SlideShare a Scribd company logo
1 of 8
Download to read offline
MODELING OF SPEECH SYNTHESIS OF
STANDARD ARABIC USING AN EXPERT SYSTEM
Tebbi Hanane1 and Azzoune Hamid2
1,2

LRIA, Option: Representation of Knowledge and systems of Inference,
USTHB, Algiers
1
2
tebbi_hanane@yahoo.fr , azzoune@yahoo.fr

ABSTRACT
In this work we present our expert system of speech synthesis based on a text written in
Standard Arabic, our work is carried out in two great stages: the creation of the sound data
base, and the transformation of the written text into speech (Text To Speech TTS). This
transformation is done firstly by a Phonetic Orthographical Transcription (POT) of any written
Standard Arabic text that to transform it into his corresponding phonetics sequence, and
secondly by the generation of the voice signal which corresponds to the chain transcribed. We
spread out the different steps of conception of the system, as well as the results obtained
compared to others manners of works studied to realize TTS based on Standard Arabic.

KEYWORDS
Engineering knowledge, modelling and representation of vocal knowledge, expert system, (TTS
Text To Speech), Standard Arab, PRAAT.

1. INTRODUCTION
Generate the voice is a complex work because of the variability intra and interlocutor of the voice
signal. In computer science the difficulty of modeling the speech signal is for the reason that we
don’t know yet how to model very well the enormous mass of knowledge and information useful
to the signal synthesis. Thus we have made a choice to use an expert system to modeling that
knowledge to build a robust system which can really read a text written in a language chosen
especially in Standard Arabic. To obtain a better organization of our work, we defined our direct
aims. We divided our modeling into three essential stages; the signal analysis, the phonetic
orthographical transcription (POT) and finally the synthesis of the textual representation written
in Standard Arabic. So the finality considered here is that the user can understand the different
phrases transcribed and synthesized which will be pronounced with a clear and high quality
manner.

2. PREVIOUS WORKS
At the current time, we can judge that works emanated in the same context of ours are still not
really colossal, and this is because of the complexity of the language itself. And if this some
works exists, they are based on the same principle as transcribers of the other languages (French,
David C. Wyld et al. (Eds) : CCSIT, SIPP, AISC, PDCTA, NLP - 2014
pp. 469–476, 2014. © CS & IT-CSCP 2014

DOI : 10.5121/csit.2014.4241
470

Computer Science & Information Technology (CS & IT)

English, etc.) [1], nevertheless, the efforts undertaken are encouraging and open a large window
to follow research task in this field.
−

TTS system MBROLA [2] which use the code SAMPA during the stage of transcription,
in this case the user
must respect the form of the SAMPA code which is not a
universal code;

−

Work of S. BALOUL [3] who represents a concrete example of transcription of words;
based on morphological analysis, and on the studies of pauses to generate the
pronunciation of the texts.

−

SYAMSA (SYstème d’Analyse Morphosyntaxique de l’Arabe), realized by SAROH [4].
According to him, "the phonetisation of the Arabic language is based in particularly on
the use of lexicons and on a morphological analyzer for the generation of the different
forms of a word. In addition, they are the phenomena of interaction among the words
(connection, elision, etc.) and the phenomena of assimilation which suggest the uses of
phonological rules "[5]. This tool ensures for each word in entry, the root which
correspond to it as well as the morphological and phonetic representations of the root.

−

The GHAZALI [6] project which was carried out within the IRSIT (Institut Régional des
Sciences Informatiques et des Télécommunications) of Tunisia, it is based on a
transcription work which fits inside the framework of the realization of a TTS system.
The characteristic of this system is shown in the use of a set of rules in the emphasis
propagation.

−

SYNTHAR+ [7] which is a tool studied by Z. ZEMIRLI within the NII (the National
Institute of Informatics of Algiers), it ensures the transcription step for a TTS system so
that it transmits the phonetic representation to the MULTIVOX synthesizer. It should be
known that SYNTHAR+ is based on a morphological analysis before realizing the
transcription.

3. STRUCTURE OF THE SYSTEM
The uses of the concept of code and the introduction of high levels in analysis (morphological,
syntactic, and pragmatic.) makes the transcription task so difficult and requires deepened studies
of the language itself. The difference in our work compared to all that exists is in the transcription
using graphemes, i.e. the uses of the Arabic characters as basic units to transcribe directly the
text, indeed modeled by an expert system.

Definition of an expert system:
An expert system is an informatics tool of AI (Artificial Intelligence), conceived to simulate the
knowhow of an human expert, in a precise and delimited field, this thanks to the good
exploitation of a set of knowledge provided explicitly by expert of the field [5].
The E.S is a system where the data (the knowledge database) is quite separated from program
which handles it (the inference engine). In our case, we built our expert system basing on the two
following components:

A knowledge database: It contains a database of facts and a database of rules, represents the
knowledge (we speaks here about the sound database) and knowledge - to make (a set of
Computer Science & Information Technology (CS & IT)

471

rewriting rules). The fact database integrate two types of facts: permanent facts of the field and
deduced facts by the inference engine which are specific to the field considered (voice synthesis).

The inference engine: It is capable to reason starting from the information contained in the
knowledge database and to make deductions. The inference engine uses the facts and rules to
produce new facts.
Thus the role of an expert system is then to infer generally rules like if/then. To build this system
our work targets the following steps:

Figure 1. General schema of our system of speech synthesis

3.1. Acquisition
It is the first step of our process of voice synthesis; it plays the role of the interface intermediate
between the user and the system. During this time the user have to enter his set of phrases to be
pronounced by the system. After that the automatic process can begin.

