Scribd
Upload
36 views22 pages

Introduction To Multimedia Compression: National Chiao Tung University Chun-Jen Tsai 2/21/2012

This document provides an introduction to multimedia compression. It discusses how multimedia data such as audio, video and images can be compressed by modeling the data and encoding the model along with residual errors. Specifically, it explains that lossless compression techniques model patterns in text data, while lossy techniques are often used for multimedia as they can achieve higher compression by tolerating small distortions. It also gives examples of how linear, differential and variable length coding can be used to compress data by modeling relationships within the data stream.

Uploaded by

Puja Astawa
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
36 views22 pages

Introduction To Multimedia Compression: National Chiao Tung University Chun-Jen Tsai 2/21/2012

This document provides an introduction to multimedia compression. It discusses how multimedia data such as audio, video and images can be compressed by modeling the data and encoding the model along with residual errors. Specifically, it explains that lossless compression techniques model patterns in text data, while lossy techniques are often used for multimedia as they can achieve higher compression by tolerating small distortions. It also gives examples of how linear, differential and variable length coding can be used to compress data by modeling relationships within the data stream.

Uploaded by

Puja Astawa
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 22

Introduction to Multimedia Compression

National Chiao Tung University Chun-Jen Tsai 2/21/2012

Data is Everything, and Nothing


Everyday, we are bombarded by all kinds of information (data)
textbooks, news papers, movies, songs, conversations, lectures, preaching,

What are the purposes of all the information?


To cause chemical reactions in your brains To duplicate chemical reactions from one brain to the other To

2/22

Three Aspects of Data


To distribute information (data) around to serve your purposes, there are three aspects one should worry about:
Quantity Reliability Security

Example: knowledge of multimedia compression

3/22

The Paper That Starts It All


In 1948, Claude E. Shannon published the revolutionary paper, A Mathematical Theory of Communication.
Later, in 1949, a book was published based on this paper, but the first word of the title was changed from A to The

The paper provides many insights into the essence of the communication problem
In particular, Shannon perceived that all communication is essentially digital !

4/22

Data Distribution Systems


Shannon was the first person to partition a communication system as follows:
Information Sources Received Information

Source Encoder

Channel Encoder
Channel

Channel Decoder

Source Decoder

Get most compact Representation of information Get most robust Representation of information

The information content of a source and the information capacity of a channel can be identified using the concept of entropy
5/22

The Origin of Information Theory


The term entropy was first used in thermodynamics and in statistical mechanics
Some people think that information theory grew out of statistical mechanics because L. Szilard applied an idea of information to solve a physical problem in 1929

However, Shannons work evolved from the field of electrical communication


entropy was used in information theory merely due to its mathematical analogy with the entropy of statistical mechanics

6/22

Entropy in Thermodynamics
In thermodynamics, entropy is a measure of thermal energy of a body of gas
Low entropy

High entropy

Statistical mechanics says that an increase in entropy means a decrease in predictability


7/22

Linking Back to Information Theory


The complexity of a system depends on our knowledge of the system; the more we know about the system, the less words we need to describe the system In information theory, the amount of information conveyed by a message increases as the amount of uncertainty as to what message actually will be produced becomes greater

8/22

Some Information
Check the entropy of the following messages
My dog cannot fly My dog runs faster than a chicken My dog is a lady dog My dog runs slower than a chicken My dog can sing

It seems that, a rare message carries more information than a common message

9/22

Frequency-based Coding
Morse code
Invented in 1838 by Morse for electrical telegraph, and expanded by Vail in 1844 To shorten the transmission of messages, English text was coded based on relative frequencies of occurrence The efficiency of Morse code can only be improved by 15% using modern theory Questions: efficient for all languages?

fig. ref.: wikipedia

J. R. Pierce, An Introduction to Information Theory, 2nd. Ed., Dover Publications, 1980.


10/22

Context-based Coding
Braille code, by Louis Braille in 1825 Grade 1 Braille
Letters and numbers

A or 1

B or 2

C or 3

D or 4

E or 5

F or 6

G or 7

H or 8

I or 9

J or 0

Symbols

or ?

( or )

Grade 2 Braille

AND

CH

SH

ST

TH

11/22

Model-based Coding
Statistical structure is not the only way of compression. Describing things using models is usually less wordy For example, what is the minimal precise description of ?
Shannons idea the unpredictability of patterns of digits in Kolmogorovs idea the size of a program that computes

12/22

How Large Is the Amount of Data?


1 second of CD audio:
44100 samples 2 channels 16 bits/sample = 1,411,200 bits

1 second of 1080p HD video:


1920 1080 pixels 3 color channels 8 bits/color sample 30 frames = 1,492,992,000 bits

Sometimes, large data amount is a technique against piracy

13/22

The Future Is Here


Super Hi-Vision (8K system)
7680 4320 = 33 Mega pixels per frame Physical data rate (video-only): 180-600 Mbps

http://www.nhk.or.jp/digital/en/super_hi/

14/22

Data Compression Concept


X original data, Xc compressed representation, y reconstruction

compression

xc

decompression

Lossless compression: when y is equal to x Lossy compression: when y is different from x Compression ratio: |X| : |Xc| or (|X| |Xc|/ |X|) *100%
For example, |X| = 65536 bytes, |Xc| = 16384 bytes, the compression ration is 4:1 or 75%.

Data rate: for time-varying data, the number of bits per second (or sample) required to represent the data
15/22

Lossless and Lossy Compressions


Text compression techniques are often lossless
Any counter examples?

Image, audio, video compression techniques are often lossy


Any counter examples?

Distortion: the difference between the original and the reconstruction


If the distortion is small, we say the quality or fidelity is high. Or, we say the reconstruction is a high-definition copy of the original
16/22

Modeling and Coding


One of the most powerful tools in data compression is called data modeling
Model a systematic way to describe data

A common data compression scheme is to encode a description of the model, and a description of how the data differ from the model (aka, residual)
By encode, we mean to put the data in binary digits
Data Source Are they similar? Model
17/22

Example 1: Linear Model


Data sequence xi,
9 11 11 11 14 13 15 17 16 17 20 21

...

Model: xn = n + 8, n = 1, 2, Residual: en = xn x n = 0, 1, 0, 1, 1, 1, 0,

18/22

Example 2: Differential Model


Data sequence xi,
27 28 29 26 27 29 28 30 32 34 36 38

...

Model: x1 = 0, xn = xn1, n = 2, Residual: en = xn x n = 27, 1, 1, 1, 2, 1, 2,

19/22

What About Speech Models?


Typical speech signals:

frequency

time

spectrogram

waveform

20/22

What About Image Models?


Typical image signals:

21/22

Example 3: Variable Length Coding


Given a sequence of symbols:

If fixed length coding (FLC) is used: 3 bits per symbol If variable length coding (VLC) is used: 2.58 bits per symbol 1.16 : 1 compression ratio

22/22

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy