Break out the computation to a separate class, and call directly into
this from channel.cc rather than going through AudioProcessing. This
circumvents AudioProcessing's sample rate limitations.
We now compute the RMS over all samples rather than downmixing to a
single channel. This makes the call point in channel.cc easier, is
more "correct" and should have similar (negligible) complexity.
This caused slight changes in the RMS output, so the ApmTest.Process
reference has been updated. Snippet of the failing output:
[ RUN ] ApmTest.Process
Running test 4 of 12...
Value of: rms_level
Actual: 27
Expected: test->rms_level()
Which is: 28
Running test 5 of 12...
Value of: rms_level
Actual: 26
Expected: test->rms_level()
Which is: 27
Running test 6 of 12...
Value of: rms_level
Actual: 26
Expected: test->rms_level()
Which is: 27
Running test 10 of 12...
Value of: rms_level
Actual: 27
Expected: test->rms_level()
Which is: 28
Running test 11 of 12...
Value of: rms_level
Actual: 26
Expected: test->rms_level()
Which is: 27
Running test 12 of 12...
Value of: rms_level
Actual: 26
Expected: test->rms_level()
Which is: 27
BUG=3290
TESTED=Chrome assert is avoided and both voe_cmd_test and apprtc
produce reasonable printed out results from RMS().
R=bjornv@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/16459004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6056 4adac7df-926f-26a2-2b94-8c16560cd09d
Select "processing" rates based on the input and output sampling rates.
Resample the input streams to those rates, and if necessary to the
output rate.
- Remove deprecated stream format APIs.
- Remove deprecated device sample rate APIs.
- Add a ChannelBuffer class to help manage deinterleaved channels.
- Clean up the splitting filter state.
- Add a unit test which verifies the output against known-working
native format output.
BUG=2894
R=aluebs@webrtc.org, bjornv@webrtc.org, xians@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/9919004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5959 4adac7df-926f-26a2-2b94-8c16560cd09d
The design of VoeNetEqStatsTest in voice_engine_unittests depended on
being able to inject a factory for the audio coding module into
voice engine. This functionality is now likely going away, which would
make this test fail to compile. Further, the functionality under test
is mostly ACM functionality, wherefore it makes better sense to test it
at ACM level.
BUG=2996
R=henrika@webrtc.org, turaj@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/12029004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5912 4adac7df-926f-26a2-2b94-8c16560cd09d
This CL brought to you by:
$ for d in $(for f in $(git ls-files '*gyp' '*gypi'); do dirname $f; done|sort|uniq|grep -v '^\.$'); do echo -e "\n# These are for the common case of adding or renaming files. If you're doing\n# structural changes, please get a review from a reviewer in this file.\nper-file *.gyp=*\nper-file *.gypi=*" >> $d/OWNERS; done
$ for d in $(for f in $(git ls-files '*gyp' '*gypi'); do dirname $f; done|sort|uniq|grep -v '^\.$'); do git add $d/OWNERS; done
(and then removed the talk/ impact)
R=niklas.enbom@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/11969004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5903 4adac7df-926f-26a2-2b94-8c16560cd09d
Replace the two versions with a single DownConvertToCodecFormat. As
mentioned in comments, this could be further consolidated with
RemixAndResample but we should write a full audio converter class in
that case.
Along the way:
- Fix the bug present in Channel::Demultiplex with mono input and a
stereo codec.
- Remove the 32 kHz max from the OnDataAvailable path. This avoids a
48 -> 32 -> 48 conversion when VoE is passed 48 kHz audio; instead we
get a straight pass-through to ACM. The 32 kHz conversion is still
needed in the RecordedDataIsAvailable path until APM natively supports
48 kHz.
- Merge resampler improvements from ACM1 to ACM2. This allows ACM to
handle 44.1 kHz audio passed to VoE and was originally done here:
https://webrtc-codereview.appspot.com/1590004
- Reuse the RemixAndResample unit tests for DownConvertToCodecFormat.
- Remove unused functions from utility.cc.
BUG=3155,3000,b/12867572
TESTED=voe_cmd_test using both the OnDataAvailable and
RecordedDataIsAvailable paths, with a captured audio format of all
combinations of {44.1,48} kHz and {1,2} channels, running through all
codecs, and finally using both ACM1 and ACM2.
R=henrika@webrtc.org, turaj@webrtc.org, xians@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/11019005
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5843 4adac7df-926f-26a2-2b94-8c16560cd09d
In the case where a network glitch causes a packet to arrive so late
that the jitter buffer has gone into expand mode, the playout timestamp
could have been increased to a value that is larger than the RTP
timestamp of the late packet when it finally arrives. This causes
the difference to be negative, and would make the value wrap (unsigned).
With this fix, the difference is set to zero when the playout
timestamp is ahead of the incoming RTP timestamp. Further down in the
method, a zero-value will lead to the averaging filter not being updated.
BUG=3080
R=henrika@webrtc.org, tina.legrand@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/10199004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5735 4adac7df-926f-26a2-2b94-8c16560cd09d
- Add ability to VoE to send Absolute Sender Time header extension.
- Refactor handling of RTP header extensions in VoE to work the same as in ViE.
- Add API to enable receiving Absolute Sender Time in VoE.
This is part of the work to include audio packets in bandwidth estimation, for
better accuracy in estimates.
BUG=
TBR=solenberg@webrtc.org,henrikg@webrtc.org,stefan@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/9509004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5654 4adac7df-926f-26a2-2b94-8c16560cd09d
Reasons for removing:
- Feels like a complete hack IMHO.
- Not used by any client.
- Unclear functionality regarding time stamp, marker bit etc.
- Causes several issues in tests due to a bad design which mainly depends on the fact that this API "breaks" an ongoing data/packet flow and it complicates the threading model and creates risks for deadlock and memory corruption. Not worth trying to fix given the very unclear benefit of maintaining the API. Better to remove the API instead.
- We also see lots of TSan races and memcheck errors related to this API.
BUG=2296,2240
R=mflodman@webrtc.org, niklas.enbom@webrtc.org, xians@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/8819004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5574 4adac7df-926f-26a2-2b94-8c16560cd09d
Voice engine shouldn't really have to manage this. Instead, have AGC
keep track of the last input volume, so that it can avoid getting stuck
under coarsely quantized conditions.
Add a test to verify the behavior.
TESTED=unittests, and observed that AGC didn't get stuck on a MacBook
where this problem can actually occur.
R=bjornv@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/8729004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5571 4adac7df-926f-26a2-2b94-8c16560cd09d