3.2. Preprocessing and standardization of the text
In this phase any form of text is transformed to it literal, perfectly disambiguated form [1]. In
particular, this module draft problems of formats, dates and hours, abbreviations, numbers,
currencies, addresses, emails…etc. It was shown that majority errors of conversion
grapheme/phoneme, for the best operational systems came from the proper names and the
exceptions that they pose [8]. Here are some examples of the processing carried out in our case:
−
−
−

The replacement of each composed character by its equivalents
Example: ‫ال‬
‫ل ا‬
Consultation of the exceptions lexicon to eliminate the special words.
The application of the transcription rules established for the language. This module must
be able to process grammatical complexes tasks to identify the rules of transcription
which will be used in a considered context.
472

Computer Science & Information Technology (CS & IT)

3.3. Analysis (the extraction of the characteristics)
The aim of the analysis of the voice signal is to extracts the acoustic vectors which will
be used in the stage of synthesis follows. In this step the voice signal is transformed into a
sequence of acoustic vectors on the way to decrease redundancy and the amount of data to be
processed. And then a spectral analysis by the discrete Fourier transform is performed on a signal
frame (typically of size 20 or 30 ms) [9]. In this frame, the voice signal is considered to be
sufficiently stable and we extract a vector of parameters considered to be enough for the good
operation of the voice signal. In the speech synthesis, the characteristics extraction step,
commonly known as the step of analysis, can be achieved in several ways. Indeed, the acoustic
vectors are usually extracted using methods such as temporal encoding predictive linear (Linear
Predictive Coding LPC) or Cepstrales methods as the MFCC encoding (Mel Frequency Cepstral
Coding), as well as the process of segmentation, etc. This process delimits on the acoustic signal
a set of segments characterized by labels belonging to the phonetics alphabet of the language
under consideration.
At present, the segmentation completely automatic of a voice signal remains a fastidious task.
Indeed, looking at the complexity acoustico-phonetics phenomena being studied, this activity
requires often a manual intervention. Generally, the methods that perform the segmentation of the
acoustic wave are divided into two great classes:
−

The first include all the methods which allow segmenting a voice signal without a priori
knowledge of the linguistics content of this signal. These methods split the voice signal
into a set of zones homogeneous and stable;

−

The second class includes all the methods which allow segmenting the voice signal
basing on a priori linguistic description (typically on phonemes) of this signal. These
methods of segmentation are revealed like methods with linguistically constraints. In our
state, we have opted for the second technique using a tool of voice signal analysis
which is PRAAT [10] so to slice manually the speech signal in a succession of segments,
each one associated with an element acoustic unit (phoneme or diaphone).

Figure 2. Decomposition in polysound of the word [ ‫]جلس‬
Computer Science & Information Technology (CS & IT)

473

3.4. Modeling of the sound database
Majority of work curried out in the field of the spoken communication required often the
recording, and the handling of corpuses of continuous speech, and that to curry out studies on the
contextual effects, one the phonetic indices, and variability intra and inter-speaker. Our corpus is
modeled by the diagram of class1 as follow:

Figure 3. The diagram of class of our sound database

3.5. The training
Here we referred to a POT module to convert each grapheme in phoneme according to the context
and that is possible using a set of rewriting rules. The principal advantage of this approach is the
possibility to modeling the human linguistic knowledge by a set of rules which can be integrated
in expert systems. Each one of these rules has the following form:
[Phoneme] = {LC (Left Context)} + {C (Character)} +
Here is concrete example of transcription rules [11]:

{RC (Right Context)}

#
is a beginning sign of the sentence,
$
is a sign of end of the sentence,
§
is extremity of a word,
C
is a Consonant,
V
is a Vowel,
SC is a Solar Consonant and LC is a Lunar Consonant
[uu]={SC}+{ُ }+ {‫} و‬
[uu]= §+{ُ }+ { ‫} و‬
[uu]={LC}+{ُ }+ {‫} و‬
[uu]= $+{ُ }+ { ‫} و‬
When the " ‫ " و‬is preceded by the vowel / ُ /
obtain the long vowel [ uu ].

and followed by a consonant, we

1: It is a diagram used in UML(Unified Modeling Language) which is used in object oriented modeling
474

Computer Science & Information Technology (CS & IT)

When the " ‫" و‬
vowel [ uu ].

is preceded by the vowel / ُ

/ in final position, we obtain the long

3.6. The synthesizer (the voice generation)
The generation of the voice signal is the real synthesis of speech. This operation consist of a
transform of the phonetic sequence (which represents the pronunciation of the written text)
resulting from the transcription step to its substance, i.e. to it acoustics realization. During this
step we choose from our sound database the units (phonemes, diphones) most suitable to build the
sentence to generate, this means that we will create a automatic function of reading, so that in the
end, the user has only to listen to the synthetic sentences. To improve the quality of the generated
synthesis, we increase each time the sound unit only. The general silhouette of our expert system
can be as follow:

Figure 4. General diagram of our expert system

4. TESTS AND RESULTS
To test the performances of our TTS system based on Standard Arabic language, we have chosen
a set of sentences which we judged like reference since they contain almost the different possible
combinations specific to the language itself. To calculate the success rate (SR) associated with
each sentence tested; we got the following formula:

SR =

Numberofphrasewellprononced
*100%
Numberofphrasestested

The system present in general a SR of 96 % for the set of the sentences tested. Results obtained
are summarized by the following table:
Computer Science & Information Technology (CS & IT)

475

Table 1. Rate of success for a sample of selected sentences
Majority content
Short vowels
Long vowels
Solar consonants
Lunar consonants
Isolated words
Sentences
Numbers
Exception Words

