A Text to Speech (TTS) System with English to Punjabi Conversion

Prabhsimran Singh Amritpal Singh

Department of Computer Science & Engineering Department of Computer Science & Engineering
Guru Nanak Dev University, Amritsar Guru Nanak Dev University, Amritsar
prabh_singh32@yahoo.com amritsinghmand@gmail.com

AbstractThe paper aims to show how an application can be

developed that converts the English language into the Punjabi
Language, and the same application can convert the Text to
Speech(TTS) i.e. pronounce the text. This application can be
really beneficial for those with special needs.

KeywordsText to Speech, Translator, Parsing, English to

Punjabi, Text Conversion.

I. INTRODUCTION
Now days the main aim of technology is to help people in
their day to day requirements and facilitate their work. People
of all age categories from young to old use computers, even
those with visual disabilities.

All this has been made possible with the development of

systems which can recognize text and speech, this makes blind
and visually impaired people to use computer systems without
the help of any other person. But if we combine these systems
with language translation systems it gives an added advantage
to those who are not so familiar with English language[1].

II. TEXT TO SPEECH

A Text to Speech (TTS) system is a computer system that
converts text into speech, i.e. read the text automatically when
asked to do so[2]. This system is a combination of both
hardware and software. Generally a Text (Sentence) is
composed of collection of words, while words are
combination of alphabets arranged in a meaningful way. The
work on (TTS) systems goes back to 1779, when the Danish
scientist Christian Kratzenstein, build a human vocal model
that can produce five long vowel sounds i.e. [a], [e], [i], [o]
and [u].

Generally a TTS system consists of 5 fundamental phases:

a. Text Analysis and Detection

b. Text Normalization and Linearization
c. Phonetic Analysis Fig. 1 Overview of TTS[2]
d. Prosodic Modeling and Intonation
e. Acoustic Processing
III. LANGUAGE TRANSLATOR
The Text to be converted is passed through these phases
A language translator is a system that converts the input
step by step in order to obtain the Speech. Figure 1, shows text from one language to another. There are various
diagrammatical representation of these phases. translators available on the internet, with most famous being
the Google Translate[3]. It support multiple language ranging
over different regionals as well as national languages. Figure

1
2, shows a screenshot of Google Translate, where English is A. Requirements for the proposed system
being converted to Spanish language. The aim of this paper is The requirements can be classified as hardware and
to convert English Language to Punjabi Language, and to give software. The Minimum Hardware requirements include Dual
additional functionality of pronouncing the text i.e. Text to Core Processor, 2GB Ram, and 500 mb Hard disk. The
speech (TTS).. minimum Software requirements are Microsoft Windows XP
service pack 3 or higher, Microsoft .Net framework 4.0.
IV. LITURATURE REVIEW AND MOTIVATION
All around the world various systems has been developed B. Architecure of Proposed System
which translate one language to another and that covert text to For each English words all possible Punjabi words
speech. Some of them which act as motivation for our system substitutions are stored in the database, and infact this creates
are as following: one of the major problem ambiguity i.e. for one English word
there is more than one substitutions. Table 1, shows some
A. English language to Arabic language convertor by
words with multiple substitutions. So proper rules need to be
Ameera Al-Rehili, Dalal Al-Juhani, Maha Al-Maimani
and Munir Ahmed. It converts English Language to applied to check which word fits best in the sentence, while
Arabic Language and Vice Versa. Moreover has the conversion of English to Punjabi.
ability to pronounce English as well as Arabic Text [1].
Table 1, Ambiguity in Punjabi for English words
B. A system that converts Hindi language to Telugu English Punjabi
Language and vice versa proposed and implemented by
You qusI
Lakshmi Sahu. This system has the ability to pronounce
hindi and telugu text in male voice and female voice [4]. You qUM
Your qyrw
C. Text to Speech: A Simple Tutorial by D. Sasirekha and Your quhwfw
E. Chandra, shows what all is required to develop a
system that has capabilities to convert Text to Speech. He
paper discuses the various TTS systems and continuous VI. WORKING OF THE PROPOSED SYSTEM
improvement that is taking place in this field [2].
The working of the entire system can be broken down into two
main parts i.e. Translation and Text to Speech(TTS).
V. PROPOSED SYSTEM
The aim of the proposed system is to develop a system A. Translation
that has capability to perform Translation (English Language
The Figure 3, shows the User interface of the system.
to Punjabi Language), Converting (Text to Speech i.e. read the Whenever button to convert the text is pressed, the entire
Text aloud). The proposed system will be developed in sentence/string is parsed(Divided) into single words/units, and
Microsoft Visual Web Developer Express edition 2010, with then corresponding word/sound in the database is searched
C#.Net as front end while SQL Server 2008 as back end [5].
which satisfy certain set of rules for each and every
The System thus developed can be used as a web application
words/units and is replaced. This process is repeated till the
or desktop application by deploying it as local host. The
entire sentence is processed/converted. Figure 4, shows the
system proposed her is developed for a small domain of
entire process in step by step fashion.
English words.

