From: Rademan, C. M. <156...@su...> - 2014-07-28 07:31:37
|
To whom it may concern, I would like to use Kaldi to transcribe speech into text. But I would like to include additional information, that being the time stamped phonemes pronounced in each spoken word. I am unaware of the readers knowledge of the TIMIT database, but I would like to create a similar, but much smaller version, of such a data set. The TIMIT database has recordings of speech which include text files with timing information on phonetic pronunciation during speech. Is it possible to create a similar output using Kaldi? Kind regards, Christo Rademan |