Acoustic Model Discussions

Flat
Bug in some of the audio files
User: ecret
Date: 3/5/2009 4:36 pm
Views: 5440
Rating: 3

Well not inherently a bug but it makes scripting more difficult.


Some of the audio files have a prompt file, not a PROMPT as 95% of the others do.  They also have a different format inside prompt,

e.g:

in prompts:

filename this is a transcription

 

in PROMPTS:

folder/mfc/filename this is a transcription

 

So my fix  would be to make the prompts renamed as PROMPTS and change it to

folder/mfc/filename this is a transcription

 

I am making those changes on my local system. Should I upload these changes as well?

--- (Edited on 3/5/2009 4:36 pm [GMT-0600] by ecret) ---

Re: Bug in some of the audio files
User: ecret
Date: 3/5/2009 4:37 pm
Views: 183
Rating: 3

if i was given svn access I could do the fix

--- (Edited on 3/5/2009 4:37 pm [GMT-0600] by ecret) ---

Re: Bug in some of the audio files
User: kmaclean
Date: 3/6/2009 8:53 pm
Views: 81
Rating: 2

Hi ecret,

>if i was given svn access I could do the fix

Done.

Please don't check everything out (only check out the files you are updating)... the repository contains "329,470 items, totalling 26.2 GB" and resides on the same server as VoxForge.

Thanks for your help on this!

Ken

--- (Edited on 3/6/2009 9:53 pm [GMT-0500] by kmaclean) ---

Re: Bug in some of the audio files
User: ecret
Date: 3/6/2009 11:56 pm
Views: 2257
Rating: 3

We got to figure out a way to sneak that 29 gigs onto some public svn :)

 

I will have it fixed in a few days.  I just butchered my /wav/ folder to get it working for sphinxtrain and need to start from the start to resolve the bugs.

--- (Edited on 3/6/2009 11:56 pm [GMT-0600] by ecret) ---

PreviousNext