more than one VCO can be difficult. Let me suggest a method that may make it easier. & Patch all the VCO's into a mixer so they may all bbe heard together or individually. It will help if they all come out of the same speaker. b. Connect the keyboard or any other control to each, ‘VCO in parallel (the same control going to each ‘VCO simultaneously) © On the mixer, turn down all but one VCO and ad- just its control voltage attenuator so that an octave on the keyboard sounds like an octave. When doing this, begin with the lowest available octave on the keyboard; then when satisfied, check the upper oc- taves. It will alo help if the VCO's are all set to a harmonically rich waveshape such as square or sawtooth. Do this for all other VCO's, listening to ‘each one individually. 4. Touch the lowest key on the keyboard (or the key corresponding to 0 volts) and turn up the gain for the second VCO. You will now hear both VCO's, probably set at some absurd interval, so tune them to 2 unison. ‘The rich waveshape will make it easier to hear the beats. £ Touch the octave key and adjust the processing pot on VCO 2 so that it is in tune with the refer- ence. Do not touch the reference VCO unless you are sure it has drifted or been changed. After the Processing pot has been adjusted to produce a uni- son, touch the 0 volt key and readjust the offset of VCO 2, as it probably will have changed a bit. g. Repeat this process for each oscillator. Now the VCO's can be offset to any intervals and should track exactly in tune. Some VCO's have phase locking capabilities which will insure precise tracking, It has been my experience that the use of phase locking to correct intonation really makes a bank of oscillators sound like one VCO as you lose much of the phasing information inherent in so ‘many interesting timbres. However, if this is what you ‘want, by all means use it. Uses of phasing can be very effective with other techniques and if you have to waste phasing by keeping the VCO's in tune, some- thing is probably out of adjustment on the VCO or the keyboard. (See figure 6.73.) Keep in mind that VCO's do not have to track in parallel. Try taking one leg of the control voltage through some various process such as inversion or slew- ing (see figure 6.7b).. Portamento. At times it may be desirable to have a gradual voltage change, as in figure 6.8. This gradual voltage change may be achieved on certain keyboards by using 2 “portamento” control. The word “porta- mento” must be understood in its literal sense (Italian— “to cary”), since the gradual change of voltage need not be applied to frequency. The portamento control setting will determine the time it takes for the voltage to change from one level to another. This rate may be 49Figure 6.7. Tracking wo VOOs pee Figure 6.9. Pressure contro! few milliseconds or sometimes as long as 10 seconds. If two adjacent keys produced the voltages of S and 4 volts respectively, and the portamento control was set at 1 second, the utilization of the keys would pro- duce a I-second rate change from one voltage to the other. If the voltages were controlling an amplifier, the effect would be a 1 second crescendo or dimimu- endo. If your keyboard does not have portamento capa- bilities, you might invest in an integrator. Just route the keyboard voltage through the integrator before at- taching it to an oscillator and you have controllable Portamento. If the integration rate is voltage control- lable, you have all the added features. Fluctuating Controls from a Keyboard. Another op- erating characteristic of many voltage keyboards is that the output voltage may be a function of the man- ual force applied to any individual key. Keyboards con- sisting of capacitance plates will produce voltages pro- portional to the applied finger pressure. This would bbe very useful if the keyboard has a separate ontput for the preset voltages and for voltages produced by finger pressure, since it would be possible to create ‘various phrasings and accents by using the preset volt- age to control frequency and using the finger-pressure voltage to control the amplitude of each frequency (figure 69). With this patching, the individual amp- litudes, along with possible cresendi and diminuendi, would ‘be a function of finger pressure, and pitch would be a function of keyboard voltages. The use of pressure as a viable control requires a ‘it of practice bat itis worth the time spent. Although the input is called pressure on touch keyboards, it is really a result of amount of skin contact. This is a correlative of finger pressure, but accurate perfor- mance requires that we really know what is going on. First, the standard “thumb under” keyboard usually doesn't work. This technique brings the thumbnail in contact with the touch plate and no voltage will be produced. Attach the pressure output to 2 VCF or ‘VCO and try just rocking back and forth om a key. ‘The key doesn't matter, as the voltage is not produced by key selection. With a VCO, see how accurate you ‘can become in playing specific intervals. Process the voltage by attenuating it to workable levels. Some keyboards allow keyboard adjustment of pressure re- ‘sponse so that it can really be tailored to your own touch. If you have the capability of voltage-controlled ‘portamento, control the portamento rate with pressure. In this way, if 2 higher voltage decreases the porta- mento time (you may have to invert the voltage to get this correlation), the firmer you press a new key the faster you arrive at a new pitch. In this manner porta- ‘mento rate can vary from pitch to pitch. The Buchla 219 Kinesthetic Input Port is designed as an interfacing instrument for use between the per- former and « microprocessor, or is able to control ap- propriate analog modules by direct front panel patch- ing. Designed in a touch sensitive keyboard format, it offers still another dynamic input. In addition to the previously described touch sensitivity, it provides posi- tive and negative controls proportional to lateral move- ‘ment of the entire keyboard. It is so mounted that there is about 1/20 of an inch movement of the keyboard on its lateral axis, One can use these voltages, summed with the regular key voltages, to control the pitch of a VCO. In this way each key would produce a dis- ‘crete pitch but it could be bent flat or sharp with lat-eral pressure, The pressure voltage could then be used to control some other dynamic parameter such as a filter. ‘Additional dynamic inputs provided by various keyboard designs include velocity and depth sensi- tivity. In the first case 2 control is produced propor- tional to how fast a key is depressed, and the latter capability produces a voltage proportional to how jar a key is depressed. All of this implies an important point which the musician should always bear in mind, Acoustic events which have been useful as artistic tools are usually complex structures with several on- going variables. If one approves of this direction (and some relevant composers often do not), the more in- putting or interfacing capabilities one has with the instrument the better. ‘A device usually associated with the keyboard is the “ribbon controller” or the “linear controller.” This is a tight band about 2 feet long that will produce a voltage proportional to where it is touched by the per- former. The linear controller is usually equipped with a pot which will determine the total output voltage, similar to the range control on the keyboard. If the total outpat is 5 volts, and it is controlling an oscil- lator, it would be possible to produce a 5 octave glis- sando by sliding one’s finger from one end of the rib- bon to the other. An ascending glissando would be achieved by sliding from left to right, while the oppo- site action would produce a descending glissando. If the total output of the ribbon was set at 1 volt, the same physical action would produce only a 1 octave lissando. This device will produce the same effects 2s the portamento control on the keyboard. The basic ‘differences are that the ribbon often has a smaller range, but the individual characteristics of the voltage sweeps can be more immediately controlled by the performer. An extended voltage range with the same Tength ribbon could result in a loss of performance accuracy. Some ribbon controllers are hardwired to ‘add to the keyboard output, limiting it to association ‘with whatever the keyboard happens to be controlling, Other ribbon controllers have undedicated outputs, and they may be used to control whatever the per- former decides. Timing Pulses. Still another application of the key- board, independent of format, is the production of triggers and gates (both referred to as “timing pulses"). ‘These terms often differ from manufacturer to manu- facturer. As discussed earlier in this text, performing on an instrument often requires a single input to con- tuol a variety of activities. Assuming the keyboard is being used for pitch selection, you might also want ‘other things to happen simultaneously. For example, the timbre and loudness might change according to some pre-programmed design, and at the same time the sound might spin around the room in various pat- terns. These are preprogrammed functions which the performer usually determines previously and accesses in various ways in modules like envelope generators (see page 64), sequencers (page 70), random voltage sources (page 83), etc. These functions then have to be called forth by some command. And it is usually desirable to call forth these commands simultaneous! with other parametric changes which might be con- tolled by a keyboard. ‘These commands are called timing pulses. The terms “trigger” and “gate” are often used interchans ably but there is a distinction in what they do. First of all we should understand that a trigger and gate have manufacturer determined levels and that the same voltage level for each is produced for every key. Key 1 produces the same trigger and gate voltage as ey 30. A trigger is a transient voltage that bas a very fast rise and fall (see figure 6.13A). A trigger is used to cause something to happen, such as incrementing 2 sequencer (page 64), calling up a random voltage, or triggering an envelope generator to begin its attack (see page 70). Trigger voltages will vary from manu- facturer to manufacturer, but they are usually high magnitude voltages. A