The paths of the protobuf output files needs to match the actual tree path
in order for Bazel builds to work.
BUG=webrtc:6412
NOTRY=True
Review-Url: https://codereview.webrtc.org/2716413004
Cr-Commit-Position: refs/heads/master@{#16912}
and handling of echo path changes.
This CL add tuning for the AEC3 that
1) Improves the handling of the initial echo suppression
before the linear filter is reliable.
2) Improves the handling of echo path changes.
There are also minor bugfixes included.
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2717353002
Cr-Commit-Position: refs/heads/master@{#16873}
This CL adds correction of the echo suppressor gain
It contains 2 changes:
-Bounds the upper value for the echo suppression gain
for bin 1 to avoid that the high-pass filter causes the
gain to be high for bin 1, which in turn may impact the
realization of any lower gains for the neighboring bins.
-Bounds the upper values for the echo suppression gains
for the higher bins to avoid any impact of the
external anti-aliasing filters.
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2718993002
Cr-Commit-Position: refs/heads/master@{#16854}
Explicit initialization of const member of new EchoCanceller 3
submodule.
NOTRY=True
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2715573003
Cr-Commit-Position: refs/heads/master@{#16802}
The test program modules/audio_processing/test/audioproc_float.cc
defined the flag 'agc_compression_gain' and had checks if the
parameter was valid (audioproc_float). The flag was also copied to
webrtc::test::SimulationSettings of audio_processing_simulator.h. The
setting was however never applied to APM.
This change applies the setting on the GainControl submodule in the
same way as the agc_target_level is applied.
This is needed for e.g. testing the AGC fixed digital limiter with the
same configuration as it is (currently) used with in AudioMixerImpl.
Also added new flag '-experimental_agc'. This flag allows disabling the
experimental AGC, which is how the AGC is used in AudioMixerImpl.
ExperimentalAgc is enabled by default, exactly as it was prior to this change.
The change has been tested locally by listening tests and diff comparisons.
BUG=None
NOTRY=True # win_dbg bot not cooperating
Review-Url: https://codereview.webrtc.org/2684983004
Cr-Commit-Position: refs/heads/master@{#16603}
canceller 3
This CL adds code to the all the delay estimation functionality that is
available for the first version of echo canceller 3. The code completes
the class EchoPathDelayEstimator.
Note that this code does not yet include any handling of clock-drift so
there will be upcoming versions of this code.
Also note that the CL includes some minor changes in other files for
echo canceller 3.
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2644123002
Cr-Commit-Position: refs/heads/master@{#16489}
AEC3 and AEC2 are separate submodules in APM. This CL ensures that AEC3
deactivates AEC2 if both are active at the same time.
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2675863004
Cr-Commit-Position: refs/heads/master@{#16443}
This CL adds code to the BlockProcessor, which basically constitutes
the second layer in echo canceller 3. The CL includes two incomplete
classes (EchoRemover and EchoPathDelayEstimator) which will be completed
in upcoming CLs. Because of this, some of the unittests are disabled
until those are added.
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2611223003
Cr-Commit-Position: refs/heads/master@{#16319}
This is a first step towards it. I plan to refactor modules_unittests before continuing with this.
BUG=webrtc:6626
NOTRY=True
Review-Url: https://codereview.webrtc.org/2626163004
Cr-Commit-Position: refs/heads/master@{#16109}
Bulk of changes done using
git grep -l 'RTC_DCHECK(false)' | \
xargs sed -i 's/RTC_DCHECK(false)/RTC_NOTREACHED()/'
peerconnection.cc also used RTC_DCHECK(false && "msg") in two places,
which were updated manually.
BUG=webrtc:6424
Review-Url: https://codereview.webrtc.org/2623313004
Cr-Commit-Position: refs/heads/master@{#16026}
Reason for reland:
Added disabling of the BlockProcessor DEATH tests due to false alarms on the trybots.
These errors cannot be reproduced locally and they occur only in the DEATH test code,
which means that the memory leaks do not occur under normal conditions.
Original issue's description:
> Reason for revert:
> Memcheck buildbot detected memory leaks in the death tests. The code looks fine
> to me, but please investigate what causes the error and reland.