POT
100%
100%
100%
100%
100%
100%
90%
100%

Synthesis
95%
95%
97%
95%
80%
75%
95%
/

by Phonemes

Synthesis by Diphones
/
/
/
/
90%
85%
100%
/

5. CONCLUSION AND PROSPECTIVE
A system for synthesizing Arabic speech has been developed based on an expert system which
uses a set of subphonetic elements as the synthesis units to allow synthesis of limited-vocabulary
speech of good quality. The synthesis units have been defined after a careful study of the phonetic
properties of modern Standard Arabic, and they consist of central steady-state portions of vowels,
central steady-state portions of consonants, vowel-consonant and consonant-vowel transitions,
and some allophones. A text-to-speech system which uses this method has also been explored.
The input of the system is usual Arabic spelling with diacritics and/or simple numeric
expressions. Synthesis is controlled by several text-to-speech rules within the rule database of the
expert system, which are formulated and developed as algorithms more suited for computer
handling of the synthesis process. The rules are required for converting the input text into
phonemes, converting the phonemes into phonetic units, generating the synthesis units from the
phonetic units, and concatenating the synthesis units to form spoken messages. The suitability of
the method for synthesizing Arabic has been shown by realizing all its functions on a personal
computer and by conducting understandability test on synthesized speech. So, we have detailed
the different components which represents the basic blocks of our TTS system based on a written
text in Standard Arabic and our modeling of it with the use of an expert system, this tool which is
a fruits of the artificial intelligence remain less used in the field of automatic speech processing.
This fact encouraged us to explore this world trying to give a little push to the research done in
this multidisciplinary field. Like prospective to our work we show the following points:
−
−
−

Improvement of the quality of voice generated by using methods of modification of the
prosody;
The use of others techniques of synthesis such as the Synthesis using the unit’s selection;
Make, if possible, the signal segmentation totally automatic.

REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]

Tebbi Hanane, « Transcription orthographique phonétique en vue de la synthèse de la parole a partir
du texte de l’Arabe », Mémoire de magister en Informatique, Blida, Algérie, 2007
http//:tcts.fpms.ac.be/ synthesis/ Mbrola.html.
Baloul, « Développement d’un système automatique de synthèse de la parole à partir du texte arabe
standard voyellé ». Thèse de doctorat d’université, Le Mans, France, 27 Mai 2003.
SAROH Abdelghani, Base de données lexicales dans un système d'analyse morpho-syntaxique de
l'arabe : SYAMSA, Toulouse 3, 1989.
Guerti Mhania, Tebbi Hanane, La Conversion Graphèmes Phonèmes En Vue D'une Lecture
Automatique de Textes en Arabe Standard, LANIA, Chlef, Algérie, 2007 S.
Dichy J. & Hassoun M.O. (1998), Some aspects of the DIINAR-MBC research programme, in
Ubaydly A., ed.,1998: 2.8.1-5.
476
[7]

Computer Science & Information Technology (CS & IT)

SYNTHAR+: Arabic speech synthesis under multivox. (SYNTHAR+: Synthèse vocale arabe sous
Multivox.) (English) RAIRO, Tech. Sci. Inf. 17, No. 6, 741-761 (1998).
[8] Mehdi Braham, Oumaya Abbes, Maroua Trabelsi, Mehdi Dahmen, Rania Nouaari, « Intelligence
artificielle : Diagnostic par système expert, Modèle d’analyse d’acteurs : MACTOR »
[9] P. Boula de Mareüil, « Synthèse de la parole à partir de courriers et évaluation de la conversion
graphème-phonème ». LIMSI-CNRS
http://www.limsi.fr/ Individu/ mareuil/
[10] Paul Boersma and David Weenink, "Praat: doing phonetics by computer" Phonetic Sciences,
University of Amsterdam Spuistraat 210, 1012VT Amsterdam The Netherlands.
[11] T. Saidane, M. Zrigui, & M. Ben Ahmed, « La transcription orthographique phonétique de la langue
Arabe ». RECITAL 2004, Fès, 19-22 Avril 2004.

More Related Content

What's hot

CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...
CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...
CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...ijcsit
 
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTK
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTKPUNJABI SPEECH SYNTHESIS SYSTEM USING HTK
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTKijistjournal
 
5215ijcseit01
5215ijcseit015215ijcseit01
5215ijcseit01ijcsit
 
Hindi digits recognition system on speech data collected in different natural...
Hindi digits recognition system on speech data collected in different natural...Hindi digits recognition system on speech data collected in different natural...
Hindi digits recognition system on speech data collected in different natural...csandit
 
International journal of signal and image processing issues vol 2015 - no 1...
International journal of signal and image processing issues   vol 2015 - no 1...International journal of signal and image processing issues   vol 2015 - no 1...
International journal of signal and image processing issues vol 2015 - no 1...sophiabelthome
 
Speaker recognition in android
Speaker recognition in androidSpeaker recognition in android
Speaker recognition in androidAnshuli Mittal
 
Bachelors project summary
Bachelors project summaryBachelors project summary
Bachelors project summaryAditya Deshmukh
 
A Survey on Speaker Recognition System
A Survey on Speaker Recognition SystemA Survey on Speaker Recognition System
A Survey on Speaker Recognition SystemVani011
 
Extractive Summarization with Very Deep Pretrained Language Model
Extractive Summarization with Very Deep Pretrained Language ModelExtractive Summarization with Very Deep Pretrained Language Model
Extractive Summarization with Very Deep Pretrained Language Modelgerogepatton
 
Investigation of Text-to-Speech based Synthetic Parallel Data for Sequence-to...
Investigation of Text-to-Speech based Synthetic Parallel Data for Sequence-to...Investigation of Text-to-Speech based Synthetic Parallel Data for Sequence-to...
Investigation of Text-to-Speech based Synthetic Parallel Data for Sequence-to...NU_I_TODALAB
 
