3

I'm using pocketsphinx python bindings on ubuntu 11.10. I pass it a grammar and an audio file and it works perfectly. I am now looking to find the timestamp of each word in the audio file. I have seen:

void ps_seg_frames(ps_seg_t *seg, int *out_sf, int *out_ef);

which I believe should do what I need. However I can't seem to find away to do this in python. Is this possible?

Richard Mosse
  • 560
  • 7
  • 18

1 Answers1

3

This feature is not implemented yet. There was a patch here but it was lost due to sourceforge outage it seems

https://sourceforge.net/tracker/?func=detail&atid=301904&aid=3480256&group_id=1904

It should be relatively easy to add this binding

Nikolay Shmyrev
  • 24,897
  • 5
  • 43
  • 87
  • How might one add this feature? – Richard Mosse Feb 22 '12 at 20:15
  • 1
    Hello. The patch to implement this feature has been just committed to subversion repository. Just download and compile the snapshot to have this functionality. See pocketsphinx/python/ps_test_seg.py for details on how to use it. – Nikolay Shmyrev Feb 24 '12 at 00:40
  • What exactly needs to be downloaded to use this functionality? I don't see this file on the cmusphinx repo (https://github.com/cmusphinx/pocketsphinx), I only see it on a mirror (https://github.com/skerit/cmusphinx/blob/master/pocketsphinx/python/ps_test_seg.py) – Adam_G Jun 21 '16 at 21:32
  • This code is not relevant anymore in latest version – Nikolay Shmyrev Jun 22 '16 at 00:02