[hts-users:00568] Re: speech segmentation

["Heiga ZEN (Byung Ha CHUN)" <zen@xxxxxxxxxxxxxxxx>]

Hi,

Tamer Fares wrote:

> i need to make speech segmentation and assign to each segment label name 
> in order to specify the duration starting and ending for each phoneme  
>
> i had a certain program where i found that there is an HTK transcription 
> (this mean that this done manually by hearing each utterance and 
> segments all possible phonemes)
>
> or there exist certain another way that determine this automatically

Do you have HMMs for your target language?

If yes, run HVite with -a option using these HMMs.

If no, you can use flat-start training (start from HERest phase).
It does not require phone boundaries.

For both cases, you need the corresponding phoneme sequence for each utterance.
Please refer HTKBook section 3.2 for detail.

Regards,

Heiga ZEN (Byung Ha CHUN)

--
------------------------------------------------
Heiga ZEN     (in Japanese pronunciation)
Byung Ha CHUN (in Korean pronunciation)

Department of Computer Science and Engineering
Nagoya Institute of Technology
Gokiso-cho, Showa-ku, Nagoya 466-8555 Japan

http://www.sp.nitech.ac.jp/~zen
------------------------------------------------