gate always accompanies a trig- ger and is a sustaining voltage which allows something to continue to happen. The gate voltage may be the same magnitude as the trigger or may be a bit lower (see figure 6.13B). Each time a key is depressed it produces a trigger and a gate at the appropriate out- ‘puts. The trigger is transient and its length camnot be varied. The gate, however, will stay “on” or “high” for as long as the key is depressed. Upon release of ‘the key the gate will immediately turn “off” or “low.” ‘This means that the trigger will activate some func- tion and the gate will allow this function to continue until the key is released. In some cases an instrument is so designed that a clear user distinction between trigger and gate is nec- essary. These systems will usually have separate trigger ‘and gate outputs and inputs. Other systems tie triggers and gates together internally, and the responding mod- ‘ule must distinguish between the two types of infor- mation and use them accordingly. More will be said about triggers and gates when discussing programmed control sources and other types of timing pulse sources. aan a ee ea) Figure 6.10. Triggers and gates 1At this point the reader should realize that most kinds of timing pulses are not controls in the sense that they ‘can be adjusted in voltage to cause different magni- tudes of response. This does not mean that you should indiscriminately plug triggers and gates into just any input. If there is 2 danger in confusing timing pulses with controls, it will so state in the user's manual, or there will be @ clear distinction with the type or color of plug used. Joysticks One of the cybernetic means of voltage production is the “joystick.” A joystick consists of two or more po- tential voltage sources which are simultaneously con- trolled by 2 vertically-positioned lever. The two volt- ages are physically controlled at an angle of 90° to each other on an X-¥ axis. Movements of the joy- stick from right to left would produce a relative change in voltage X, while movements of the stick toward or away from the body would produce relative changes in voltage Y. The advantage of the joystick is that it provides simultaneous but independent control of any ‘two voltage-controllable parameters. If voltage X is being used to control amplitude and voltage Y is being - used to control frequency, an endless number of amp- Iitude-frequency relationships can be realized. Move- ment of the stick at ¢ 90° angle in relation to the body ‘will vary the frequency independent of amplitude, and a right-left movement of the stick will vary the amp- litude independent of frequency. Moving the stick at ‘an ascending angle would result in an abrupt change in frequency with a relatively slow change in ampli ‘tude (figure 611A). The opposite effect (abrupt ‘change in amplitude with a relatively slow change in frequency) could be produced by moving the stick at a right angle (figure 6.118). A circular rotation of the joystick would produce 2 continually varying change in the two parameters in constant opposite re- Istionships (figure 6.11C). ‘Some joystick shafts are mounted on a vertically positioned pot to provide a Z axis voltage. The voltage is then proportional to the up-down movement of the stick. This Z axis voltage, of course, can be attached to any desired parameter. This may be combined with ‘the previous patch to control a filter cut-off as illus- ‘trated in figure 611D. 52 Figare 6.11, JoysticksEven more useful, and more expensive, joystick de- signs provide a fourth voltage produced by a rotary handle. As the stick is being turned, the handle or en- tire stick is capable of independent rotation, thereby producing a fourth dynamically variable control. Such a four-axis joystick can provide control of four inde- pendent parameters with a single hand. Keep in mind that this is not easy and requires practice—but what instrument capable of significant structures doesn't? Envelope Detectors Many performance situations are enhanced and often simplified by interfacing external acoustic or elec- tronic signals with electronic instruments. The circuits which provide for direct control from external instra- ments are generally known as Envelope Detectors (ot Envelope Followers) and Pitch-to-Voliage concerters (PVC). Basically, the envelope detector produces @ control voltage proportional to the amplitude or in- ipient loudness of a sound, and the PVC produces @ control voltage proportional to the fundamental fre- quency or pitch of a sound. The following section will discuss both instruments in some detail. ‘An “envelope” is commonly known as the loudness curve of a sound. Figure 6.12 illustrates some gen- eralized envelopes from various familiar acoustic events. It must be pointed out that envelopes from ‘acoustic sound sources are much more complex than these illustrations represent. If you have access to an oseillascope, observe the voltage behavior of various kinds of acoustic waveshapes. It must also be stressed, that these envelopes are a function of the loudness of the instrument, not the pitch. There is a certain amount of correlation between loudness and pitch, but in order to keep things understandable, don't be concerned with the correlation at this time. Most en- vvelope detectors are designed to accept a line or high level signal. Details of this are given in chapter 14. In ‘this present discussion it means that most acoustic sources such as electric guitars, microphones, some electronic pianos, etc., will have to be pre-amplified, as their signal is too low to be effectively detected by the circuit. Therefore, in trying out some of the sug- gested patches, have your studio technician adjust Your instrument up to line level (if you don't know how to do it), or tum ahead and read chapter 14, age 295) ‘Working from a simple patch will facilitate the understanding of an envelope detector. If you have the resources, patch together the instrument dllus- trated in figure 6.13. After the audio signal is brought up to the correct level with a pre-amp, it is patched to the input of the ED. Its output is then patched to the control input of a VCO. It will be easier to hear ‘what is happening if the waveshape has minimal har- ‘monic content (sine or triangle). The ED will usu- —_ DAE HA0) = eens ston ie mas ally have control marked “sensitivity” or “response.” It may be a pot or switch which determines the pro- portion of output control voltage to the amplitude of the input signal. With this control at minimum, a “mf” sound will produce little or no control voltage. Play some different articulations on the instrument and ex- periment with how this control affects the VCO. If the sensitivity is at maximum, a moderately loud sound will produce 2 control that drives up the pitch: of the VCO a proportional amount. As the sensitivity is Jowered, the same sound will not have as much effect ‘on the pitch of the VCO, since the output voltage has een reduced. The sensitivity control is, in fact, an attenuator which allows you to tailor the ED's response to your own needs. This same thing can be accom- plished by lowering the control voltage attenuator on the VCO, but leave it at maximum for now so the various effects of the ED can be beard. Now patch up the instrument as in figure 6.14. In this case the output of the ED is patched to the con- ‘vol input of a VCA. Trying the same experiments, you will find that the loudness and articulation of the in- ppat signal determines the loudness and articulation of the signal from the VCA. Most ED's will have a sec ‘ond control marked “decay,” “lag,” or even “slew.” Figure 6.14. Gain conto! wih an envelope detectoraaa aoe Ley Figure 6.15. Envelope decay time Figure 6.16. Triggers and gates from an envelope detector The control determines how closely the circuit will follow or reconstruct the input signal as shown in fig- ure 6.15. With this control at maximum, 2 staccato sound produces 2 control envelope with a somewhat longer decay. If the ED has a pot specifically marked “decay,” this will usually refer to a downward slew, and sharp attacks will be maintained. If other terms are used, it may mean that the entire envelope is gen- erally smoothed out. Listen to the response of the patch ‘and the function will be evident. ‘An ED may respond linearly or exponentially, and some have switch or jack selection between the two types of detector slopes. In the linear mode, the ED control voltage changes in direct pro- ED is in exponential going through the VCA will have the same envelope as the signal going into the ED. If ‘A second output produced by most ED's is a trig- ger and/or gate voltage. Most ED's contain a circuit called a Schmitt Trigger. The Schmitt trigger examines the input or output envelope and creates a trigger or gate (a timing pulse) when it reaches pre-set level known as the threshold. The threshold is determined by @ comparator circuit and in some designs this term muy be used. Figure 6.16 illustrates trigger and gate generation with different threshold values. The timing pulse output may respond either of two ways. In one mode of operation the gate will stay high for as long as the envelope is above the threshold level; this is the case with a true comparator. In certain instances the gate length may be determined by front panel con- trols. Some circuits will not generate a timing pulse until the envelope falls below the threshold level and a new attack is initiated Other designs make it pos- cy sible to generate new timing pulses with any amount of transient activity above the established threshold. Different instruments solve this problem in various ‘ways ranging from simple to comples—check your in- strument manual. The patches in figure 6.17 suggests some initial applications of envelope detectors. Except in one case, the use of externally generated triggers will be left until later when they can be discussed in terms of de- vvices that can effectively make use of them. In each case the patches can be used with any properly pre~ amplified instrument—voice, guitars, barking dogs, etc. Figure 6.174 illustrates envelope transfer between two instruments. Each instrument is taken in parallel to a VCA and an ED. With the VCA offsets at zero, the signal of the respective instrument will not be audible (through the VCA) until it is opened up by a con- trol voltage. The control