I have initialized decoder with keyword search, fed 48128 samples to it until hypotesis appeared.
Information returned by ps_seg_iter shows that recognized word is at frames 887-940.
Frame rate is 100, sample rate is 8000. I want to get last sample of hypotesis.
940 / 100 * 8000 = 75200, which is more that number of samples fed. What do I do wrong?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
ps_seg_framesreturns offsets of word start and end frames. But what is the base of these offsets? It seems it is not utterance start.It is a start of the continuous stream. You can restart the stream with
ps_start_streamcall.I have initialized decoder with keyword search, fed 48128 samples to it until hypotesis appeared.
Information returned by
ps_seg_itershows that recognized word is at frames 887-940.Frame rate is 100, sample rate is 8000. I want to get last sample of hypotesis.
940 / 100 * 8000 = 75200, which is more that number of samples fed. What do I do wrong?
Maybe you didn't set samprate properly, you need to provide more information on what your situation is.