Automatic speech recognition system using deep learning
Automatic speech recognition system using deep learningAutomatic speech recognition system using deep learning
Automatic speech recognition system using deep learningAnkan Dutta
 
IRJET - Language Linguist using Image Processing on Intelligent Transport Sys...
IRJET - Language Linguist using Image Processing on Intelligent Transport Sys...IRJET - Language Linguist using Image Processing on Intelligent Transport Sys...
IRJET - Language Linguist using Image Processing on Intelligent Transport Sys...IRJET Journal
 
EXTRACTIVE SUMMARIZATION WITH VERY DEEP PRETRAINED LANGUAGE MODEL
EXTRACTIVE SUMMARIZATION WITH VERY DEEP PRETRAINED LANGUAGE MODELEXTRACTIVE SUMMARIZATION WITH VERY DEEP PRETRAINED LANGUAGE MODEL
EXTRACTIVE SUMMARIZATION WITH VERY DEEP PRETRAINED LANGUAGE MODELijaia
 
IRJET - Analysis of Paraphrase Detection using NLP Techniques
IRJET - Analysis of Paraphrase Detection using NLP TechniquesIRJET - Analysis of Paraphrase Detection using NLP Techniques
IRJET - Analysis of Paraphrase Detection using NLP TechniquesIRJET Journal
 

What's hot (14)

CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...
CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...
CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...
 
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTK
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTKPUNJABI SPEECH SYNTHESIS SYSTEM USING HTK
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTK
 
5215ijcseit01
5215ijcseit015215ijcseit01
5215ijcseit01
 
Hindi digits recognition system on speech data collected in different natural...
Hindi digits recognition system on speech data collected in different natural...Hindi digits recognition system on speech data collected in different natural...
Hindi digits recognition system on speech data collected in different natural...
 
International journal of signal and image processing issues vol 2015 - no 1...
International journal of signal and image processing issues   vol 2015 - no 1...International journal of signal and image processing issues   vol 2015 - no 1...
International journal of signal and image processing issues vol 2015 - no 1...
 
Speaker recognition in android
Speaker recognition in androidSpeaker recognition in android
Speaker recognition in android
 
Bachelors project summary
Bachelors project summaryBachelors project summary
Bachelors project summary
 
A Survey on Speaker Recognition System
A Survey on Speaker Recognition SystemA Survey on Speaker Recognition System
A Survey on Speaker Recognition System
 
Extractive Summarization with Very Deep Pretrained Language Model
Extractive Summarization with Very Deep Pretrained Language ModelExtractive Summarization with Very Deep Pretrained Language Model
Extractive Summarization with Very Deep Pretrained Language Model
 
Investigation of Text-to-Speech based Synthetic Parallel Data for Sequence-to...
Investigation of Text-to-Speech based Synthetic Parallel Data for Sequence-to...Investigation of Text-to-Speech based Synthetic Parallel Data for Sequence-to...
Investigation of Text-to-Speech based Synthetic Parallel Data for Sequence-to...
 
Automatic speech recognition system using deep learning
Automatic speech recognition system using deep learningAutomatic speech recognition system using deep learning
Automatic speech recognition system using deep learning
 
IRJET - Language Linguist using Image Processing on Intelligent Transport Sys...
IRJET - Language Linguist using Image Processing on Intelligent Transport Sys...IRJET - Language Linguist using Image Processing on Intelligent Transport Sys...
IRJET - Language Linguist using Image Processing on Intelligent Transport Sys...
 
EXTRACTIVE SUMMARIZATION WITH VERY DEEP PRETRAINED LANGUAGE MODEL
EXTRACTIVE SUMMARIZATION WITH VERY DEEP PRETRAINED LANGUAGE MODELEXTRACTIVE SUMMARIZATION WITH VERY DEEP PRETRAINED LANGUAGE MODEL
EXTRACTIVE SUMMARIZATION WITH VERY DEEP PRETRAINED LANGUAGE MODEL
 
IRJET - Analysis of Paraphrase Detection using NLP Techniques
IRJET - Analysis of Paraphrase Detection using NLP TechniquesIRJET - Analysis of Paraphrase Detection using NLP Techniques
IRJET - Analysis of Paraphrase Detection using NLP Techniques
 

Similar to Modeling of Speech Synthesis of Standard Arabic Using an Expert System

SMATalk: Standard Malay Text to Speech Talk System
SMATalk: Standard Malay Text to Speech Talk SystemSMATalk: Standard Malay Text to Speech Talk System
SMATalk: Standard Malay Text to Speech Talk SystemCSCJournals
 
Robust Speech Recognition Technique using Mat lab
Robust Speech Recognition Technique using Mat labRobust Speech Recognition Technique using Mat lab
Robust Speech Recognition Technique using Mat labIRJET Journal
 
CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...
CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...
CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...AIRCC Publishing Corporation
 
Speech to text conversion for visually impaired person using µ law companding
Speech to text conversion for visually impaired person using µ law compandingSpeech to text conversion for visually impaired person using µ law companding
Speech to text conversion for visually impaired person using µ law compandingiosrjce
 
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARSYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARijcseit
 
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARSYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARijcseit
 
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARSYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARijcseit
 
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARSYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARijcseit
 
Survey On Speech Synthesis
Survey On Speech SynthesisSurvey On Speech Synthesis
Survey On Speech SynthesisCSCJournals
 
Acceptance Testing Of A Spoken Language Translation System
Acceptance Testing Of A Spoken Language Translation SystemAcceptance Testing Of A Spoken Language Translation System
Acceptance Testing Of A Spoken Language Translation SystemMichele Thomas
 