used to drive the VCA is the control voltage produced by the other instrument. In- strument 1 will not be heard unless instrument 2 is playing, and vice versa. If 1 is playing sustained notes, and 2 is playing a series of staccato notes, the sustained note of 1 will actually be heard as a series of staccato articulations. If the instruments can be directly heard, such as with voice, this type of patch is made more effective by removing the instruments from the imme- diate enviroument, or by tape recording the events. You may wish either to tape record the instruments before they are processed or to tape the final proc- cessed result In this way only the processed events will be heard. Previous taping has the advantage of elimi- nating the pre-amplifiers, as tape outputs are already up to line level. Figure 6.17B is a variation on the previous patch; in this case the loudness of instrument 2 controls the inal timbre of instrument 1 by using a detected en- velope to control 2 filter. Figure 617 uses an ex- ternal envelope to control the loudness of an electronic signal. The oscillator may be generating a series of events from a keyboard or any other controller but is not heard unless the acoustic instrument is generat- ing a control for the VCA. Figure 6.17D uses the‘opposite logic, in that the control envelope is inverted. If the instrument is not playing, the processed control is at maximum. As the instrument produces an en- velope, the control voltage is proportionally attenu- ated. Hence the VCO is heard only while the acoustic instrument is not playing. As soon as the instrument begins to play, the inverted envelope tums down the YVCA gain proportionally. If this sort of “if” “and,” = ro) Pst “neither,” “nor” logic interests the reader, tum to the score for Robert Ashley's String Quartet Describing the Motions of Large Real Bodies at the end of this chapter (page 92). Pitch-to-Voltage Converters If the process of envelope detection is understood, the process of pitch-to-voltage conversion will be relatively simple to grasp but not always so easy to accomplish A pitch-to-voltage converter (PVC) is exactly what the term implies: it detects the fundamental of a note being played and converts it to an analogous contro! voltage. For example, suppose a flute were to play a series of arpeggios, as in example 6.18. The PVC would extract the fundamental of each note and convert it to corresponding control voltages. Assuming the 1 volt per octave standard, the control voltages could then be patched to a VCO and, if offset to the correct pitch, would reproduce the pitches played by the flute. The VCO could also be offset at a different pitch and would track with the flute at that interval. With cer- tain instruments, PVC’s present some problems. Most acoustic signals have a transient harmonic state, ¢s- pecially on the attack. What this amounts to is that the fundamental of a pitch is not always the strongest for most apparent component of a sound’s spectrum. ‘When this is the case, it can be very difficult for the PVC to perceive what the fundamental frequency ac- ‘ually is. Depending on the nature of the transients, the PVC may switch back and forth between harmon- ics, or even noise, until the waveshape has settled down enough to present clearly the fundamental for conversion. This, of course, takes too much time for adequate musical applications. There are several meth- ods a manufacturer may use to solve these problems, but this also takes time for the circuitry to do the job. ‘Current surveys of PVC users indicate various degrees of satisfaction and dissatisfaction, most of them relat- ing to this type of problem. We can assume, however, that technology will perfect the PVC, and the follow- {ng discussion will be based upon this assumption. Figure 6.18. Pitch-to-votage conversion‘Since most of the PVC applications are to make an electronic sound source track at set intervals, it is con- sistent that envelope duplication would also be of use. Most PVCs have built-in envelope detectors and ‘Schmitt triggers. Beyond this, one can think of the PVC as just an alternate to the keyboard. Miscellaneous Controllers ‘The EML Poly-Box is a combination one-octave key- ‘board and PVC. This is an accessory designed to pro- vvide limited polyphonic capabilities to an instrument. ‘An output of any VCO is patched to the Poly-Box Since in this case the PVC is given a relatively steady- state electronically generated signal, the conversion is ‘uncomplicated and very accurate. The converted volt- age then controls an internal oscillator in parallel with the extemal source and offset as desired. By mean of an electronic organ technique of top-octave divi sion, the C-to-C chromatic polyphonic keyboard is tuned to the pitch of the master oscillator. If the master oscillator is tuned to “A,” a C major scale will sound like an A major scale. The term “polyphonic” ‘must really be qualified in this case as all available 18 pitches come from a single output and are subject to the same enveloping, filtering, etc. (see fig. 6.19). The percussionist is accommodated by several kinds of percussion interfaces and percussion synthesizers. Figure 6.20 shows Star Instruments’ Synare 2 Percus- sion Synthesizer. This is the basic synthesizer voice ‘with a monophonic tunable keyboard taking the form of rabber pads much like practice pads. The padboard is purposefully laid out so as not to resemble an organ format. Each pad may be assigned a pitch and oc- tave, so that pitch pattems can be set up to be ac- cessed in the most convenient manner. Another switch is used to preset the pads to @ chromatic sequence from C to B. The pads also generate triggers for acti- vating other functions built into the instrament, The Symare 2 is designed as an independent instrument, and therefore its control voltages and triggers are not available to be interfaced with other synthesizers. The Moog Percussion Controller is yet another type of source of control for the drummer which reacts in much ‘the same manner as a touch keyboard. Its output volt- age is determined by how hard the drum is hit, and ‘the voltage can be patched to any usable voltage con- | Figure 6.20. Snare 2 Diagram (From the ‘Synare 2 Percussion Sythesizor/Diptal ‘Sequencer Manwal. Courtesy of Star Inston, ln. Used by permieaion)ied Aa poon ‘044 evo|nauo Jo Keonnop) g}oe¥eD wIndueg pu Wounsten HOOUKT euL “LE eunBLd quewnaysuj uoopATtrolled parameter. The percussionist may control the pitch of 2 VCO, loudness by means of a VCA, etc. ‘Two manually set controls determine the drum’s sensi- tivity (how hard one has to hit to get a response) and scale or output voltage range. These voltages, of course, may then be further altered by whatever proc- essing modules are needed and available. Wind players cannot be neglected; they are accom- modated by instruments like the Lyricon (figure 6.21). Designed in much the same manner as a clarinet with fingerings based on the classic Boehm system, the in- strument gives direct control over pitch, loudness, ar- ticulation, and timbre. iological Controls, One of the more interesting areas of performer input is the sensing and transduction of direct biological func- tions and reactions. Keep in mind that anything which can be measured can be turned into a voltage, and if a voltage i compatible with an instrament, practic. ally anything is fair game. Various types of brainwave activity, skin temperature, muscle tension, periods be tween different modes of physical activities, etc., can be used and have been used as real-time performance input for electronic instruments. One of the earliest works of this nature is Alvin Lucier’s Music for Solo Performer. Here the perform ers alpha activity* is sensed through special ampli fiers and amplified to be used as sub-audio and low frequency activators for various resonating objects. Since this is such a landmark work in the development of input concepts, the score is reproduced here as a matter of documentation. A newer version of Music for Solo Performer was developed in 1975 with the assistance of Nicholas Collins (figure 6.22). In this version the “comparator” and “retriggerable mono- stable” generate timing pulses when the alpha voltage 2 An alpba wave is 2 very low magnitede voltage of from to 1 He generated by certain brain activity. ~ ya Figure 6.22. Avia Lucie’s Masic for Soto Performer (Used by permission ofthe composer)exceeds a preset threshold—really an envelope detec- tor and Schmitt trigger. The pulses are then used to tix and route the low frequencies to the resonating objects, This version was performed at the Paula Cooper gallery in May of 1975—the tenth anniversary of the piece. It should be interesting to be at its Silver anniversary in 1990 to hear/see what new orchestra- tions technology will contribute! Music for Solo Performer (1965) ‘The idea for Music for Solo Performer (1965) came cout of a series of conversations | had in 1964 with physi cist Edmond Dewan of the Air Force Cambridge Re- ‘search Laboratory in Bedtord, Massachusetts. At that time, Dewan was engaged in’ brainwave research par- ticularly as it pertained to fying: it was believed that cenain periodic visual rhythms of slow propelior speeds were locking onto corresponding brainwave frequencies of aircraft pilots, causing dizziness, blackouts, and epi= leptic fits. Dewan, an accomplished amateur organist, was eager to share his ideas and equipment with any ‘composer interested in exploring this hitherto uncharted ‘egion. Inspired by the imagery and technology of elec- ‘woencephalography. | immediately set to work to dis- ‘cover all | could about alpha. Working ong hours alone in the Brandeis University Electronic Music Studio with Dewan's equipment (two Tektronix Type 122 preamplifiers in series, one Model ‘330M Kronhite Bandpass Filter, which had been set for 2 range of from 9 to 15 Hz, one integrating threshold ‘Switch, electrodes. appropriate connectors etc.) plus the studio's conventional equipment, 1 learned to produce alpha fairy consistent. | found that success could be ‘attained by setting the gain on the audio amplifier to = Point just below oscillation so that even a relatively weak ‘pha signal would come through. Often, | could pro- ‘duce alpha only in shor bursts; it took precisely the ght physical and psychological conditions to sustain it in longer phrases. | did not attempt any experiments in k aS such but was aware of the reinforce: ment of my own alpha-producing ability white monitoring in reabtime the sounds that came out of the studio 1oud- ‘Speakers. | observed that over long periods of time, for ‘example while recording alpha for storage material for tse in performances, or when tired, related or slightly Bored, the alpha would tend to drift somewhat down ‘ward and sett. From the beginning, | was determined to make a five Performance work despite the delicate uncertainty of the uipment, ciffcult to handle even under controled fabo- Fatory conditions. I realized the value of the EEG situa ‘Son as a theatre element and knew from experience that live Sounds are more interesting than taped ones. 1 wae aso touched by the image of the immobile if not para lyzed human being who, by merely changing states of ‘sual attention, can activate a large configuration of com- ‘uncation equipment with what appears to be power ‘rom spiritual realm. 