VoxForge
VoxForge
I created the wordlist with UTF-8 encoding. That part worked fine and acoustic model get created. Then I try to test it sampleapp but it didn't identified the charater encoding. So it displays some junk characters rather than the actual representation. Is there a way to overcome this problem?
Hi kmaleen,
Just to clarify...
>I created the wordlist with UTF-8 encoding.
You used HTK Perl script prompts2wlist to create a word list (wlist) file?
>That part worked fine and acoustic model get created.
You created an HTK acoustic model?
>Then I try to test it sampleapp but it didn't identified the charater
>encoding.
You are trying to test the acoustic model with Julius or HTK? Which sampleapp are you talking about?
Ken
Hi kmaclean,
No I didn't use the prompts2wlist script. I just created a small word list manually.
>You created an HTK acoustic model?
Yes. I followed voxforge How-To tutorial.
>You are trying to test the acoustic model with Julius or HTK? Which sampleapp are you talking about?
I tested the model with Julius. In Julius there is a simple GUI application as sampleapp. I tried that application. It didn't display the characters. I also tried the command line application. It also didn't work.
I hope these information would give clarification about my problem.
Thanks