Approach To Build A Marathi Text-To-Speech System Using Concatenative Synthes...
Approach To Build A Marathi Text-To-Speech System Using Concatenative Synthes...Approach To Build A Marathi Text-To-Speech System Using Concatenative Synthes...
Approach To Build A Marathi Text-To-Speech System Using Concatenative Synthes...IJERA Editor
 
IRJET - Audio Emotion Analysis
IRJET - Audio Emotion AnalysisIRJET - Audio Emotion Analysis
IRJET - Audio Emotion AnalysisIRJET Journal
 
Approach of Syllable Based Unit Selection Text- To-Speech Synthesis System fo...
Approach of Syllable Based Unit Selection Text- To-Speech Synthesis System fo...Approach of Syllable Based Unit Selection Text- To-Speech Synthesis System fo...
Approach of Syllable Based Unit Selection Text- To-Speech Synthesis System fo...iosrjce
 
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTK
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTKPUNJABI SPEECH SYNTHESIS SYSTEM USING HTK
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTKijistjournal
 
Emotional telugu speech signals classification based on k nn classifier
Emotional telugu speech signals classification based on k nn classifierEmotional telugu speech signals classification based on k nn classifier
Emotional telugu speech signals classification based on k nn classifiereSAT Publishing House
 
Emotional telugu speech signals classification based on k nn classifier
Emotional telugu speech signals classification based on k nn classifierEmotional telugu speech signals classification based on k nn classifier
Emotional telugu speech signals classification based on k nn classifiereSAT Journals
 
A Review On Speech Feature Techniques And Classification Techniques
A Review On Speech Feature Techniques And Classification TechniquesA Review On Speech Feature Techniques And Classification Techniques
A Review On Speech Feature Techniques And Classification TechniquesNicole Heredia
 
Rendering Of Voice By Using Convolutional Neural Network And With The Help Of...
Rendering Of Voice By Using Convolutional Neural Network And With The Help Of...Rendering Of Voice By Using Convolutional Neural Network And With The Help Of...
Rendering Of Voice By Using Convolutional Neural Network And With The Help Of...IRJET Journal
 

Similar to Modeling of Speech Synthesis of Standard Arabic Using an Expert System (20)

SMATalk: Standard Malay Text to Speech Talk System
SMATalk: Standard Malay Text to Speech Talk SystemSMATalk: Standard Malay Text to Speech Talk System
SMATalk: Standard Malay Text to Speech Talk System
 
Robust Speech Recognition Technique using Mat lab
Robust Speech Recognition Technique using Mat labRobust Speech Recognition Technique using Mat lab
Robust Speech Recognition Technique using Mat lab
 
CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...
CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...
CURVELET BASED SPEECH RECOGNITION SYSTEM IN NOISY ENVIRONMENT: A STATISTICAL ...
 
H010625862
H010625862H010625862
H010625862
 
Speech to text conversion for visually impaired person using µ law companding
Speech to text conversion for visually impaired person using µ law compandingSpeech to text conversion for visually impaired person using µ law companding
Speech to text conversion for visually impaired person using µ law companding
 
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARSYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
 
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARSYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
 
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARSYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
 
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMARSYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
SYLLABLE-BASED SPEECH RECOGNITION SYSTEM FOR MYANMAR
 
Survey On Speech Synthesis
Survey On Speech SynthesisSurvey On Speech Synthesis
Survey On Speech Synthesis
 
Acceptance Testing Of A Spoken Language Translation System
Acceptance Testing Of A Spoken Language Translation SystemAcceptance Testing Of A Spoken Language Translation System
Acceptance Testing Of A Spoken Language Translation System
 
Approach To Build A Marathi Text-To-Speech System Using Concatenative Synthes...
Approach To Build A Marathi Text-To-Speech System Using Concatenative Synthes...Approach To Build A Marathi Text-To-Speech System Using Concatenative Synthes...
Approach To Build A Marathi Text-To-Speech System Using Concatenative Synthes...
 
IRJET - Audio Emotion Analysis
IRJET - Audio Emotion AnalysisIRJET - Audio Emotion Analysis
IRJET - Audio Emotion Analysis
 
Approach of Syllable Based Unit Selection Text- To-Speech Synthesis System fo...
Approach of Syllable Based Unit Selection Text- To-Speech Synthesis System fo...Approach of Syllable Based Unit Selection Text- To-Speech Synthesis System fo...
Approach of Syllable Based Unit Selection Text- To-Speech Synthesis System fo...
 
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTK
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTKPUNJABI SPEECH SYNTHESIS SYSTEM USING HTK
PUNJABI SPEECH SYNTHESIS SYSTEM USING HTK
 
Emotional telugu speech signals classification based on k nn classifier
Emotional telugu speech signals classification based on k nn classifierEmotional telugu speech signals classification based on k nn classifier
Emotional telugu speech signals classification based on k nn classifier
 
Emotional telugu speech signals classification based on k nn classifier
Emotional telugu speech signals classification based on k nn classifierEmotional telugu speech signals classification based on k nn classifier
Emotional telugu speech signals classification based on k nn classifier
 
A Review On Speech Feature Techniques And Classification Techniques
A Review On Speech Feature Techniques And Classification TechniquesA Review On Speech Feature Techniques And Classification Techniques
A Review On Speech Feature Techniques And Classification Techniques
 
K010416167
K010416167K010416167
K010416167
 
Rendering Of Voice By Using Convolutional Neural Network And With The Help Of...
Rendering Of Voice By Using Convolutional Neural Network And With The Help Of...Rendering Of Voice By Using Convolutional Neural Network And With The Help Of...
Rendering Of Voice By Using Convolutional Neural Network And With The Help Of...
 