Fig 2. Screenshot of Google Translate [3]

2
 correct. Hence the final result will be ieh iks ny kIqw? .
B. Detailed working of the translation System
The description is shown in figure 5.
Suppose we want to translate an English sentence Who
did this? into Punjabi Language. First of all the entire string
Table 2, Punjabi words for corresponding English words
will be parsed, and will result in 4 independent unit/words.
English Punjabi
Then for each english word corresponding Punjabi word will
be selected from the database as shown in Table 2. Who iks ny
did kIqw
Now if these words are combined to form a Punjabi
this ieh
sentence, it will result in iks ny kIqw ieh? , which is ? ?
grammatically incorrect. Here the rules come in play, and
rearrangement of sentence take place to make it grammatically

Fig 3, User Interface of the System

Fig 4, Stepwise process of Language Translation

3
 VII. COMPARISON OF PROPOSED SYSTEM WITH FEW OTHER
SYSTEMS
In this section the proposed system will be compared
with few other systems based on various parameters like
language used, database used etc as shown in table 3.

Table 3, Comparison of Different System

System
Proposed System
Proposed by Ameera Proposed
System Al-Rehili by Lakhsmi
and Sahu[4]
group[1]
Front End
(Language C#.Net Vb.Net C#.Net
Used)
Back End SQL Server MS Access
---
(Database) 2008 2007
Language
Conversion High Moderate High
Efficiency
Text to
Yes Yes Yes
Speech (TTS)
TTS other Yes Yes (Hindi
Yes (Arabic)
than English (Punjabi) & Telugu)
Pronunciation
High High High
Efficiency
Web /
Desktop Both Desktop Desktop
Application

VIII. CONCLUSION AND FUTURE WORK

Fig 5, Detail Working of Translation Process In this paper we presented a system that can convert the
English sentence into corresponding Punjabi sentence, as well
as read the text aloud. The system is capable to be used as a
C. Text to Speech(TTS)
web application or desktop application (deploying it as local
To convert Text to Speech we use the inbuilt .Net host), hence making it more flexible. This system can be really
Library/Class System.Speech.Synthesis[6]. The Code for helpful for those who are not so familiar with English
this is shown is figure 6. The System is coded to speak both language as well as for those with visually disabilities.
English as well as Punjabi languages. When we select a
language and press the Speak (picture) button the text is read In Future, various advancements can be made to the system
aloud. The system accurately pronounces the Punjabi like adding the facility of Speech to Text (STT) so that user do
language. The working interface of TTS is shown in the figure not have to enter text manually, it can be done simply by
7 and 8. speaking only. Current system can only supports the
conversion of English language to Punjabi language, but we
can add support for conversion of more languages to it so that
it can become more useful. Moreover adding the facility of
inbuilt dictionary i.e. giving meaning of different words
available in the sentence.

Fig 6, Code of TTS [6]

4
 Fig 7, Select Language Option

Fig 8, Speak Button

REFERENCES
1. Ameera Al-Rehili, Dalal Al-Juhani, Maha Al-Maimani
and Munir Ahmed, A Novel Approch to convert
speech to Text and Vice-Versa and Translate from 4. Lakshmi Sahu, Hindi & Telugu Text-to-
English to Arabic Language, IJSAIT, Volume 1, Speech Synthesis (TTS) and inter-language
No.2, May-June 2012. ISSN:2278-3083 text Conversion,International Journal of
Scientific and Research Publications, Volume
2. D. Sasirekha, E. Chandra, Text to Speech: A Simple 2, Issue 4, April 2012. ISSN:2250-3153
Tutorial, IJSCE, Volume 2, Issue 1, March 2012.
ISSN:2231-2307 5. Microsoft Visual Web Developer 2010,
http://www.visualstudio.com/en-
3. GoogleTranslate, https://translate.google.co.in us/products/visual-studio-express-vs.aspx

5
6. Text to Speech(TTS) Code,
http://msdn.microsoft.com/en-us/library/
system.speech.synthesis (v=vs.110).aspx