Mitigates an issue with sample rate conversion in Windows Audio Session API (WASAPI) which leads to audio glitches.

Mitigates an issue with sample rate conversion in Windows Audio Session API (WASAPI) which leads to audio glitches.

before being started.

Uses UTF-8 instead of ASCII in Core Audio on OS X in order to potentially resolve issues with non-ASCII characters. (The solution is unlikely to be complete though because of the mixes usage of original UTF-8, modified UTF-8 and strlen.)

Fixes an audio quality degradation observed with Core Audio on OS X when the Audio tab in Preferences is open during a call.

Fixes a case which may lead to ignoring non-distinguishing but important format parameters/SDP fmpt such as H.264 sprop-parameter-sets.

Updates AEC for CoreAudio using webrtc library, recompiled with latest webrtc, was crashing when running from source with unresolved reference "___sincos_stret". Adds some basic instructions for compiling.

Filters RTP packets in accord with the MediaStream direction. Fixes problems with new ReceiveStream-s being created while on hold.

Makes sure that the file is copied even if the jitsi repo has a version with a newer date (which could happen after a git pull).

from the MediaStream. Fixes a failure to re-start streams in certain situations
(sometimes after hold), resulting in no audio.

Updates fmj.jar to r40 (Fakes a TimeoutEvent when a ReceiveStream is replaced with a PassiveSSRCInfo (e.g. when an RTCP Receiver Report is received, instead of a Sender Report, for the ReceiveStream).).

AudioMixer. Improves handling of ReceiveStream timeout events in
MediaStreamImpl (handles TimeoutEvent-s with a null ReceiveStream). Calls
getDataSource().disconnect() on ReceiveStream-s when they are removed.

For RTCP FIR messages, use the SSRC from the packet FCI, instead of the 'SSRC of media source' field (see RFC5104 sections 4.3.1.1 and 4.3.1.2).