Recently uploaded

APIForce Zurich 5 April Automation LPDG
APIForce Zurich 5 April  Automation LPDGAPIForce Zurich 5 April  Automation LPDG
APIForce Zurich 5 April Automation LPDGMarianaLemus7
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr LapshynFwdays
 
Story boards and shot lists for my a level piece
Story boards and shot lists for my a level pieceStory boards and shot lists for my a level piece
Story boards and shot lists for my a level piececharlottematthew16
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationRidwan Fadjar
 
My INSURER PTE LTD - Insurtech Innovation Award 2024
My INSURER PTE LTD - Insurtech Innovation Award 2024My INSURER PTE LTD - Insurtech Innovation Award 2024
My INSURER PTE LTD - Insurtech Innovation Award 2024The Digital Insurer
 
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024BookNet Canada
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsMark Billinghurst
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfAddepto
 
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Mark Simos
 
SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024Lorenzo Miniero
 
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek SchlawackFwdays
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfAlex Barbosa Coqueiro
 
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)Wonjun Hwang
 
Developer Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLDeveloper Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLScyllaDB
 
Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 3652toLead Limited
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...shyamraj55
 
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024BookNet Canada
 
DevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsDevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsSergiu Bodiu
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Patryk Bandurski
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Commit University
 

Recently uploaded (20)

APIForce Zurich 5 April Automation LPDG
APIForce Zurich 5 April  Automation LPDGAPIForce Zurich 5 April  Automation LPDG
APIForce Zurich 5 April Automation LPDG
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
 
Story boards and shot lists for my a level piece
Story boards and shot lists for my a level pieceStory boards and shot lists for my a level piece
Story boards and shot lists for my a level piece
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 Presentation
 
My INSURER PTE LTD - Insurtech Innovation Award 2024
My INSURER PTE LTD - Insurtech Innovation Award 2024My INSURER PTE LTD - Insurtech Innovation Award 2024
My INSURER PTE LTD - Insurtech Innovation Award 2024
 
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR Systems
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdf
 
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
 
SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024
 
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdf
 
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
 
Developer Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLDeveloper Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQL
 
Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
 
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
 
DevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsDevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platforms
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!
 

