ISCA Archive WSLP 2003
ISCA Archive WSLP 2003

Comparative analysis of Hindi retroflex and dental CVsyllables and their synthesis

Rajesh Verma, Puneet Chawla

This paper describes in detail the analysis results of the Hindi Retroflex consonants /t./, /t.h/, /d./ & /d.h/ and the Dental consonants /t/, /th/, /d/ and /dh/ analyzed by using PC based Sensimetrics Speech Station Software. These sounds were analyzed in five long vowel contexts /a/, /i/, /u/, /e/ and /o/ for a very accurate description of their acoustic characteristics/features and the differences between the corresponding cognate sounds in the two classes. Various parameters like duration of closure/voice bar, duration of burst, voice onset time, duration of aspiration, rate of second formant transition and burst frequencies and amplitudes have been studied in details.

The analyzed data was further used to generate the synthetic CV syllables using a cascade/ parallel formant synthesizer simulated on a PC. For the synthesis purpose, the source and vocal tract parameters of the synthesizer configuration were selected very carefully. Special attention was paid to the parameters like formant frequencies and their relative amplitudes, which play an important role in making distinction between cognate sounds like /t/ and /t./. The overall burst amplitude also plays a crucial role to make clear distinction in dental and retroflex cognate sounds.

The parametric doc files were modified iteratively, until a satisfactory quality of synthetic sound was obtained. The quality of synthetic speech was evaluated not only by subjective listening but also by matching the spectra of synthetic speech with original speech.

Cite as: Verma, R., Chawla, P. (2003) Comparative analysis of Hindi retroflex and dental CVsyllables and their synthesis. Proc. Workshop on Spoken Language Processing, 163-167

  author={Rajesh Verma and Puneet Chawla},
  title={{Comparative analysis of Hindi retroflex and dental CVsyllables and their synthesis}},
  booktitle={Proc. Workshop on Spoken Language Processing},