Data to download

Click here to download the parallel ASCII(Roman) transliteration data for three Indian languages.

For demo, please follow below link:
Indic-search: http://srikanthr.in/indic-search


This page contains following samples:

  1. Natural: Originally recorded voice.
  2. Vocoded: Use STRAIGHT to extract 60-dimensional Mel-Cepstral Coefficients (MCCs), 25 band aperiodicities (BAPs) and fundamental frequency (F0), and then pass these parameters directly to STRAIGHT-vocoder to reconstract waveform.
  3. Uni-Grapheme: Synthesized voice from informal transliteration using graphemes.
  4. Multi-Grapheme: Synthesized voice from informal transliteration using multi-grapheme model.
  5. G2P: Synthesized voice from informal transliteration using G2P model.
  6. Benchmark: Synthesized voice from automated transliterated text(from native text) using common phone set notation.



Date: 10/04/2015
Text-To-Speech Synthesis
(Database: Telugu)


Natural Vocoded Uni-Grapheme Multi-Grapheme G2P Benchmark

Links :