GitLab

Then, successfully unpacketized NAL units will be passed to the decoder

That did not take into account the bitrate target given by Mediastreamer

Indeed, it looks that, and it seems generally admitted, that there are 1000ms in one second.

In order to be able to apply equlization at once on the mouth output and
the speaker input, two MSEqualizer filters are instancited at audiostream
starting.

The old 'eq_active' and 'eq_gains' linphonerc parameters have been disabled
as they are no more sufficient to set different parameters on the speaker and
microphone equalizers. Those have been replaced by four parameters that
explicitly specify the equalizer to set up.

These new parameters are 'mic_eq_active' and 'mic_eq_gains' for the microphone equalizer and
'spk_eq_active' and 'spk_eq_gains' for the speaker equalizer.

An extra parameter has also been added in audio_stream_enable_equalizer() and
audio_stream_equalizer_set_gain() in order to indicate which equalizer is
concerned by the modification.

+fix compile bug when ZRTP is not available

It was not built anymore because of a cmake conditionnal options that
depended of a variable set further in CMakeLists.txt

Set DSCP value after the remote address has been set (this is needed for DSCP to be set correctly on Windows).

Fixing bug which was preventing libv4l2 from being really disabled when using the option --disable-libv4l2.

Make the VideoToolbox decoder sends the MS_DECODER_DECODING_ERRORS notification when it fails to decode a frame