search

alumae / kaldi-gstreamer-server

Real-time full-duplex speech recognition server, based on the Kaldi toolkit and the GStreamer framwork.
BSD 2-Clause "Simplified" License
1.07k stars 342 forks source link

How to use 8khz sampling frequency trained chain TDNN model #193

Closed boleamol closed 5 years ago

boleamol commented 5 years ago

Now, we are working on Telephonic speech, which is having sampling frequency 8khz. If we want chain TDNN model which is trained on 8khz model then what are the changes I have to do?

My Target is live speech decoding using my model..

boleamol commented 5 years ago

I changed the sampling frequency in kaldi/src/feat/pitch-functions.h , then kaldi/src/gst-plugin and kaldi/src/onlinebin and recompiled kaldi once again. Now it's working with 8khz audio. Also used "nnet-mode : 3" in .yaml file. Thank @alumae for your gstreamer.