Juha Heinanen wrote on 08/28/2014 02:34 PM:
> I changed in opus.c channels param from 1 to 2:
> 
>   BEGIN_PAYLOADS
>     PAYLOAD( -1, "opus", _OPUS_RATE, 48000, 2, CODEC_OPUS, 
> AMCI_PT_AUDIO_FRAME )
>   END_PAYLOAD
> 
> That caused sems to accept offer
> 
> a=rtpmap:96 opus/48000/2.
> a=fmtp:96 stereo=1;sprop-stereo=1.
> 
> and generate same in reply.
> 
> Then I did both playback and recording tests and didn't notice any audio
> quality problems.
those two don't use the receive (playout) buffer, thus you don't see
the issue there.

> 
> What kind of tests show bad audio quality if the above change is made?
try calling into the echo or conference app from e.g. baresip. What
you'll see is that the PLC functions are called all the time because
there's audio 'missing'.

But you're right it's probably not really the number of channels that
trick sems (they're ignored in AmAudioFormat anyway) but somehow
coincidence that it worked for me with channels=1.

My analysis so far is that when reading from the receive/playout
buffer, too much is read due to frame size miscalculations and thus
PLC functions are called which actually overwrite proper received audio.

Stefan

_______________________________________________
Semsdev mailing list
[email protected]
http://lists.iptel.org/mailman/listinfo/semsdev

Reply via email to