> > Original issue's description:
> > Added first layer of the echo canceller 3 functionality.
> >
> > This CL adds the first layer of the echo canceller 3.
> > All of the code is as it should, apart from
> > block_processor.* which only contains placeholder
> > functionality. (Upcoming CLs will add proper
> > functionality into those files.)
> >
> >
> >
> > BUG=webrtc:6018
> >
> > Review-Url: https://codereview.webrtc.org/2584493002
> > Cr-Commit-Position: refs/heads/master@{#15861}
> > Committed:
> 38fd1758e9
>
> TBR=ivoc@webrtc.org,aleloi@webrtc.org,henrik.lundin@webrtc.org,peah@webrtc.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=webrtc:6018
BUG=webrtc:6018
TBR=ivoc@webrtc.org,aleloi@webrtc.org,henrik.lundin@webrtc.org
Review-Url: https://codereview.webrtc.org/2608233002
Cr-Commit-Position: refs/heads/master@{#15884}
Reason for revert:
Memcheck buildbot detected memory leaks in the death tests. The code looks fine to me, but please investigate what causes the error and reland.
Original issue's description:
> Added first layer of the echo canceller 3 functionality.
>
> This CL adds the first layer of the echo canceller 3.
> All of the code is as it should, apart from
> block_processor.* which only contains placeholder
> functionality. (Upcoming CLs will add proper
> functionality into those files.)
>
>
>
> BUG=webrtc:6018
>
> Review-Url: https://codereview.webrtc.org/2584493002
> Cr-Commit-Position: refs/heads/master@{#15861}
> Committed: 38fd1758e9TBR=ivoc@webrtc.org,aleloi@webrtc.org,henrik.lundin@webrtc.org,peah@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2603293002
Cr-Commit-Position: refs/heads/master@{#15877}
This CL adds the first layer of the echo canceller 3.
All of the code is as it should, apart from
block_processor.* which only contains placeholder
functionality. (Upcoming CLs will add proper
functionality into those files.)
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2584493002
Cr-Commit-Position: refs/heads/master@{#15861}
that the level of the output in the audio processing
module is monitored. This CL adds that.
BUG=webrtc:6181, webrtc:6183, webrtc:6220
Review-Url: https://codereview.webrtc.org/2549143004
Cr-Commit-Position: refs/heads/master@{#15718}
The previous CL that added the ability to add
and artificial nearend signal had an issue with
null pointer access.
This is addressed in this CL.
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2573033003
Cr-Commit-Position: refs/heads/master@{#15600}
It will be followed by a number of other CLs that extends this framework.
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2567513003
Cr-Commit-Position: refs/heads/master@{#15593}
orders of the ProcessStream and ProcessReverseStream API calls.
This CL adds the ability to specify that call order in a file.
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2561843003
Cr-Commit-Position: refs/heads/master@{#15503}
is the one stored in the aecdump.
During AEC development, it is handy to be able to simulate different
doubletalk scenarios. This CL adds the ability to add an
artificial nearend on top of that present in the aecdump, which
allows for the developer to artificially customize the scenario
being tested
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2562593003
Cr-Commit-Position: refs/heads/master@{#15502}
This moves some GN check configurations out of .gn to individual targets.
The now checked target is:
"//webrtc/modules/audio_processing/*",
BUG=webrtc:6828
NOTRY=True
R=kjellander@webrtc.org
Review-Url: https://codereview.webrtc.org/2558813003
Cr-Commit-Position: refs/heads/master@{#15475}
Reason for revert:
The downstream problem is now fixed, and this should be good to land again.
Original issue's description:
> Revert of APM: Change 3 UMA metrics to fewer but linearly distributed buckets (patchset #2 id:20001 of https://codereview.webrtc.org/2547593002/ )
>
> Reason for revert:
> Breaks down-stream dependencies.
>
> Original issue's description:
> > APM: Change 3 UMA metrics to fewer but linearly distributed buckets
> >
> > In this change WebRTC.Audio.ApmCaptureInputLevel{Average,Peak} are
> > changed to 64 buckets, while WebRTC.Audio.AgcLevel is changed to 50
> > buckets. All three are changed to have linear spacing between buckets.
