Puts the whole block in contiguous memory and reduce pointer look-up.
The change has been verified to be bit-exact.
Bug: webrtc:14089
Change-Id: I264aaf764bf53a29f23249105f704b2fdbd7e51c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/263203
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Gustaf Ullberg <gustaf@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36983}
The high-band gain is corrected by fixing the computation of the
low-band energy
Bug: webrtc:14108
Change-Id: I5033287de57aedcd91bb71623ca2862519ffb35b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/263201
Commit-Queue: Gustaf Ullberg <gustaf@webrtc.org>
Reviewed-by: Jesus de Vicente Pena <devicentepena@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36972}
This change adds a Block class to reduce the need for std::vector<std::vector<std::vector<float>>>. This make the code
easier to read and less error prone.
It also enables future changes to the underlying data structure of a
block. For instance, the data of all bands and channels could be stored
in a single vector.
The change has been verified to be bit-exact.
Bug: webrtc:14089
Change-Id: Ied9a78124c0bbafe0e912017aef91f7c311de2ae
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/262252
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Gustaf Ullberg <gustaf@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36968}
Adds two metrics for stereo detection:
- An enum indicating whether the last 10 seconds contained persistent stereo content or not, logged every 10 seconds.
- An enum indicating whether any persistent stereo content at all has been detected, logged at the end of the AEC lifetime.
These metrics allow us to assess:
- What proportion of all audio is treated as stereo.
- What proportion of sessions encounter any significant stereo content. If this is unexpectedly high, the stereo detection code may need fine tuning.
Metrics are only logged for component lifetimes exceeding 5 seconds: This is to avoid brief AEC lifetimes due to internal resets etc within APM.
Corresponding Chrome CL for XML histogram declarations:
https://crrev.com/c/3579317
Bug: chromium:1295710
Change-Id: I93e2bf74588cf4bb2a8922dbfad079bccab01456
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/258760
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36537}
During temporary stereo content when the AEC3 uses a mono reference signal, the signal is downmixed by averaging instead of using only the left channel.
Additionally, temporary stereo content is flagged as an echo path change.
Tested: Modified local build: Verified stereo mode entered / left in accordance with hysteresis and timeout thresholds. Verified temporary stereo detected during temporary stereo playout. Made an aecdump and inspected content.
Bug: chromium:1295710
Change-Id: I6bd53e615dfb3ec39bc1c73275b7d6d599ac7c57
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/258481
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36504}
Even if playout audio is only very briefly stereo, the AEC will enter stereo processing mode. To save CPU and improve AEC performance, this CL adds a hysteresis period before treating playout as stereo.
The feature is enabled by default in the AEC3 config.
Bug: chromium:1295710
Change-Id: I29116ab2e7823e25a02aa3b66a1c619f1d966d9e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/258479
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36503}
If playout audio is temporarily stereo, the AEC will currently enter stereo processing mode indefinitely. To save CPU and improve AEC performance, this CL adds support for falling back to mono after a period of no stereo.
The feature is enabled by default in the AEC3 config.
Bug: chromium:1295710
Change-Id: I690b5b22f8407f950bf41f3bcaa9ca0138452157
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/258421
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36502}
The features have two safety fallbacks:
- multichannel config has a killswitch WebRTC-Aec3SetupSpecificDefaultConfigDefaultsKillSwitch
- stereo detection has a killswitch WebRTC-Aec3StereoContentDetectionKillSwitch
Both features are enabled by default in the AEC3 config.
Tested: Bitexact on a large number of aecdumps.
Bug: chromium:1295710
Change-Id: I340cdc9140dacd4ca22d0911eb9f732b6cf8b226
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/258129
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36482}
Apart from making the construction more straightforward, this change allows recreating the BlockProcessor at runtime. This is used to change parameterization at runtime in an upcoming CL [1].
[1] https://webrtc-review.googlesource.com/c/src/+/258129
Tested: Bitexact on a large number of aecdumps.
Bug: chromium:1295710
Change-Id: I2e0275c5c97044cb4370042633266b193c06b960
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/258100
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36473}
This CL removes even more top-level const from parameters in function
declarations. This change is safe because top-level const in function
declarations (not function definitions) are ignored by the compiler
and so change is just a no-op cleanup.
