MILE TTS for Tamil and Kannada for blizzard challenge 2013

We participated in the Blizzard Challenge 2013 for Tamil and Kannada, using our unit selection based concatenative speech synthesis system. Sentence level viterbi search is used to select the reliable speech units among a set of candidate units. The same Text-to-Speech synthesis framework is used to synthesize speech in both the Indian languages. The given Wikipedia and semantically unpredictable test sentences are synthesized using IIIT-H Indic corpus and the listening test results reported by the blizzard evaluation team is discussed. The letter code for MILE TTS is “R”.

[1]  Alan W. Black,et al.  Unit selection in a concatenative speech synthesis system using a large speech database , 1996, 1996 IEEE International Conference on Acoustics, Speech, and Signal Processing Conference Proceedings.

[2]  P Arulmozhi,et al.  Continuity metric for unit selection based text-to-speech synthesis , 2010, 2010 International Conference on Signal Processing and Communications (SPCOM).