> >
> > Also, the metrics are renamed to avoid stats conflicts because of different bucket schemes:
> > - WebRTC.Audio.AgcLevel -> WebRTC.Audio.AgcSetLevel
> > - WebRTC.Audio.ApmCaptureInputLevelAverage -> WebRTC.Audio.ApmCaptureInputLevelAverageRms
> > - WebRTC.Audio.ApmCaptureInputLevelPeakRms -> WebRTC.Audio.ApmCaptureInputLevelPeakRms
> >
> > BUG=webrtc:6622
> > CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_chromium_rel_ng
> >
> > Committed: https://crrev.com/49715fe3be17d8579586d5bc954d626126d53415
> > Cr-Commit-Position: refs/heads/master@{#15418}
>
> TBR=peah@webrtc.org,rkaplow@chromium.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=webrtc:6622
>
> Committed: https://crrev.com/63407a9b6ae6f3fc096e01d64e46c6d21d86b517
> Cr-Commit-Position: refs/heads/master@{#15420}
TBR=peah@webrtc.org,rkaplow@chromium.org
BUG=webrtc:6622
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_chromium_rel_ng
Review-Url: https://codereview.webrtc.org/2551863003
Cr-Commit-Position: refs/heads/master@{#15442}
flaky, with many false results and with a huge
variance.
This CL addresses that by changing the way the
API call durations are measured, using a warmup
period and a longer interval for computing the
timing estimates.
Furthermore, this CL reduces the number of tests
to compensate for the fact that the tests now are
more expensive, as well as to reduce the number
of regressions further.
BUG=webrtc:6614,webrtc:6685,666725
Review-Url: https://codereview.webrtc.org/2549403002
Cr-Commit-Position: refs/heads/master@{#15440}
The Bazel build format doesn't support having separate
lists of compilation flags for C and C++; it just has a single
copts list for cc_library:
https://bazel.build/versions/master/docs/be/c-cpp.html#cc_binary.copts
This makes it hard to convert our GN targets to Bazel when there are
compiler warnings that aren't supported for C (like -Woverloaded-virtual
being added in bugs.webrtc.org/6653).
The solution for this is to move all .c files to their own targets
and remove C++-only compiler flags during conversion.
New targets:
//webrtc/common_audio:common_audio_c
//webrtc/common_audio:common_audio_neon_c
//webrtc/modules/audio_coding:g711_c
//webrtc/modules/audio_coding:g722_c
//webrtc/modules/audio_coding:ilbc_c
//webrtc/modules/audio_coding:isac_c
//webrtc/modules/audio_coding:isac_fix_c
//webrtc/modules/audio_coding:isac_test_util
//webrtc/modules/audio_coding:pcm16b_c
//webrtc/modules/audio_coding:webrtc_opusj_c
//webrtc/modules/audio_device:mac_portaudio
//webrtc/modules/audio_procssing:audio_processing_c
//webrtc/modules/audio_procssing:audio_processing_neon_c
This CL also adds a PRESUBMIT.py check that will throw an error
if targets are mixing .c and .cc files, to preven this from regressing.
BUG=webrtc:6653
NOTRY=True
Review-Url: https://codereview.webrtc.org/2550563003
Cr-Commit-Position: refs/heads/master@{#15433}
This change enables experimentation with the clipping minimum level
parameter in the gain control.
BUG=webrtc:6622
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_chromium_rel_ng
Review-Url: https://codereview.webrtc.org/2543753006
Cr-Commit-Position: refs/heads/master@{#15426}
Reason for revert:
Breaks down-stream dependencies.
Original issue's description:
> APM: Change 3 UMA metrics to fewer but linearly distributed buckets
>
> In this change WebRTC.Audio.ApmCaptureInputLevel{Average,Peak} are
> changed to 64 buckets, while WebRTC.Audio.AgcLevel is changed to 50
> buckets. All three are changed to have linear spacing between buckets.
