input wave file sampling rate 8kHz
GoogleCodeExporter opened this issue · 7 comments
GoogleCodeExporter commented
What steps will reproduce the problem?
1. Run vid -i RIFF-little-endian-16bit-8kh-wave-file.wav
Error: Traceback (most recent call last):
File "/usr/local/bin/vid", line 132, in <module>
thrd_n=multiprocessing.cpu_count() * 5)
File "/usr/local/lib/python2.7/dist-packages/voiceid/sr.py", line 899, in extract_speakers
self._to_wav()
File "/usr/local/lib/python2.7/dist-packages/voiceid/sr.py", line 605, in _to_wav
fm.file2wav(self.get_filename())
File "/usr/local/lib/python2.7/dist-packages/voiceid/fm.py", line 93, in file2wav
+ "wavenc ! filesink location=" + name + ".wav ")
File "/usr/local/lib/python2.7/dist-packages/voiceid/utils.py", line 67, in start_subprocess
raise err
OSError: Subprocess <subprocess.Popen object at 0xb7490c2c> closed unexpectedly
[gst-launch filesrc location='/opt/resources/silence.wav' ! decodebin !
audioresample ! 'audio/x-raw-int,rate=16000' ! audioconvert !
'audio/x-raw-int,rate=16000,depth=16,signed=true,channels=1' !wavenc ! filesink
location= RIFF-little-endian-16bit-8kh-wave-file.wav]
What is the expected output? What do you see instead?
Expected: It should work with 8kHz sampled wave files as well. Or atleast give
an exception that the file doesn't match the sampling rate.. and exits.
Actual: it gives an exception, doesn't exit cleanly and truncates the input
wave file!
What version of the product are you using? On what operating system?
Revision: 189. URL: http://voiceid.googlecode.com/svn/trunk
Ubuntu 12.04 32bit wheezy/sid
Original issue reported on code.google.com by Millenni...@gmail.com on 2 Aug 2012 at 7:41
GoogleCodeExporter commented
It's a bad thing :)
We knew this problem but never had the need to fix, until your request!
Now with the revision 192 it should be fixed.
Let me know if it's ok.
Original comment by maurome...@gmail.com on 2 Aug 2012 at 9:46
- Changed state: Fixed
- Added labels: Priority-High
- Removed labels: Priority-Medium
GoogleCodeExporter commented
It doesn't truncate the input wave file now. Rest is the same. When would
voiceid be able to handle 8kHz files? Since LIUM_SpkDiarization framework uses
8kHz as well as 16kHz, voiceid (built on top of LIUM_SpkDiarization), should be
able to support 8kHz sampled files.
Original comment by Millenni...@gmail.com on 2 Aug 2012 at 9:55
GoogleCodeExporter commented
Can you attach a 8kHz file to try, with some 8kHz waves I tried it works.
Original comment by maurome...@gmail.com on 2 Aug 2012 at 10:00
- Changed state: Started
GoogleCodeExporter commented
8kHz sample file attached.
Original comment by Millenni...@gmail.com on 2 Aug 2012 at 10:07
Attachments:
GoogleCodeExporter commented
Well I am at revision 192. And here is what I get:
converting_file
Traceback (most recent call last):
File "/usr/local/bin/vid", line 132, in <module>
thrd_n=multiprocessing.cpu_count() * 5)
File "/usr/local/lib/python2.7/dist-packages/voiceid/sr.py", line 899, in extract_speakers
self._to_wav()
File "/usr/local/lib/python2.7/dist-packages/voiceid/sr.py", line 605, in _to_wav
fm.file2wav(self.get_filename())
File "/usr/local/lib/python2.7/dist-packages/voiceid/fm.py", line 93, in file2wav
+ "wavenc ! filesink location=" + name + ".wav ")
File "/usr/local/lib/python2.7/dist-packages/voiceid/utils.py", line 67, in start_subprocess
raise err
OSError: Subprocess <subprocess.Popen object at 0xb7404c2c> closed unexpectedly
[gst-launch filesrc location='sample-8kHz.wav' ! decodebin ! audioresample !
'audio/x-raw-int,rate=16000' ! audioconvert !
'audio/x-raw-int,rate=16000,depth=16,signed=true,channels=1' !wavenc ! filesink
location=sample-8kHz.wav ]
Original comment by Millenni...@gmail.com on 3 Aug 2012 at 5:20
GoogleCodeExporter commented
For what I see in the traceback, you are not using the revision 192 but still
the revision 189, because the line numbers don't match the code in that
revision.
For example in line 605 of sr.py (rev 192) there is not what traceback says but
you can find that very line in the revision 189.
You have to install the new version after svn up and be sure to have the right
revision by looking at the installed files in
/usr/local/lib/python2.7/dist-packages/voiceid/
Original comment by maurome...@gmail.com on 3 Aug 2012 at 8:41
GoogleCodeExporter commented
For some reason it was giving an error on installing the new version. I
recompiled my python installation and it worked fine :)
Original comment by Millenni...@gmail.com on 7 Aug 2012 at 7:47