1 found the alpha's quiet thunder extremely beautiful and, instead of spoiling it by proc essing, chose to use it as an active force in the same way One uses the power of a river. twsed the alpha to resonate a large battery of per- ‘ession instruments including cymbals, gongs, bass drums, timpani, and other resonant found objects. In most cases, ‘twas necessary physically to couple the loudspeaker to the instrument, although in the case of highly resonant bass drums and timpani, the loudspeaker could be an inch or s0 away. Placing loudspeakers in trash cans or cardboard boxes worked extremely well, a= did using ‘cheap small speakers face down on snare drums or taped against windows. | learned that by varying both short bursts and longer sustained phrases of alpha plus mak- ing musical decisions as to placement of loudspeakers, choice of resonant instruments or objects, volume. con. ‘ol, channelling and mixing, | was able to get a wide varity of sonortes as well as retain the natural phys. ical quality that seemed asked for by the sound source itselt ‘In conjunction with the threshold switch, 1 used the alpha as 2 control signal to operate a stereo tape re- Corder upon which was stored transposed versions of re-recorded alpha accelerated up to five times. These Figher phantoms relieved the sameness of the iow: frequency originals and were used Doth by themselves and to impart contrasting resonances to whatever instru- ments they were coupled to. My original intention was to develop the idea of contol to include more sophisticated systems of lights, alarms, television sets, radios, whole ‘environmen ‘Aithough an assistant is usually needed to operate the preamplifier controls, | éid pertorm Music for Solo Per- former (1965) by myset on the “Visions of the Present” festival in Stockholm in 1966. | succeeded in producing alpha by leting my hands operate the amplifier controls 235 randomly 25 possible to avoid visualization caused by decision-making with reference to channeling and place- ment of loudspeakers. | have always wanted to have situation in which the alpha could perform all the con- trol functions by means of a code; for example, a cern number of bursts of certain durations could tigger cer- ‘ain mixtures of channels, | have not pursued further development of brainwaves 125 a musical Fesource in order to let myself move. on to other works involving other ideas including echolo~ ‘cation, underwater sound, resonant characteristics of rooms, and the alteration of vocal identities. | am happy to see that many other composers are using alpha in creative and imaginative ways ‘music for Solo Performer (1965) is dedicated to John Gage who assisted me in the rst pertormance on May 5, 41965 at the Rose Art Museum, Brandeis University and to whom | am grateful for encouragement greatly appre- ciated at that time as well as now. ‘ Alvin Lucier September 8, 1971 Middletown, Connecticut More technically complex and differing applica- tions of bio-control were pioneered and continue to bbe developed by composer/performer David Rosen- ‘boom. Much of his early work is documented in Bio- feedback ond the Arts: Results of Early Experiments Rosenboom’s Portable Gold and Philosophers’ Stones (Music from Brains in Fours) uses control voltages derived from galvanic skin response and body temper- ature, These voltages are used to control a bank of 3 Published by ARC. Publication, P.O. Box 3044, Van- ceouver, B.C., Canada, VOB 3X5.PORTABLE GOLD AND PHILOSOPHERS’ STONES (Music From Brains In Fours) [crocs ane arma monn caves ae mache moter ne ea wees SSeS ee teeta Semon cn cane wns pat ees mt Stoecey san be ety son» voren too kee Sap eo =e David Rosenboom Sir inten romeo set arma naonssP m# SSS Eee Figure 6.23. Davie Rosenboom's Portable Gold and Philosopher's Stones (Masic om Brains in Fours) frequency dividers, and the controls are further proc- essed to control filters. The score is reprinted here and the patch for performance is illustrated in figure 6.23. Richard Teitelbaum's In Tune is still another ap- proach to bio-control, utilizing visual display and amp- Jified throat and beart sounds (figure 6.24). Two fil- ters in series allow only brainwave activity between 8 and 12 Hz to be conveyed to controls by an envelope follower (detector). The detected voltages control two ‘VCO's and the trigger output allows the loudness and timbre to be shaped by an envelope generator. This work was composed specifically for a Moog. instra- ment. 1 am generally trying to avoid aesthetic discussions, but biofeedback suggests some relevant aesthetic ques- tions and answers. I would strongly recommend that the reader read Rosenboom’s Biofeedback and the Arts.ot amet oO 7 Programmed Control Voltage Sources ‘So far the discussion of controls has been restricted to direct physical input from the player. In spite of their seeming complexity and range of possibilities they still are not totally adequate for the unleashed minds of musicians. “Playing” is a very complex matter— even on the most simple instrament. As yet there does not exist a single circuit that can replicate the amount of information which a traditional player transmits to his instrument. At the same time the nature of a synthesizer requires that we have the ability to or- ganize and temper parametric responses to specific needs. The next step is to have the possibility of pre- ‘set responses or voltages which can be accessed by ‘the performer as needed. The following sections will ‘explore circuits and modules which perform preset and. Strictly speaking, a function is the activity of a “thing,” and the “things” described here have special- ized and/or general functions. In terms of our appli- cations, the activities or functions are voltage patterns. Voltage functions must have a beginning, ongoing state, ‘and end, and the performer must have some means of defining these time points. Timing pulses, triggers, and fates are used to initiate, sustain and terminate func- tions. ‘The voltage outputs of programmed voltage sources. such as envelope generators (EG), sequencers, ran- dom voltage sources, sample and hold modules (S/H), ete., are usually associated or correlated with some ‘other parametric activity. For this reason most kin- ‘esthetic inputs, like keyboards and envelope detectors, have associated timing pulse outputs. When 2 per- former calls up a pitch voltage for a VCO, a timing alse can simultaneously cue another voltage to do Figure 6.24. Richard Teteaun's i Tune (From Bicteodback and the Arts—resuts of early ‘experiments. Ed. by David Rosenboor, Aesthetic Research Centre of Canada, Used by permission) whatever it is supposed to do-spin a sound around the room, play @ programmed arpegsio, sweep a VCA, or filter, ete Review of Timing Pulses: Activation of Events ‘As @ quick review, let's look again at how manu- ally activated timing cues and Schmitt triggers be- have. Figure 625 illustrates these relationships. As soon as the keyboard is touched, the trigger voltage is generated. A trigger is usually a high voltage in the form of a transient spike lasting about 100 microsec- fonds. On the Moog modular systems, there is a dis- tinction between “switch-closing” or switch trigger (S-TRIG) and a fast rising positive pulse known as the “voltage trigger” (V-TRIG). They accomplish the ssame thing, so that the user only has to be concerned, ‘with which modules accept “S” or “V" triggers. As long as the key is depressed (or touch plate is being touched) the gate voltage is present. This gate allows the various function generators to continue with their respective activities. As soon as the key is released, the gate is switched off and the activity will stop, or begin to stop. ‘Envelope detectors have similar rationale, that is, the generation of a trigger as soon as the input ex- ceeds a set threshold, and the gate will remain on as _ =a t ‘ Figure 6.25. cstes nS meet ee oee renewsFigure 6.26. Gate and tigger functions Jong as the input voltage is above the threshold. As soon as the input voltage is below the threshold, the gate will go off. The reason for the distinction between ‘gate and trigeer information is that it allows one in- pput device such as a key closure to separately activate and sustain different events. A Key closure may acti- vate a repetitive pattern of voltages and simultane- ously call up a random voltage pattern. If the voltage pattern is to continue while a new random voltage is called up, the gate information would maintain the patter’s activity and the trigger would call up the ran- dom voltage—all by just touching another key. It could ‘even be done by very rapidly releasing or touching the same key, as long as it is done fast enough so that the interruption in the gate voltage is not apparent (figure 626). ‘Some srstems such as the Buchla instruments elimi- nate the necessity of separate trigger and gate outputs and inputs. The gate and trigger information are set at two different voltage levels and may be transmitted simultaneously with a single patcheord as illustrated in figure 67. With this approach the function gen- erators can be told, usually by means of switching. whether or not to respond to the gate information. Timing Pulse Generators In many cases, ongoing timing pulses may be called for which are independent of any manual or detecting input. This would be the case when using timing pulses for generating