Bug: webrtc:13610
Change-Id: Icf6868c27b1fdb9d9915b3a7020eb34bdcf07a09
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/249989
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Ali Tofigh <alito@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35866}
This is a safe cleanup change since top-level const applied to
parameters in function declarations (that are not also
definitions) are ignored by the compiler. Hence, such changes do
not change the type of the declared functions and are simply
no-ops.
Bug: webrtc:13610
Change-Id: Ibafb92c45119a6d8bdb6f9109aa8dad6385163a9
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/249086
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Ali Tofigh <alito@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35802}
Changing to an index for-loop (instead of using std::max_element & std::distance) tracking even & odd elements separately allows the compiler to produce code with less pipeline stall.
Bug: None
Change-Id: Iaa3e820a3a3b61e2eb276f0dac9106c848db1891
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/240061
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Christian Schuldt <cschuldt@google.com>
Cr-Commit-Position: refs/heads/main@{#35729}
Manually unrolling the multiply-and-accumulate loop of the matched filter allows interleaving of instruction, which gives a significant saving.
Bug: None
Change-Id: Ie7a7d92bd453d81e9dd61812781a7b6d62e1f1f4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/240321
Reviewed-by: Ivo Creusen <ivoc@webrtc.org>
Commit-Queue: Christian Schuldt <cschuldt@google.com>
Reviewed-by: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35566}
Changing to an index for-loop (instead of a range for-loop) allows the compiler (clang for x86 at least) to unroll it x2.
Bug: None
Change-Id: I9b9612a8513a06e8aa3b12ae39f6911217da55fa
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/239741
Reviewed-by: Ivo Creusen <ivoc@webrtc.org>
Commit-Queue: Christian Schuldt <cschuldt@google.com>
Cr-Commit-Position: refs/heads/main@{#35478}
Changing to an index for-loop (instead of using std::transform) allows the compiler (clang for x86 at least) to use 3 different registers in the loop rather than just 1, resulting in less pipeline stall (I'd assume). Interestingly, the compiler unrolls the loop(s) completely in both cases.
Bug: None
Change-Id: I586773bc525e91bb6eb6638d5399928482306b9f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/239364
Reviewed-by: Ivo Creusen <ivoc@webrtc.org>
Commit-Queue: Christian Schuldt <cschuldt@google.com>
Cr-Commit-Position: refs/heads/main@{#35453}
Reducing pointer following. This will allow the compiler to optimize more efficiently with the "-fno-strict-aliasing" flag.
Bug: None
Change-Id: Ic126bd2d53969a7e9d15e1c1081d5278e27a816c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/238664
Reviewed-by: Per Åhgren <peah@webrtc.org>
Reviewed-by: Ivo Creusen <ivoc@webrtc.org>
Commit-Queue: Christian Schuldt <cschuldt@google.com>
Cr-Commit-Position: refs/heads/main@{#35414}
Reducing pointer following. This will allow the compiler to optimize more efficiently with the "-fno-strict-aliasing" flag.
Bug: None
Change-Id: I7cde835161e2d3e85fc7c919556fa9a9e87ef6df
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/238169
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Christian Schuldt <cschuldt@google.com>
Cr-Commit-Position: refs/heads/main@{#35393}
Reducing pointer following. This will allow the compiler to optimize more efficiently with the "-fno-strict-aliasing" flag.
Bug: None
Change-Id: Ib1fd3a1cf3f89471b0ec87404650a6061eec5e2d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/237782
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Christian Schuldt <cschuldt@google.com>
Cr-Commit-Position: refs/heads/main@{#35374}
Add implementation of RTC_DCHECK_NOTREACHED equal to the RTC_NOTREACHED.
The new macros will replace the old one when old one's usage will be
removed. The idea of the renaming to provide a clear signal that this
is debug build only macros and will be stripped in the production build.
Bug: webrtc:9065
Change-Id: I4c35d8b03e74a4b3fd1ae75dba2f9c05643101db
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/237802
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35348}
This change improves echo canceller transparency by enabling the use
of a non-capped ERLE when computing the residual echo spectrum for
dominant nearend detection.
Experimentation has shown that the feature improves echo canceller
transparency and user ratings.
