- Change how the transmission offset is calculated, to
incorporate the time since the frame was captured.
- Break out RtpRtcpClock and move it to system_wrappers.
- Use RtpRtcpClock to set the capture time in ms in the capture module.
We must use the same clock as in the RTP module to be able to measure
the time from capture until transmission.
- Enables the RTP header extension for packet transmission time offsets.
BUG=
TEST=trybots
Review URL: https://webrtc-codereview.appspot.com/666006
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2489 4adac7df-926f-26a2-2b94-8c16560cd09d
We previously had an error when a mono capture device was used with
a stereo codec. This is prevented by avoiding any remixing in
AudioProcessing. Instead, capture side downmixing is now done before
resampling. Upmixing can now be handled properly by AudioCoding,
since the AudioProcessing error condition has been removed.
On the render side, downmixing now occurs before resampling. Ideally
this would be handled still earlier in the chain. Similarly, downmixing
for the AudioProcessing reference data occurs before resampling. This
code has been refactored into RemixAndResample, with a comprehensive
unittest added in output_mixer_unittest.cc.
BUG=issue624
TEST=manually through voe_cmd_test, by using mono and stereo capture
and render devices with mono and stereo codecs. voice_engine_unittest,
voe_auto_test.
Review URL: https://webrtc-codereview.appspot.com/676004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2448 4adac7df-926f-26a2-2b94-8c16560cd09d
Previously I uploaded patch 640007 to address issue 471. Today, while discussing that patch with Andrew, we noticed this patch should do the job. Leo is not here to verify it, but Andrew did some test to verify it. I'll ask Leo to do some testing.
We don't want to abandon patch 640007 as it will save some complexity.
Review URL: https://webrtc-codereview.appspot.com/648004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2405 4adac7df-926f-26a2-2b94-8c16560cd09d
Now we would get a failure if we do the same thing and the file would not be played out. Then GTalk/Hangout also reported this failure to us.
This CL is to restore the original function.
BUG = Issue 490
TEST = Manual test and voe_auto_test->FileBeforeStreamingTest
Review URL: https://webrtc-codereview.appspot.com/569016
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2347 4adac7df-926f-26a2-2b94-8c16560cd09d
Description:
Ths cl makes voe_cmd_test work on android by excluding some code
that are availabel on android today, some highlights
1. change maxnumofchannles
2. disable audio device selection
3. disable set/get volume
BUG=
TEST=test on try bots
Review URL: https://webrtc-codereview.appspot.com/584009
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2300 4adac7df-926f-26a2-2b94-8c16560cd09d
We now only set the volume when starting playout if the user has called
SetSpeakerVolume while we weren't playing. This now also ensures it will
actually get set to what the user requested rather than being overridden
by a default value.
Add tests to voe_auto_test.
BUG=6140661
TEST=voe_auto_test
Review URL: https://webrtc-codereview.appspot.com/566006
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2188 4adac7df-926f-26a2-2b94-8c16560cd09d
not per interface. This simplifies things a bit, reduces code and makes it
possible to implement reference counting (if we ever want) without the
static Delete() method. (Reference counted objects are traditionally
implicitly deleted via the last Release())
* Since the reference counting code is now simpler, there's no need for the
RefCount class so I'm removing it.
* Also removing the optional |ignoreRefCounters| variable from the VoiceEngine::Delete
method. The justification is that it's no longer used and the reason it was there
in the first place was to avoid bugs in third party code, so it's an indication that
something is wrong elsewhere.
* Fix a crash in voe_network_impl and voe_extended_test that would happen on machines with IPv6 support.
* Fix an assert (handle the case) in the extended audio tests when SetCNAME is called with a NULL pointer.
* As a side effect of this change, hopefully the footprint of VoE will be slightly smaller :)
BUG=10445 (Coverity)
Review URL: https://webrtc-codereview.appspot.com/521003
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2127 4adac7df-926f-26a2-2b94-8c16560cd09d
In this CL I have rewritten the way we handle stereo packets in VoE.
Before this change we split the packets in the RTP module and added two packets to ACM, one for the left channel and one for the right. This lead to timing problems caused when a different thread called RecOut in between the two calls to add stereo packet to ACM. (RecOut is called to pull audio data, decode packets, on the receiving side).
While doing the change I also took the opportunity to changed some functions so that the data stream is uint8 everywhere.
The list of files in this CL is long, but should be fairly easy to review. It is difficult to see what has been changed in some of the tests, but I can explain offline.
Reviewers:
Björn - /src/modules/audio_coding
Patrik - /src/modules/rtp_rtcp
Patrik -/src/modules/utility
Henrik A - /src/voice_engine
BUG=410
TEST=voe_cmd_test
Review URL: https://webrtc-codereview.appspot.com/473003
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2012 4adac7df-926f-26a2-2b94-8c16560cd09d