different levels of tempo and various repetitive rhythmic activities. Modules called Pulsers and/or Low Frequency Oscillators (LFO) are gen- erally used for this task. Most smaller performance in- struments such as the ARP Odyssey, Roland System 100, ete., will contain an oscillator dedicated to a fre- quency range of about .2 to 20 Hz. These sub-audio waveshapes are usually used for different types of modulation (see chapter 7), but also may be used for the generation of triggers. The way this is accom- plished is that the squarewave output is fed to 2 62 ttitil | = iam 2D. nto | yo Figure 627. Gate and trigger on one line ‘Schmitt trigger and corresponding triggers are gener- ated. Some function generators will not require a trig- ‘ger as such and will trigger with the leading edge of the squarewave. Unless the oscillator has some means of duty-cycle control or pulse-width modulation, there will be no way to vary the gate time. With the Schmitt trigger technique, any sub-audio waveform can be tumed into a trigger, and non-switching waveshapes such as sine, triangle, etc, can produce variable gate times by adjusting the threshold. On smaller perfor- ‘mance instruments, the LFO is usually not voltage con- trollable. On larger instruments, oscillators with sub- audio ranges can be fed to Schmitt triggers, and volt- age control of the VCO results in voltage controlled ‘trigger rates. ‘LFO’s dedicated to the generation of timing pulses are referred to as Pulsers or Timing Pulse Generators (TPG). Although no longer in production, the Buchla Series Timing Pulse Generator is a useful approach to timing pulse production and is still in operation in studios. The Buchla 100 Series instruments did not sum internal offsets with controls, so that the period of voltage controlled. At a pulse period pulse width indicates that the pulse voltage will be ‘high for the entire period. A 50% pulse width (seeFigure 6.28. Pulse lngth figure 6.28) indicates that the pulse is high for only ‘one half a second. Pulse width in no way affects the nite at which pulses are being generated. Using voltage control of pulse width can offer some interesting cor- relations. These applications will be covered when dis- ‘cussing envelope generators (see page 64). The Buchla 100 Series Timing Pulse Generator has two alternat- ing outputs, and these output pulses are alternately assigned to each output, allowing pulse division by 2 or alternate triggering of functions. With the mode switch in the “repetitive” position, the TPG will pro- duce timing pulses at 2 period proportional to the ‘manual offset or an external control voltage. Certain oscillators can be used to perform most of the above tasks. What is needed is a low frequency squarewave oscillator with pulse width modulation eo 8 4 & 05 SS few e-é soo Flours 6.29. Eu 2200 VCO (Courtesy Eu Systems, lnc. Used by emission) ‘and gate inputs. The Eu 2200 has these specifications and is pictured in figure 6.29. The Eu 2200 VCO os- cillates as slow as 03 Hz. That means an incipient pulse period of 33 seconds—certainly long enough for most applications. When the “gate” input receives 2 volt- age of about 25 volts or more (from another gate or any other controller), the oscillation will stop, and as soon as the gate voltage drops below this threshold, the oscillator will start again. This is close enough to a start-stop mode to be workable for the knowledge- able user. The pulse width offset and pulse width modulation input determine the duty-cycle from 0 to 100%. This is comparable to width or gate time. On some oscillators the pulse width can virtually be 0 or 100%, With 0¢ pulse width all you have is 0 volts, and 100% pulse width means non-flactuating DC—hence no ‘oscillation in either case. If a pulse width variable os- cillator appears not to be working, check out the pulse width controls. Whether the oscillator has to be patched through an ED to give the correct voltage levels will vary from instrument to instrument. Usu- ally, as long as an incipient trigger and gate voltage are above the required levels, everything should work. OF course not all VCO's have all these features; but a little thought and investigation of the specs will usu- ally bring latent possibilities to light. Electronic Switches Electronic, Analogue or Sequential Switches are not limited to timing pulse information, but they will be discussed briefly so that various function generators ‘can be presented in a variety of applications. Figure 6.30 illustrates a logic module capable of processingboth analogue (control) and digital (timing) voltages. For example, imagine that three different keyboard players were producing independent controls for 2 VCO, and the three kevboard voltages were taken to the three “signal” inputs of the switch. The voltage which is present at the output depends on the condition of the input pulse information. If a pulse input is act vated, the associated control voltage input is con- nected to the output. Thus a player may select which of the ongoing controls will be sent to the VCO by routing pulses to the desired input or by manually activating a front panel switch. The switch also has corresponding trigger outputs which are activated whenever a corresponding input signal is selected. In- struments such as the Moog 962 Sequential Switch can switch control or audio voltages. and the switeh- ing may be pulse or manually activated in any chosen order. They may be sequentially switched (1-2-3-1-2 3 or 1.2:1-2, etc.) by applying timing voltages to the “switching input.” Timing Pulse Delays A timing pulse is binary information—it is either there or not there—and just about all one can do is to de- fine conditions in which they will or will not be gen- erated or allowed to have some effect. There is, how- ever, one other possible timing pulse process, and this is to delay it. Some function generators have built-in trigger delays. With a pot or control voltage, the user may define a delay time. Once this delay time has been defined, 2 pulse can be applied and the function will not begin until after that delay time has been com- pleted. Figure 6.31 shows the Moog $11-A Dual Trig- ger Delay. In this design a “coupling mode” switch determines the routing of the input and output pulses. In the “parallel” position the trigger inputs of both delays are connected. A’ single pulse at either input will activate both delay circuits, each still having the possibility of independent delay time for each circuit. ‘When each circuit's delay time is complete, a pulse will be generated at its respective output. When the switch is in the “series” position, the trigger output of the upper circuit is connected to the input of the lower circuit. When the delay eycle of the upper cir- cuit is complete, it will automatically trigger the lower circuit, In the “off” position both circuits are com- pletely independent. In addition to dedicated trigger delay modules, there are a few “tricks” that one may vse to make other function generators delay and di- vide timing pulses. These techniques will be described in connection with the appropriate module. ‘The preceding paragraphs by no means present ex- tensive information on timing pulse logic. But there 4s enough information to “trigger” the question, “so what?” The rest of this chapter will be dedicated to a survey of most types of function generators. Having 64 aati Figure 6.31. Moog 911A Trigger Delay ‘a basic understanding of timing pulse logic at this ‘point makes it possible to explore the many applica tions of control voltage generation without confusing the role of a timing voltage and a control voltage. A timing pulse can initiate, sustain, or terminate a func- tion; @ function is a control that defines « parametric response. Voltage Sources Envelope Generators. The most common function gen- erator is known under various names, such as En- ‘velope Generator (EG), Attack Generator (AG), ‘Transient Generator (TG), Contour Generator (CG), Function Generator (FG), AR, ADSR, and probably some more I have omitted. Such a divergence in terms has to do with their assumed application by the mann- facturer or with attempts to describe circuit opera- tion. Each term has come under semantic attack by the competitors with both significant and insignificant arguments. Rather than contribute to the situation of 2 confused terminology, I will arbitrarily use the gen- eral term Encelope Generator (EG), simply because this is universally understood.‘The EG, upon receiving a timing pulse, generates, slti-staged evolutionary voltage. The most simple EG 4 the “attack-decay” genre (figure 632). Upon re ceipt of a trigger, the module will produce a voltage which rises to its maximum voltage output (the attack), then falls back to 0 volts (the decay mode). The amount of time the voltage spends in the attack and decay is determined by front panel offsets, and may range from a few milliseconds to 30 seconds or longer. depending on the manufacturer (although the range of attack time on an EG is usually shorter than avail- able decay time). ‘The most common application of EG's is illustrated in figure 633. A keyboard is controlling 2 VCO which is patched through a VCA. With the VCA offset at 0. some control is required to raise the gain so the os- cilator may be heard. This may be an unnecessary statement, but it would be convenient if the gain was tumed up and down simultaneous with pitch choice by the keyboard! (This, however, also may not be the case.) Assuming that this is what is to happen, simply take a pulse issued by the keyboard to the EG and patch the EG control to the VCA. Now every time a key is depressed for pitch selection, a simultaneous trigger tells the EG to produce its voltage function, and we hear the loudness of the sound evolve in pro- portion to the control voltage. Keep in mind that the envelope voltage is summed with the VCA’s internal offset voltage, so for the signal to decay to silence the ‘offset must be at 0. Referring to figure 6.32, it was shown that these envelopes may have exponential or linear slopes— either set be the designer or user selectable. If the VCA has a linear response. it may be best to use ex- ‘ponential slope, and vice versa. If there are no switches or specs about this on your instrument. don't worry about it. Figure 6.344 shows a patch variation that reverses the control routing. Here the