Hi Sergey,

>Is there any time estimate when Sequitur G2P models will be trained with the

>VoxForge Pronunciation dictionary?

Here they are

Ken

--- (Edited on 5/1/2008 11:07 am [GMT-0400] by kmaclean) ---

Hi Ken,

That is a great news!

Is there any tutorial or readme file on how to apply it to Automated Audio Segmentation?

Thanks,

Sergey

--- (Edited on 5/1/2008 10:37 am [GMT-0500] by Sergey) ---

Hi Sergey,

>Is there any tutorial or readme file on how to apply it to Automated Audio

>Segmentation?

Note: G2P will only help us to automatically create pronunciations for words that are not in the VoxForge pronunciation dictionary (which is based on CMUDitc version 0.6).

From G2P the README file: 

4. Transcribe new words.
   Prepare a list of words you want to transcribe as a simple text
   file words.txt with one word per line (and no phonemic
   transcription), then type:
       g2p.py --model model-3 --apply words.txt

I am now working on a script to try to automate the entire segmentation process, using G2P and the models I created. 

Ken 

--- (Edited on 5/1/2008 12:27 pm [GMT-0400] by kmaclean) ---

Hi Ken,

> ...g2p.py --model model-3 --apply words.txt

Thanks. I will try it.

> I am now working on a script to try to automate the entire segmentation process...

That would be fantastic!

Thanks,

--Sergey

--- (Edited on 5/1/2008 3:21 pm [GMT-0500] by Sergey) ---