>
> Also, the metrics are renamed to avoid stats conflicts because of different bucket schemes:
> - WebRTC.Audio.AgcLevel -> WebRTC.Audio.AgcSetLevel
> - WebRTC.Audio.ApmCaptureInputLevelAverage -> WebRTC.Audio.ApmCaptureInputLevelAverageRms
> - WebRTC.Audio.ApmCaptureInputLevelPeakRms -> WebRTC.Audio.ApmCaptureInputLevelPeakRms
>
> BUG=webrtc:6622
> CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_chromium_rel_ng
>
> Committed: https://crrev.com/49715fe3be17d8579586d5bc954d626126d53415
> Cr-Commit-Position: refs/heads/master@{#15418}
TBR=peah@webrtc.org,rkaplow@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:6622
Review-Url: https://codereview.webrtc.org/2548333002
Cr-Commit-Position: refs/heads/master@{#15420}
In this change WebRTC.Audio.ApmCaptureInputLevel{Average,Peak} are
changed to 64 buckets, while WebRTC.Audio.AgcLevel is changed to 50
buckets. All three are changed to have linear spacing between buckets.
Also, the metrics are renamed to avoid stats conflicts because of different bucket schemes:
- WebRTC.Audio.AgcLevel -> WebRTC.Audio.AgcSetLevel
- WebRTC.Audio.ApmCaptureInputLevelAverage -> WebRTC.Audio.ApmCaptureInputLevelAverageRms
- WebRTC.Audio.ApmCaptureInputLevelPeakRms -> WebRTC.Audio.ApmCaptureInputLevelPeakRms
BUG=webrtc:6622
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_chromium_rel_ng
Review-Url: https://codereview.webrtc.org/2547593002
Cr-Commit-Position: refs/heads/master@{#15418}
This CL is in preparation to move the AudioFrame into webrtc/api. The
AudioFrame is a POD type used for representing 10ms of audio. It
appears as a parameter and return value of interfaces being migrated
to webrtc/api, in particular AudioMixer.
Here, methods operator+=, operator>>=, Mute are
moved into a new target webrtc/audio/utility/audio_frame_operations,
and dependencies are changed to use
the new versions. The old AudioFrame methods are marked deprecated.
The audio frame utilities in webrtc/modules/utility:audio_frame_operations
are also moved to the new location.
TBR=kjellander@webrtc.org
BUG=webrtc:6548
NOPRESUBMIT=True
Review-Url: https://codereview.webrtc.org/2424173003
Cr-Commit-Position: refs/heads/master@{#15413}
These tests have been causing a large number of false alerts, which is
tons of work for the perf sheriff. It's infeasible to test every single
permutation, so we must choose carefully. This CL reduces the number
of RESULT lines from the test from ~450 to ~50. I attempted to choose
interesting permutations, but you probably know better what's
interesting...
https://chromeperf.appspot.com/report?
sid=a7193c96f708018848ca07ad9c78ac657cadab3c70b3939c42bd7d70a092d61a
also suggests most of the metrics have enormous standard deviations,
so maybe you could look into how stable the metrics really are
and remove/stabilize the ones that aren't?
BUG=chromium:666725
Review-Url: https://codereview.webrtc.org/2529393006
Cr-Commit-Position: refs/heads/master@{#15329}
This CL adds WebRTC.Audio.ApmCaptureInputLevelAverage and
WebRTC.Audio.ApmCaptureInputLevelPeak. The metrics are updated once
every 10 seconds.
BUG=webrtc:6622
Review-Url: https://codereview.webrtc.org/2534473004
Cr-Commit-Position: refs/heads/master@{#15300}
This change rewrites RMSLevel, making it accept an ArrayView as input,
and modify the implementation somewhat. It also makes the class keep
track of the peak RMS in addition to the average RMS over the
measurement period.
New tests are added to cover the new functionality.
BUG=webrtc:6622
Review-Url: https://codereview.webrtc.org/2535523002
Cr-Commit-Position: refs/heads/master@{#15294}
There's no longer any need to make the two arguments have the same
signedness, so we can remove a bunch of superfluous (and sometimes
dangerous) casts.
It turned out I also had to fix the safe_cmp functions to properly handle
enums that are implicitly convertible to integers.
NOPRESUBMIT=true
BUG=webrtc:6645
Review-Url: https://codereview.webrtc.org/2534683002
Cr-Commit-Position: refs/heads/master@{#15281}