keyboard determines the loudness and the EG causes the pitch of the VCO torise and fall with each trigger. Figure 6.34B is a fur- ther variation which inverts the control envelope. In this case the VCO is sonic until a trigger is received. The rising and falling envelope is then inverted to @ falling and rising envelope to effect a diminuendo and ‘crescendo. In figure 634C an envelope is quantized before being patched to 2 VCO and also patched in its normal state to a VCA. The result is a correlation ‘between @ scale or arpeggio (depending on quantiza- tion points) and loudness. Most AD envelope gener- ators will have a “duration” control. The setting of 2 ‘pot will determine how long the envelope will remain at its peak level before the decay cycle begins. This setting would define the sustain qualities of 2 sound but is not to be confused with the “sustain” level of an ADSR (see page 67). Figure 6.35 shows the “Attack-Release” or “AR” type of EG. The difference between this and the “AD” is that the AR is not allowed to complete its attack if the gate is not present, or to begin its decay cycle until the gate voltage is low. If the timing pulses come from a keyboard, this means that the envelope will sus- tain at its maximum voltage until the key is released, at which point the programmed decay cycle begins. ‘The Buchla 100 and 200 Series Attack Generators and the 200 Series Quad Function Generator have front panel switches which tell the circuit whether or not to respond to the gate or sustain information, although with this module the attack will not be interrupted. In the “sustained” mode, the envelope will not decay until the gate is released. In the “transient” mode, the sustain time is preprogrammed and will not depend on gate duration, On most designs the sustain time will continue to evolve even if the AR is being held in sustain by a gate. When the gate is released and the sustain time has gone past its setting, the envelope will begin to decay. Extremely short duration sounds and interesting “clicks” are almost impossible to achieve if an EG is reading the gate information—es- ‘pecially if from a keyboard. ‘Examine the patch diagram in figure 6.36. A key- board is controlling a VCO whose output is split into two parallel VCA’s. Each controlling EG is triggered by the keyboard pulse. EG 1 is in the transient mode with a rapid attack and decay (about .02 millisec- onds). EG 2 is in the sustain mode, so that its output voltage will stay high as long as key is depressed. The signal from VCA 1 will be short clicks of sounds with the pitch identity depending on the frequency. Low frequencies have to be sonic longer than high frequencies, as it takes a crtain number of “cycles” before our ear registers a definite pitch. This will probably be more effective with a rich waveshape. Since EG 1 does not register gate information, the en- velope will always be the same length, independent of gate time or how long the key is depressed. EG 2, however, will sustain until the gate is released by re- leasing the key. The final mixed result is a “click” at the beginning of the sound, accompanied by a sus- tained sound of the same pitch. You should be able to hold down any key and effect the clicks by touch- ing other keys (without releasing the depressed key). ‘Remember that the pitch will either be the lowest or last key depressed, depending on keyboard design. On some instruments, gate and trigger information are carried by the same patchcord, and with other in- struments will have to be separately patched in Figure 6.36. Paral envelopesRemember that any suggestion appearing ct any point in this text can usually be combined with other information provided previously. For a text to docu- ment and notate every possible patch would be im- possible. It is recommended here that the reader begin «@ cotalogue of cross-referenced patches. Whenever sew information pertaining to your resources is pre- sented, add it to the patch book and turn back through the older patches to see how it can be incorporated. ‘The most common type of EG is the ADSR, sach as the Eu 2350 Transient Generator pictured in figure 6.374. With acoustically produced events, the inertia required initially to get a string, reed, lips, ete,, mov- jing contains much more energy than what is required to sustain the event. When driving a car the gas pedal ‘is held down until the desired speed is reached and, at least on flat land, the pedal can be released a bit and the speed will be maintained. This “attack” tran- sient results in a lot of helter-skelter with fluctuating harmonic and non-harmonic content and noise. The ADSR is an attempt to simulate some of this activity with a high magnitude attack transient. When trig- gered by an external pulse, the envelope generator will produce a voltage which rises to a maximum DC level with a rise time as fast as 1 millisecond or as slow as 10 seconds, depending on the design. This is the “A” or Attack stage. As soon as the voltage level has reached maximum, it begins an “initial decay” ‘whichis also manually set to last between 1 millisecond and 10 seconds. During this time the initial decay volt- age approaches 2 manually set “sustain” level that may range from zero DC to the system's maximum level. This sustain level will be maintained until the gate voltage is released. The sustain control is not a tem- oral device like the rise time and initial decay pots, Dat is used to determine the amount or magnitude of the sustain voltage. The time period of the sustain voltage is dependent on the length of the applied trig- ger pulse. After the trigger voltage is released, there ‘is a “final decay” or “release,” “R,” which is also man- ually preset for any time period between 1 millisecond and 10 seconds. All four events automatically take lice in sequence upon receiving a single timing pulse. Figure 6.37B is a graphic representation of a generated ADSR envelope. Getting the right kind of responses from various timing pulses takes some thought and prac- tice. The music in figure 6.38 requires series of en- velopes with differing decays. It i, of course, usually ot practical to play the release pot manually while Playing the keyboard. The solution here is to know ‘that the initial decoy in most ADSR’s is terminated With the release of the gate voltage. By holding down. 2 key, the attack and initial decay will cycle as pro- ‘rammed. If the key is just tapped, the initial decay ‘will immediately fall to the sustain level, no se Rms Soe Figure 6.378. The ADSR Figure 6.28. Envelope aniculations ‘matter where itis in its cycle. To produce the articula- tions notated in the example, set the functions as fol- lows: attack = minimum; initial decay = the longest re- quired note value: sustain level = 0 volts: release = minimum. As shown in figure 6.39, 2 sustained key will let the initial decay evolve through its entire cycle, in effect simulating an AD function. If the key (gate voltage) is released at any time before the cycle is complete, the voltage will immediately drop to zero. In this way one can approximate piano-like articula- tions, Remember, this technique only works if the in- itial decay can be interrupted by the release of the gate. More real-time variation in envelope shapes can be accomplishd by voltage control of all the functions. These manipulations are usually only possible on larger instruments, but they certainly are worth the extra expense for the meticulous composer or performer. Eu Instruments manufacture an adjunct to their en- velope generator, called the 2355-Voltage Controlled ‘Transient Generator Input Unit. It is connected to the ery <0 = Figure 6.39. Various (ete deratons used Te ES SSO et ; fees envelope generator by internal connectors. Input volt- ‘ages to the module may be front panel attenuated as needed and then summed with the envelope gener- ators offsets. Thus any or all of the voltage functions may be controlled dynamically. One very common application is to set the EG for short envelopes and use an inverted form of the voltage going to the VCO (see figure 6.404). Since low notes require relatively longer envelopes, the low voltage producing low pitches is inverted to a high voltage and produces pro- Portionally longer decays. Other voltages could be used to control other parameters as in figure 640B. Here the X axis of a joystick is patched to the initial decay input, the Y axis is patched to sustain level input, and some kinesthetically generated keyboard voltage (velocity, pressure, etc.) is inverted to con- trol the attack. This is probably 2 hypothetical patch for the resources of most players, but it does point ‘out the possibilities. A sharp keyboard attack produces 2 short attack, Iow notes have longer final decays, and the overall loudness and initial decay can be controlled simultaneously by the joystick. ‘The applications of trigger delay will come up in future patches but should be briefly discussed here in terms of delayed onset times. All of the above ex- ‘amples involve relatively slow time patterns. In ac- tual usage, envelope generators are usually pro- ‘grammed to function at very rapid speeds, The on- ‘set behavior or attack of most acoustical events has ifferent effects on the various components of a wave. Figure 641 shows the amplitude characteristics of the fundamental and first four overtones in the first 130 milliseconds of a violin attack (at a frequency of 435 Hz). It can be seen that the fundamental has a com- ‘paratively slow rise time when compared to the third ‘overtone. The third overtone also has a very rapid in- itial decay time. In the electronic creation of attacks, the individual control of these various characteristics can be readily directed with an envelope generator reacting on a voltage controlled amplifier. Observing the many amplitude changes in figure 6.41, one can imagine how an envelope generator with an unlimited umber of programmable voltage levels can be of great use in the control of transients. The various voltage levels of the initial rise times are achieved with the ‘use of a voltage attenuator or a control voltage proc- cessor (see page 37). By controlling a bank of envelope {generators from 2 single trigger pulse, all of the com- ponents in figure 6.41 can be controlled. The delayed ‘onset of the fifth overtone is accomplished by using a “trigger delay.” A pulse would