[asterisk-dev] [Code Review] Change sample slinear frame to contain non-zero data so that translation calculations for speex works

Jeff Peeler jpeeler at digium.com
Wed Jun 2 11:24:50 CDT 2010

This is an automatically generated e-mail. To reply, visit:

(Updated 2010-06-02 11:24:50.896040)

Review request for Asterisk Developers.


Created hex dump from sound prompt as requested. The old data from slin_speex_ex.h present in 1.6.1 would have probably worked fine, but this replaces what is present with the same amount of data.


This is literally a %s/0x00/0x11/g for slin.h. The problem is that the raw data used to generate sample frames was changed to contain all zeros. When codecs.conf is configured to use both voice activity detection and preprocessing, the frames are not being returned (in lintospeex_frameout) causing the encoder to keep on trying to get more samples forever. I didn't see a good way to improve the code to be able to generate samples correctly with this "silent" data, so I changed the sample frame to contain data so voice frames were detected and passed back up.

This addresses bug 17240.

Diffs (updated)

  /trunk/include/asterisk/slin.h 266924 

Diff: https://reviewboard.asterisk.org/r/682/diff


I tried a call using gsm (just because it's easier, also uses the data from slin.h) and it worked as expected.



More information about the asterisk-dev mailing list