Modeling of Speech Synthesis of Standard Arabic Using an Expert System

  • 1. MODELING OF SPEECH SYNTHESIS OF STANDARD ARABIC USING AN EXPERT SYSTEM Tebbi Hanane1 and Azzoune Hamid2 1,2 LRIA, Option: Representation of Knowledge and systems of Inference, USTHB, Algiers 1 2 tebbi_hanane@yahoo.fr , azzoune@yahoo.fr ABSTRACT In this work we present our expert system of speech synthesis based on a text written in Standard Arabic, our work is carried out in two great stages: the creation of the sound data base, and the transformation of the written text into speech (Text To Speech TTS). This transformation is done firstly by a Phonetic Orthographical Transcription (POT) of any written Standard Arabic text that to transform it into his corresponding phonetics sequence, and secondly by the generation of the voice signal which corresponds to the chain transcribed. We spread out the different steps of conception of the system, as well as the results obtained compared to others manners of works studied to realize TTS based on Standard Arabic. KEYWORDS Engineering knowledge, modelling and representation of vocal knowledge, expert system, (TTS Text To Speech), Standard Arab, PRAAT. 1. INTRODUCTION Generate the voice is a complex work because of the variability intra and interlocutor of the voice signal. In computer science the difficulty of modeling the speech signal is for the reason that we don’t know yet how to model very well the enormous mass of knowledge and information useful to the signal synthesis. Thus we have made a choice to use an expert system to modeling that knowledge to build a robust system which can really read a text written in a language chosen especially in Standard Arabic. To obtain a better organization of our work, we defined our direct aims. We divided our modeling into three essential stages; the signal analysis, the phonetic orthographical transcription (POT) and finally the synthesis of the textual representation written in Standard Arabic. So the finality considered here is that the user can understand the different phrases transcribed and synthesized which will be pronounced with a clear and high quality manner. 2. PREVIOUS WORKS At the current time, we can judge that works emanated in the same context of ours are still not really colossal, and this is because of the complexity of the language itself. And if this some works exists, they are based on the same principle as transcribers of the other languages (French, David C. Wyld et al. (Eds) : CCSIT, SIPP, AISC, PDCTA, NLP - 2014 pp. 469–476, 2014. © CS & IT-CSCP 2014 DOI : 10.5121/csit.2014.4241
  • 2. 470 Computer Science & Information Technology (CS & IT) English, etc.) [1], nevertheless, the efforts undertaken are encouraging and open a large window to follow research task in this field. − TTS system MBROLA [2] which use the code SAMPA during the stage of transcription, in this case the user must respect the form of the SAMPA code which is not a universal code; − Work of S. BALOUL [3] who represents a concrete example of transcription of words; based on morphological analysis, and on the studies of pauses to generate the pronunciation of the texts. − SYAMSA (SYstème d’Analyse Morphosyntaxique de l’Arabe), realized by SAROH [4]. According to him, "the phonetisation of the Arabic language is based in particularly on the use of lexicons and on a morphological analyzer for the generation of the different forms of a word. In addition, they are the phenomena of interaction among the words (connection, elision, etc.) and the phenomena of assimilation which suggest the uses of phonological rules "[5]. This tool ensures for each word in entry, the root which correspond to it as well as the morphological and phonetic representations of the root. − The GHAZALI [6] project which was carried out within the IRSIT (Institut Régional des Sciences Informatiques et des Télécommunications) of Tunisia, it is based on a transcription work which fits inside the framework of the realization of a TTS system. The characteristic of this system is shown in the use of a set of rules in the emphasis propagation. − SYNTHAR+ [7] which is a tool studied by Z. ZEMIRLI within the NII (the National Institute of Informatics of Algiers), it ensures the transcription step for a TTS system so that it transmits the phonetic representation to the MULTIVOX synthesizer. It should be known that SYNTHAR+ is based on a morphological analysis before realizing the transcription. 3. STRUCTURE OF THE SYSTEM The uses of the concept of code and the introduction of high levels in analysis (morphological, syntactic, and pragmatic.) makes the transcription task so difficult and requires deepened studies of the language itself. The difference in our work compared to all that exists is in the transcription using graphemes, i.e. the uses of the Arabic characters as basic units to transcribe directly the text, indeed modeled by an expert system. Definition of an expert system: An expert system is an informatics tool of AI (Artificial Intelligence), conceived to simulate the knowhow of an human expert, in a precise and delimited field, this thanks to the good exploitation of a set of knowledge provided explicitly by expert of the field [5]. The E.S is a system where the data (the knowledge database) is quite separated from program which handles it (the inference engine). In our case, we built our expert system basing on the two following components: A knowledge database: It contains a database of facts and a database of rules, represents the knowledge (we speaks here about the sound database) and knowledge - to make (a set of
  • 3. Computer Science & Information Technology (CS & IT) 471 rewriting rules). The fact database integrate two types of facts: permanent facts of the field and deduced facts by the inference engine which are specific to the field considered (voice synthesis). The inference engine: It is capable to reason starting from the information contained in the knowledge database and to make deductions. The inference engine uses the facts and rules to produce new facts. Thus the role of an expert system is then to infer generally rules like if/then. To build this system our work targets the following steps: Figure 1. General schema of our system of speech synthesis 3.1. Acquisition It is the first step of our process of voice synthesis; it plays the role of the interface intermediate between the user and the system. During this time the user have to enter his set of phrases to be pronounced by the system. After that the automatic process can begin. 3.2. Preprocessing and standardization of the text In this phase any form of text is transformed to it literal, perfectly disambiguated form [1]. In particular, this module draft problems of formats, dates and hours, abbreviations, numbers, currencies, addresses, emails…etc. It was shown that majority errors of conversion grapheme/phoneme, for the best operational systems came from the proper names and the exceptions that they pose [8]. Here are some examples of the processing carried out in our case: − − − The replacement of each composed character by its equivalents Example: ‫ال‬ ‫ل ا‬ Consultation of the exceptions lexicon to eliminate the special words. The application of the transcription rules established for the language. This module must be able to process grammatical complexes tasks to identify the rules of transcription which will be used in a considered context.