directly trigger four envelope generators and a trigger delay would initiate the fifth overtone about 22 milliseconds later (figure 642). Each envelope generator would be programmed to produce the various rise, duration, and decay times for the particular wave component it is controlling. The trigger delay may be a function of the enviope generator, as with Eu Instruments, or may be accom- plished through a delay module as on the Moog studio Snstruments. The Buchla 281 Quad Function Generator (figure 643A) has a feature which would be quite difficult to patch externally (see figure 6.49B). By means of 1 switch, the top and/or bottom two function gen- erators are put in “quadrature.” In essence this means that the two functions are 90° out-of-phase with each other, independent of individual time constants. When a timing pulse is received, FG 1 begins its attack. When the attack peak is reached, FG 2 begins its attack. FG 2s then not allowed to begin its decay until FG 1 has ended its decay. The two envelopes are available {as separate outputs or they may be mixed internally. ‘The Buchla 281 also issues an output trigger when the decay reaches 0 volts. This trigger may be patched to any external module and/or patched by means of ‘@ switch back to the trigger input for cyclical firing. ‘With a patch cord one can also attach the output trig- ger to the input of the next FC. All four FG's can beFigure 6.41. Violin attack at 435 Hz (Amplitude charactoristics of violin attack fom Music, Sound and Sonsation: A Moder Expostion by Fritz Wineks, Dover Publications, in, New York, 1967. Reprinted trough permission ofthe publisher) OT 1 Ee oe Figure 6.42. Transiont waveform synthesis with tigger dolay Patched to cycle and sequence so as to produce 2 multitude of envelope relationships. These patterns can be set up with any number of envelope generators with output triggers. The following are some ideas for envelope patches, including some function processing possibilities A. A single EG to control timbre and amplitude si- multaneously. B. Independent control of timbre and amplitude with two EG’. In this patch be sure the attack time for the VCA is faster than the attack time for the CF, or the initial filtering effect will not be heard. Figure 6.438. Buca Series 251 Ouse Fenction Generator Figure 6.438. Guedratire ‘elgonships In these patches use an inverter on the output of any of the envelopes patched to the filter. C. The filter and VCA are put in parallel and re- mixed. In this way the final gain is not entirely dependent on the filter or VCA, but is a combi- nation of both spectrum and amplitude. ‘The following are examples of multiplication patches (if the reader does not have AC/DC coupled VCAs for multiplication, it may help to tum to page $7 and read about specialized processors). If a control volt- age multiplier is not available, put the audio signal through two VCA's in series (compare figures D and E). D. This is 2 diminuendoing echo effect. The LFO sawtooth or a recycling short envelope is multi- plied by a longer envelope. The same effect can ‘be produed with two VCA’s in series, one with the repetitive function and the other with the pri- mary envelope. (See figure E.) 69F. An envelope's overall magnitude can be controlled ‘by multiplying it with another voltage. Here a pressure or velocity voltage from the keyboard is ‘used, so hard or fast key depressions determine the total magnitude of the envelope—thus harder key articulations produce louder sounds. G. The attack and decay of an AR function is extended by integration. Take care here that integration time is not so slow that it prevents the voltage ‘peak from being reached. HL A variation on patches A, B, or C, in which the spectral sweep can be quantized into harmonic oteps” Sequencers. The sequencer, sequential voltage source, sequential controller, ec. is a programmable memory of non-fluctuating control voltages which can be taught to handle a variety of tasks. Since the sequencer hhas a multitude of applications, it is probably the most complex of all voltage sources. The main func- tion of the sequencer is to supply the composer with 70 -{ Seggeege be Sas mnrea Bae a repetitious stage of preset voltages. The various de- signs may allow for as few as 2 or as many as 256 individual DC voltages to be produced in sequence at varying speeds. The basic design for a sequencer fs ilustrated in figure 6.45. Each individual voltage may be manually set by using the individual voltage pots. By controlling the speed of the sequencer with trigger pulses, it is possible to produce the voltages at a constant rate, or the rate may be changed by varying the speed of the control trigger. The most commonly used source of trigger pulses with the se- quencer is the timing pulse generator or a clock os- Cillator, because the rate of the pulses can be con- trolled’ manually or by Voltage control. Some se- ‘uencers are equipped with an internal speed controlic = — foe = Figure 6.46. Sequencer ostinato ‘and do not necessarily have to depend on external trig- ger pulses. ‘An ostinato passage, as in figure 646 would very easily lend itself to sequencer applications. The se- ‘quencer would be patched to a voltage controlled os- Gilator and each increment of the sequencer would then be set to produce the desired frequency of the ostinato passage. If the tempo is J = 180, the 32nd ‘Rotes must be at a speed of 24 per second. The se- quencer can be programmed to fire at this rate by ‘supplying it with trigger pulses with a period of 4. The ‘sequencer may continue to produce this pattem for any number of repetitions, or it may be programmed to stop after the first repetition. Sequencers with in- ternal speed controls will produce the same tempo by manual setting of the pot. Figure 6.47. Quantized sequences ‘The voltage offset for each increment is usually ‘an analog pot. It is continuously variable throughout its entire voltage range. There are several models avail- able which feature quantization, such as the EML 400 Series Sequencer and the Buchla 248 Multiple Arbi- trary Function Generator. The quantizer will divide the infinitely variable pot voltage into discrete volt- age levels, usually linear voltage divisions to be used in various kinds of equal interval tuning. By quantiz- ing a bank of controls, equal temperament tuning can bbe accomplished faster and more accurately (see fig- sures 647A and B). ‘The programmed voltages in a sequencer bank are non-fluctuating. If controlling 2 VCO, it would nor- mally be impossible to glissando between pitches. Volt- age slides can be produced by using an integrator. The sequencer voltages are processed through the in- 71aeatee Figure 6.48. itegrated sequencer pattems wan _[oocoocos EY | 7 [3888333 B= —— I * Figure 649, Mt-bank or matrix sequencers tegrator and the time constants are set the composer chooses to utilize the other banks is his appropriate (see figure 6.48). The Buchla 248 provides the option of individual increment selection for integration. In this way some pitches can have accompanying glis- sandi and others can be non-glissandi. If applied to 2 VCF, certain increments would cause timbral sweeps ‘while others would produce abrupt timbre changes. Just as with the voltage keyboards, the sequencer may have from one to three (or more) banks of in- ividually controlled outputs. This is sometimes called a matrix sequencer. Each bank can have a different sequence of preset voltages, but the fiting speed for all three banks will remain constant. When the first increment for the first bank fires, so does the first increment for all of the other banks. Whether or not own decision. With multiple banks it is possible to rogram any sequence of frequencies with one set of pots and control the individual amplitudes with an- other set (figure 649A). By using a third output, it is possible to program very complex rhythmic pat- tems. Patching an output to 2 sequencer’s speed con- trol, via a timing pulse generator or its own internal firing control, the player will change each progression to the next increment, with the speed proportional to the voltage of that particular increment (figure 649B). ‘As an example, the relatively simple pattem below could be programmed in the following manner: fe 1 Dion‘The first increment on the sequencer would be set to produce a voltage which, when applied to the se- quencer’s speed control, would advance to the next increment at a rate of once every 2 seconds. The sec- ‘ond and third increments would have to produce a higher voltage which would advance the sequencer at a rate of twice in 1 second, and the fourth increment would have to advance the sequencer at a rate of every 1 seconds, continuing in the same manner for the time value of each note. If this process were being controlled by the third bank of voltages, with fre- quency controlled by the first bank and amplitude by the second, the three basic parameters of musical com- position could be subjected to sequential program- ming Sequencers are particularly, but not exclusively, ap- plicable to serial techniques because of the possibility of transforming any parameter to a voltage and con- ‘tolling it with 2 sequential (serial) source. The standard design of sequencers today consists of either 8, 12, or 16 increments. These are only ar- bitrary numbers decided upon by the manufacturers and are of no special benefit to the composer. By us- ing an “increment switch,” it is possible to fire any number of successive increments in the bank If a ‘composed wishes to have only five increments in a articular pattern, he can set the increment switch to "5" and only the first five voltages would fire as 4 repetitive pattern. Another method of increment se- lection is with individual switches that allow for any ‘umber of individual voltages to be eliminated from the sequence. By switching out the unwanted incre- ments, the composer is not always forced to use the preset successive voltages. In certain designed se- ‘Quencers, this switching may be manual or controlled with triggers. ‘Two other techniques of incrementation available on some instruments are pulse address and analog ad- dress. With pulse address, each increment will have a accompanying pulse input. When a trigger is set to an increment's pulse address input, the sequencer will switch to that increment. The composer may then program 2 