0

I'm using Kaldi and I have a problem, the transcription of my audio file stops after 7 seconds. There is no error returned by kaldi. An idea of why?

Here is my command line:

sudo /home/kaldi/src/online2bin/online2-wav-nnet3-latgen-faster   --online=false   --verbose=3   --do-endpointing=false   --frame-subsampling-factor=3   --config=exp/tdnn_7b_chain_online/conf/online.conf   --max-active=17000   --beam=15.0   --lattice-beam=6.0   --acoustic-scale=1.0   --word-symbol-table=exp/tdnn_7b_chain_online/graph_pp/words.txt   exp/tdnn_7b_chain_online/final.mdl   exp/tdnn_7b_chain_online/graph_pp/HCLG.fst   'ark:echo utterance-id1 utterance-id1|'   'scp:echo utterance-id1 /home/test-audio/batman.wav|'   'ark:/dev/null'

Here is my audio file: http://www.wavsource.com/snds_2018-06-03_5106726768923853/tv/batman/ordinary.wav

Thank you !

antoine grandiere
  • 1
  • 1

1 Answers1

0

Kaldi automatically stops on silence, you can not change it with this particular binary. You can use gstreamer server to process the whole file or you can use Kaldi API directly. Or you can split file on chunks and process each one individually.

Nikolay Shmyrev
  • 24,897
  • 5
  • 43
  • 87