Hi,
I'm building a speech server using Sphinx3 API, but i can't make it recognize speech in raw format. I've taken livedecode as example. I've build and test spanish acustic and language models with SphinxTrain.
The program run fine but it doesn't recognize correctly any speech sample.
It looks like you incorrectly extracted the features during the model training and as a result the model is not usable. Check the make_feats once again. Make sure you are using proper wav files, that you have enough data and that you are using proper arguments.
Also, your program crashes in the middle because of
strcpy(hypstr,"");
The hypstr is obvously undefined, you can't copy the data there.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
I'm building a speech server using Sphinx3 API, but i can't make it recognize speech in raw format. I've taken livedecode as example. I've build and test spanish acustic and language models with SphinxTrain.
The program run fine but it doesn't recognize correctly any speech sample.
The small program, the models and a couple of test are at
http://www.mediafire.com/file/2lmfjy2hyiv/simple.zip
I'm using sox to convert the audio to raw 16bites 16kHz mono, encode signed and unsigned (i've tested both).
Thansk in advance,
Andres
It looks like you incorrectly extracted the features during the model training and as a result the model is not usable. Check the make_feats once again. Make sure you are using proper wav files, that you have enough data and that you are using proper arguments.
Also, your program crashes in the middle because of
The hypstr is obvously undefined, you can't copy the data there.