Implementation CL:
https://webrtc-review.googlesource.com/c/src/+/221920
Bug: webrtc:12870
Change-Id: I7dc66810e8300cd35321bcd5b9fae9bc3386836d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/234841
Reviewed-by: Jesus de Vicente Pena <devicentepena@webrtc.org>
Commit-Queue: Gustaf Ullberg <gustaf@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35186}
The dominant nearend detector uses the residual echo spectrum for
determining whether in nearend state. The residual echo spectrum in
computed using the ERLE. To reduce the risk of echo leaks in the
suppressor, the ERLE is capped. While minimizing echo leaks, the
capping of the ERLE can affect the dominant nearend classification
negatively as the residual echo spectrum is often over estimated.
This change enables the dominant nearend detector to use a residual
echo spectrum computed with a virtually non-capped ERLE. This ERLE
is only used for dominant nearend detection and leads to increased
transparency.
The feature is currently disabled by default and can be enabled
with the field trial "WebRTC-Aec3UseUnboundedEchoSpectrum".
Bug: webrtc:12870
Change-Id: Icb675c6f5d42ab9286e623b5fb38424d5c9cbee4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221920
Reviewed-by: Jesus de Vicente Pena <devicentepena@webrtc.org>
Commit-Queue: Gustaf Ullberg <gustaf@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34270}
This change enables the use of two different adaptation speeds of the
matched filter of the delay estimator of AEC3.
One speed is used when no delay has been found, and one is used after a
reliable delay has been found. The purpose is to use a slower adaptation
speed to reduce the risk of divergence during double-talk without
slowing down the search for the initial delay.
The CL prepares for experimentation by adding field trials for
controlling the two adaptation speeds.
Bug: webrtc:12775
Change-Id: I817a1ab5ded0f78d20de45edcf04c708290173fc
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/219083
Commit-Queue: Gustaf Ullberg <gustaf@webrtc.org>
Reviewed-by: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34055}
The ERLE is used to estimate residual echo for echo suppression. The
ERLE is reduced during far-end offset to avoid echo leakage. When there
is a strong near-end present this can cause unnecessary transparency loss.
This change adds an ERLE estimation that does not compensate for onsets and
uses it for residual echo estimation when the suppressor considers the near-end to be dominant.
Bug: webrtc:12686
Change-Id: Ida78eeacf1f95c6e62403f86ba3f2ff055898a84
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/215323
Commit-Queue: Gustaf Ullberg <gustaf@webrtc.org>
Reviewed-by: Jesus de Vicente Pena <devicentepena@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33786}
Erle Uncertainty changes the residual echo computation during saturated
echo. However, the case of saturated echo is already handled by the
residual echo estimator causing the ErleUncertainty to be a no-op.
The change has been tested for bit-exactness.
Bug: webrtc:8671
Change-Id: I779ba67f99f29d4475a0465d05da03d42d50e075
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/215072
Reviewed-by: Jesus de Vicente Pena <devicentepena@webrtc.org>
Commit-Queue: Gustaf Ullberg <gustaf@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33719}
This CL adds functionality in AEC3 that allows the computational
complexity to be reduced when the output of APM is not used.
Bug: b/177830919
Change-Id: I08121364bf966f34311f54ffa5affbfd8b4db1e2
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/211341
Reviewed-by: Gustaf Ullberg <gustaf@webrtc.org>
Commit-Queue: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33476}
Currently the echo canceller reference signal is high-pass filtered to
avoid the need of modeling the capture-side high-pass filter as part of
the echo path.
This can lead to the lowest frequency bins of the linear filter
diverging as there is little low-frequency content available for
training. Over time the filter can output an increasing amount of
low-frequency power, which in turn affects the filter's ability to
adapt properly.
Disabling the high-pass filtering of the echo canceller reference solves
this issue, resulting in improved filter convergence.
Bug: webrtc:12265
Change-Id: Ic526a4b1b73e1808cfcd96a8cdee801b96a27671
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/208288
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Gustaf Ullberg <gustaf@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33322}
We need to be able build chromium with rtc_include_tests = true. It
reveals a lot of targets that are not compatible with chromium but
aren't marked so.
`rtc_include_tests=true` has been considered a way to disable targets for the Chromium build, causing an overload on rtc_include_tests while the meaning of the two GN args (rtc_include_tests and build_with_chromium) should be kept separated.
Bug: webrtc:12404
Change-Id: I2f72825445916eae7c20ef9338672d6a07a9b9ff
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/203890
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Andrey Logvin <landrey@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33124}
