In this paper, we present the unit selection
based concatenative text-to-speech synthesis. Here the unit
of selection is syllable. The paper presents system for text
to speech synthesis for Marathi language. This paper gives
complete idea that how to convert Marathi text into speech
right from text processing to audio processing. The
proposed system requires audio data base and very limited
text data base. The TTS system has been developed on
Matlab® R2014a. Matlab® is Unicode software therefore
UTF-8 encoding has used to read the Marathi text.
Darshna Badhe : Digital Systems, Rajarshi Shahu College of Engineering, Savitribai Phule Pune University,
Pune, Maharashtra-41033, India
P. M. Ghate : Digital Systems, Rajarshi Shahu College of Engineering, Savitribai Phule Pune University,
Pune, Maharashtra-41033, India
Text To Speech
Unicode
transliteration
Syllabification
Structure
UTF-8(universal characterset
transformation function 8 bit)
It can be concluded from the results of the MOS test for
checking the intelligibility, shows that around 86%
accuracy was reported. However in terms of naturalness
the level of accuracy was around 56%, which was lower,
but acceptable. Hence prosody generation is the desired
future scope from the study to come closer to higher
level of naturalness, which cannot be calculated with
this system. The main advantage of the proposed system
is that it requires very less text data base. Further the
memory required for audio database is trade-off
amongdiphone, phoneme, and words as a unit of
selection. Directions for future work: The prosody
generation is an important module in TTS for increasing
the naturalness of the output of the speech.
[1] Mr.S.D.Shirbahadurkar, “Marathi Language Speech
Synthesizer Using Concatenative Synthesis Strategy
(Spoken in Maharashtra, India)”, Second IEEE
International Conference on Machine Vision 2009.
[2] Mrs. Madhavi R. Repe, “Natural Prosody Generation
in TTS for Marathi Speech Signal”, IEEE International Conference on Signal Acquisition and
Processing 2010.
[3] Mrs. Madhavi R. Repe,” Prosody Model for Marathi
Language TTS Synthesis with Unit Search and
Selection Speech Database”, IEEE International
Conference on Recent Trends in information,
Telecommunication and Computing 2010
[4] H. Segi, R. Takou, N. Seiyama and T. Takagi, “An
automatic broadcast system for a weather report
radio program”, IEEE Trans. on broadcasting, vol.
59, no 3, September 2013.
[5] Shruti Gupta, “Comparative study of text to speech
system for Indian language”,International Journal Of
Advances In Computing And Information
Technology
[6] Shruti Gupta, “Hindi Text To Speech System”,
Computer Science And Engineering Department
Thapar University Patiala June 2012
[7] Tapas Kumar Patra, “Text to Speech Conversion
with Phonetic Concatenation”, International Journal
of Electronics Communication and Computer
Technology (IJECCT) Volume 2 Issue 5 (September
2012)
[8] MrsMinaksheepatil, “ Syllable” Concatenation for
Text to Speech Synthesis for Devnagari Script”,
International Journal of Advanced Research in
Computer Science and Software Engineering,
Volume 2, Issue 9, September 2012
[9] Shreekanth.T, “ An Unit Selection based Hindi Text
To Speech Synthesis System Using Syllable as a
Basic Unit”, IOSR Journal of VLSI and Signal
Processing (IOSR-JVSP) Volume 4, Issue 4, Ver. II
(Jul-Aug. 2014)
[10] Mr. S. B. Chaudhari, “ A Review on Multilingual
Text to Speech Synthesis by Syllabifying the
Words of Devanagari and Roman”, International
Journal on Recent and Innovation Trends in
Computing and Communication
ISSN: 2321-8169,Volume: 2 Issue: 11
[11] Snehali. K. Nandurkar, ZakirM.Shaikh, “Speech
Generation Of Transliterated Hindi
Text”,International Journal of Application or
Innovation in Engineering & Management
(IJAIEM), Volume 3, Issue 10, October 2014 ISSN
2319 - 4847
[12] Hiroyuki Segi, “An Automatic Broadcast System for
a Weather Report Radio Program”, IEEE
Transactions on Broadcasting, Vol. 59, No. 3,
September 2013