series of sub-sequences and call them up as needed. Figure 650 illustrates one possible appli- cation. A 16-stage sequencer is programmed for four different arpeggio patterns to control 2 VCO. The in- ] Sosc0088The term “correlated” random is used because it is 0s- sumed that the sequencer will also control another parameter such as pitch. If a voltage controls both pitch and duration of that pitch, there is an estab- lished and predictable relationship. If the clock is an LFO, a high random voltage will produce a high pitch and a correspondingly shorter duration (higher clock speed). If the clock is of the “pulser” genre, a higher voltage will produce a longer duration as a high volt- age will result in a long period. To adjust an LFO to behave like a pulser, invert the incoming control velt- age —and vice versa. Other methods of random address will be discussed in connection with their application of random voltage sources (see page §3). The number of increments in a bank can be ex- tended if there are multiple banks and if there is the availability of an electronic switch, A 16 x 3 matrix sequencer can be turned into a 45 increment sequencer by using the patch in figure 652. The output voltages, from the three banks are taken to the inputs of an electronic switch. The trigger output of the lst in- crement is taken into the “sequential switching input” of the switch. Every time the sequencer fires trigger #16, the switch will sequentially route the next con trol voltage bank to its output. ‘The sequencer may be tumed on and off by two different methods, depending on its manner of control. Wf it is being triggered by pulses from an external source, it may be stopped simply by stopping the source. Start-top control of the timing pulse gener- ator is extremely useful in this way. If the TPG is in the single pulse mode, the sequencer may be fired at will by the manual depression ofthe firing button. Se- ‘quencers with internal firing and control have self- contained trigger inputs for starting and stopping. The individual increments may_also have a switch which will stop the firing action when that particular incre- ment in the series is reached. One application would be to have one sequencer start and stop another. In figure 6.53 both sequencers’ controls are taken to a VCO (other parameters are also dealt with in the patch—analyze them!) Sequencer “A” is set for 8 stages, and supplies control logic for an arbitrary series of 2800- 2000» 20004 at “Pf — programmed to output 2 pentatonic scale function ‘The first increment of sequencer A sends a trigger to stop sequencer B. The function is as follows: A pro- ces 2 sequence of pitches. At the eighth event, 2 trigger is sent to start B (clocking at its own rate), thus imposing a pentatonic scale on the last pitch. When A recycles back to increment #1, 2 pulse is issued to stop sequencer B. Sequencers with digital memories are often de- signed to “read” externally generated voltages and store them for future recall. Manufacturers. are re- uctant to call such instruments “sequencers,” since they usually are capable of wider applications than the initial sequencer concept. One such instrument is the Roland MC-8 MicroComposer; the EMS AKS and 100, Sequential Systems Instruments, Eu 2500 mod- ules are other examples. With read-in capabilities, ex- ternal voltages, their durations and timing pulses are ‘tumed into numerical information by what is called an ADC or analog-to-digital converter. Once in mu- merical form, they are stored in a digital memory. Application of timing pulses from a clocking device then reads out the numbers to a DAC or digital-to- analog converter which converts the information into the original control voltages. Since the voltage levels are stored as numbers, various types of information ‘Processing, such as multiplication and division, can ‘be applied As the voltages are stored as numbers, they can be accessed or read out at different rates without chaiging their value. Manipulation or voltage control of the clock will only affect the rhythm or speed of retrieval and will not alter the value of the outputted voltage. Digital memories are as varied and offer as many specialized features as there are manufacturers, andFigure 6.53. Sequencer “star-stop” patch therefore it would be impossible to cover-every op- tion in this context. There are, however, some basic ‘capabilities which should be understood by the pros- pective purchaser and user. Most digital memories are volatile. This means that when the power is shut off, the information stored in memory is erased. Micro- pprocessor-based instruments have the possibility of “dumping” the stored information into a long-term storage device such as regular audio tape, floppy discs, ete. With the increasing accessibility of digital stor- age, many home-built instruments and a few commer- ial instruments are employing PROM storage (Pro- grammable Read Only Memory). A PROM is a small integrated circuit on which the user or design can permanently store a sequence of information that is commonly used. This can be non-traditional tuning for Keyboards, envelope functions, functions for panning, sound (see chapter 13), or whatever. Digital memories and many analog memories do not have to be read in sequential order, so that the information may be read out in any order (see analog and pulse address, page 73). A great advantage to PROM storage is that they are non-volatile, and. thus the information is not lost when the instrument is tuned off. Erasable PROM’s (EPROM) can be used for long-term storage and re- Programmed as desired One problem with digital storage can be the tempo and rhythm. Memories are not infinite. The number of events which can be stored is usually some binary number such as 256, 512, 1024, etc. Rhythm and rests are stored events, and depending on design, can use up.a significant amount of that memory. What we shall call design “A” presents the problems. In this format a clock rate is set, for example, at 10 Hz. Every tenth of 2 second the memory looks at whatever volt- age is present at its input and stores it. Everything is satisfactory until information is given faster than every tenth of a second, such as rapid tris (see figure 6.54). ‘The memory will only be able to store those portions of the trill coinciding with the tenth of a second read command. The seemingly logical solution is to speed up the clock so it can get all of the information. So we now set the clock for perhaps 30 Hz. We do get the trill but the memory is being clocked through its total capacity at a faster rate so the total time is sig- nificantly shorter. Using this type of memory then calls for a balanced compromise of information rate and clock speed. Simple measured sequences use less mem- cory and provide longer storage times, while more com- plex rhythms at faster tempi result in shorter storage times. With design “B’, rhythm and tempo do not impose restrictions on sequence length. In this method two separate words of information are stored for each event. The ADC stores the voltage information and the number of clock counts associated with each volt- age. The clock is free running at a high frequency and external to the actual memory. This is exemplified in figure 6.55. Clock speeds usually run at higher rates, but let's hypothetically assume that the reference is sending out clock pulses at a frequency of 600 Hz, or expressed another way, the memory can make 600 ob- servations each second. At a tempo of J = 60, a quar- ter Hote is equivalent to 600 counts, an eighth note equal 300 counts, 2 sixteenth note equals 150, each division of a quarter note triplet receives 200 counts, etc. As the player reads in the information, the pitch {information is put in one register (a place where digi- 75Figure 6.54. Digital memory Figure 6.55. Digital memory tal information is stashed), and the duration or num ‘ber of counts that the voltage is to be present is put in a separate but parallel counting register. Register A now contains a series of digital words to be used as pitch information without using that memory for any rhythmic information, Register B has a string of associated numbers that will be translated back into durations for each voltage. In reading the informa- tion back, the first pulse of the clock calls out the “F” voltage and that voltage will be present at the ‘output for 600 clock counts. On clock count 601, the memory switches to the second address, putting out the “A” voltage and will hold that voltage for 300 counts, etc. The clock is only putting out pulses so that it has an unlimited number; the counting is being done by the memory, and hence the durations can be of virtually any length and the number of possible stored events is not eaten into, These types of mem- ories are finite (again usually « binary number), but the possible number of storable events is not changed by tempo or durations. The articulation of @ note is controlled by a VCA and an envelope generator, as is usually done with analog sequencers. Consequently the digital memory must produce timing pulses whose period is equal to the period of each voltage change— again the same as analog sequencers. Timing pulse accessibility on a sequencer greatly defines its range of applications, and an openly de- signed sequencer may be looked at as a programmable timing pulse source. The way sequencer pulse out- puts are accessed varies from instrument to instrument ‘but can be put in three general categories: individual pulse outputs, bus outputs, and multiple bus outputs. These three access methods are illustrated in figure 76 Figure 6.56. Sequencer pulse oxtputs 656. Figure 657 illustrates how two sequencers can be used to structure polyphonic events. Sequencer 1 produces controls to generate an eight note ascending scale (VCO. 1). On increments 1, 3, 6 and 8 a trigger is taken to start sequencer 2. This sequencer is clocked at a much faster speed and generates a four note em- Dbellishment figure. All of sequencer 2s triggers are used to fire an envelope generator so that the pitches are heard only when the sequencer is running. The trigger from increment 4 is taken to the stop input, therefore the sequencer will stop on the last stage. The bus output can be set up in two ways. The pulses may be hardwired to the output as in figure 6584, or may be user-switched as in figure 658B. In