Hi, I'm working on making training models for Sphinx with SphinxTrain. I'm running make_feats.pl on my wav file, and it's been running for at least an hour. Is this normal?
The file is 15 MB and consists of about three minutes of audio.
Thanks,
Rachel
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
It seems to find the audiofile ok, it is when it is running the wave2feat.c program that it seems to hang. There aren't any errors that occur that I can see - it just seems to take forever and never return. This is the last thing the script says:
Hi, I'm working on making training models for Sphinx with SphinxTrain. I'm running make_feats.pl on my wav file, and it's been running for at least an hour. Is this normal?
The file is 15 MB and consists of about three minutes of audio.
Thanks,
Rachel
so I think my problem was that i put wav files in the wav directory, but it looks like i need SPHERE or raw format.
Wrong path of the audio? -a
It seems to find the audiofile ok, it is when it is running the wave2feat.c program that it seems to hang. There aren't any errors that occur that I can see - it just seems to take forever and never return. This is the last thing the script says:
INFO: wave2feat.c(139): wav/show24_supershort.wav.sph
Are there any sample wav files out there that are known to work that I could try for comparison?
I see,
Try to check out our recommended training setup.
http://www.speech.cs.cmu.edu/sphinx/tutorial.html
Use the an4 setup, you will be able to find files in sph format. This should allow you to do some comparison.
If you still couldn't figure out, post us again. In the Bugs page, also check-in one sample utterance. We will be able to fix it in short time.
-gj
Thanks gj, I will try that.
Rachel