  • 4. 472 Computer Science & Information Technology (CS & IT) 3.3. Analysis (the extraction of the characteristics) The aim of the analysis of the voice signal is to extracts the acoustic vectors which will be used in the stage of synthesis follows. In this step the voice signal is transformed into a sequence of acoustic vectors on the way to decrease redundancy and the amount of data to be processed. And then a spectral analysis by the discrete Fourier transform is performed on a signal frame (typically of size 20 or 30 ms) [9]. In this frame, the voice signal is considered to be sufficiently stable and we extract a vector of parameters considered to be enough for the good operation of the voice signal. In the speech synthesis, the characteristics extraction step, commonly known as the step of analysis, can be achieved in several ways. Indeed, the acoustic vectors are usually extracted using methods such as temporal encoding predictive linear (Linear Predictive Coding LPC) or Cepstrales methods as the MFCC encoding (Mel Frequency Cepstral Coding), as well as the process of segmentation, etc. This process delimits on the acoustic signal a set of segments characterized by labels belonging to the phonetics alphabet of the language under consideration. At present, the segmentation completely automatic of a voice signal remains a fastidious task. Indeed, looking at the complexity acoustico-phonetics phenomena being studied, this activity requires often a manual intervention. Generally, the methods that perform the segmentation of the acoustic wave are divided into two great classes: − The first include all the methods which allow segmenting a voice signal without a priori knowledge of the linguistics content of this signal. These methods split the voice signal into a set of zones homogeneous and stable; − The second class includes all the methods which allow segmenting the voice signal basing on a priori linguistic description (typically on phonemes) of this signal. These methods of segmentation are revealed like methods with linguistically constraints. In our state, we have opted for the second technique using a tool of voice signal analysis which is PRAAT [10] so to slice manually the speech signal in a succession of segments, each one associated with an element acoustic unit (phoneme or diaphone). Figure 2. Decomposition in polysound of the word [ ‫]جلس‬
  • 5. Computer Science & Information Technology (CS & IT) 473 3.4. Modeling of the sound database Majority of work curried out in the field of the spoken communication required often the recording, and the handling of corpuses of continuous speech, and that to curry out studies on the contextual effects, one the phonetic indices, and variability intra and inter-speaker. Our corpus is modeled by the diagram of class1 as follow: Figure 3. The diagram of class of our sound database 3.5. The training Here we referred to a POT module to convert each grapheme in phoneme according to the context and that is possible using a set of rewriting rules. The principal advantage of this approach is the possibility to modeling the human linguistic knowledge by a set of rules which can be integrated in expert systems. Each one of these rules has the following form: [Phoneme] = {LC (Left Context)} + {C (Character)} + Here is concrete example of transcription rules [11]: {RC (Right Context)} # is a beginning sign of the sentence, $ is a sign of end of the sentence, § is extremity of a word, C is a Consonant, V is a Vowel, SC is a Solar Consonant and LC is a Lunar Consonant [uu]={SC}+{ُ }+ {‫} و‬ [uu]= §+{ُ }+ { ‫} و‬ [uu]={LC}+{ُ }+ {‫} و‬ [uu]= $+{ُ }+ { ‫} و‬ When the " ‫ " و‬is preceded by the vowel / ُ / obtain the long vowel [ uu ]. and followed by a consonant, we 1: It is a diagram used in UML(Unified Modeling Language) which is used in object oriented modeling
  • 6. 474 Computer Science & Information Technology (CS & IT) When the " ‫" و‬ vowel [ uu ]. is preceded by the vowel / ُ / in final position, we obtain the long 3.6. The synthesizer (the voice generation) The generation of the voice signal is the real synthesis of speech. This operation consist of a transform of the phonetic sequence (which represents the pronunciation of the written text) resulting from the transcription step to its substance, i.e. to it acoustics realization. During this step we choose from our sound database the units (phonemes, diphones) most suitable to build the sentence to generate, this means that we will create a automatic function of reading, so that in the end, the user has only to listen to the synthetic sentences. To improve the quality of the generated synthesis, we increase each time the sound unit only. The general silhouette of our expert system can be as follow: Figure 4. General diagram of our expert system 4. TESTS AND RESULTS To test the performances of our TTS system based on Standard Arabic language, we have chosen a set of sentences which we judged like reference since they contain almost the different possible combinations specific to the language itself. To calculate the success rate (SR) associated with each sentence tested; we got the following formula: SR = Numberofphrasewellprononced *100% Numberofphrasestested The system present in general a SR of 96 % for the set of the sentences tested. Results obtained are summarized by the following table:
  • 7. Computer Science & Information Technology (CS & IT) 475 Table 1. Rate of success for a sample of selected sentences Majority content Short vowels Long vowels Solar consonants Lunar consonants Isolated words Sentences Numbers Exception Words POT 100% 100% 100% 100% 100% 100% 90% 100% Synthesis 95% 95% 97% 95% 80% 75% 95% / by Phonemes Synthesis by Diphones / / / / 90% 85% 100% / 5. CONCLUSION AND PROSPECTIVE A system for synthesizing Arabic speech has been developed based on an expert system which uses a set of subphonetic elements as the synthesis units to allow synthesis of limited-vocabulary speech of good quality. The synthesis units have been defined after a careful study of the phonetic properties of modern Standard Arabic, and they consist of central steady-state portions of vowels, central steady-state portions of consonants, vowel-consonant and consonant-vowel transitions, and some allophones. A text-to-speech system which uses this method has also been explored. The input of the system is usual Arabic spelling with diacritics and/or simple numeric expressions. Synthesis is controlled by several text-to-speech rules within the rule database of the expert system, which are formulated and developed as algorithms more suited for computer handling of the synthesis process. The rules are required for converting the input text into phonemes, converting the phonemes into phonetic units, generating the synthesis units from the phonetic units, and concatenating the synthesis units to form spoken messages. The suitability of the method for synthesizing Arabic has been shown by realizing all its functions on a personal computer and by conducting understandability test on synthesized speech. So, we have detailed the different components which represents the basic blocks of our TTS system based on a written text in Standard Arabic and our modeling of it with the use of an expert system, this tool which is a fruits of the artificial intelligence remain less used in the field of automatic speech processing. This fact encouraged us to explore this world trying to give a little push to the research done in this multidisciplinary field. Like prospective to our work we show the following points: − − − Improvement of the quality of voice generated by using methods of modification of the prosody; The use of others techniques of synthesis such as the Synthesis using the unit’s selection; Make, if possible, the signal segmentation totally automatic. REFERENCES [1] [2] [3] [4] [5] [6] Tebbi Hanane, « Transcription orthographique phonétique en vue de la synthèse de la parole a partir du texte de l’Arabe », Mémoire de magister en Informatique, Blida, Algérie, 2007 http//:tcts.fpms.ac.be/ synthesis/ Mbrola.html. Baloul, « Développement d’un système automatique de synthèse de la parole à partir du texte arabe standard voyellé ». Thèse de doctorat d’université, Le Mans, France, 27 Mai 2003. SAROH Abdelghani, Base de données lexicales dans un système d'analyse morpho-syntaxique de l'arabe : SYAMSA, Toulouse 3, 1989. Guerti Mhania, Tebbi Hanane, La Conversion Graphèmes Phonèmes En Vue D'une Lecture Automatique de Textes en Arabe Standard, LANIA, Chlef, Algérie, 2007 S. Dichy J. & Hassoun M.O. (1998), Some aspects of the DIINAR-MBC research programme, in Ubaydly A., ed.,1998: 2.8.1-5.
  • 8. 476 [7] Computer Science & Information Technology (CS & IT) SYNTHAR+: Arabic speech synthesis under multivox. (SYNTHAR+: Synthèse vocale arabe sous Multivox.) (English) RAIRO, Tech. Sci. Inf. 17, No. 6, 741-761 (1998). [8] Mehdi Braham, Oumaya Abbes, Maroua Trabelsi, Mehdi Dahmen, Rania Nouaari, « Intelligence artificielle : Diagnostic par système expert, Modèle d’analyse d’acteurs : MACTOR » [9] P. Boula de Mareüil, « Synthèse de la parole à partir de courriers et évaluation de la conversion graphème-phonème ». LIMSI-CNRS http://www.limsi.fr/ Individu/ mareuil/ [10] Paul Boersma and David Weenink, "Praat: doing phonetics by computer" Phonetic Sciences, University of Amsterdam Spuistraat 210, 1012VT Amsterdam The Netherlands. [11] T. Saidane, M. Zrigui, & M. Ben Ahmed, « La transcription orthographique phonétique de la langue Arabe ». RECITAL 2004, Fès, 19-22 Avril 2004.