The following algorithmic functionality was added:
-Add support for an exact regressor power to be computed
which avoids the issue with the updating of the filter
sometimes being unstable.
-Lowered the fixed step size of the adaptive filter to 0.05
which significantly reduces the sensitivity of the
adaptive filter to near-end noise, nonlinearities,
doubletalk and the unmodelled echo path tail. It also
reduces the tracking speed of the adaptive filter but the
chosen value proved to give a sufficient tradeoff for the
requirements on the adaptive filter.
To allow the new functionality to be selectively applied the following was done:
-A new Config was added for selectively activating the functionality.
-Functionality was added in the audioprocessing and echocancellationimpl classes
for passing the activation of the functionality down to the AEC algorithms.
To make the code for the introduction of the functionality clean,
the following refactoring was done:
-The selection of the step size was moved to a single place.
-The constant for the step size of the adaptive filter in extended filter mode was
made local.
-The state variable storing the step-size was renamed to a more describing name.
When the new functionality is not activated, the changes
have been tested for bitexactness on Linux.
TBR=minyue@webrtc.org
BUG=webrtc:5778, webrtc:5777
Review URL: https://codereview.webrtc.org/1887003002
Cr-Commit-Position: refs/heads/master@{#12384}
are active in the module and its submodules.
BUG=webrtc:5778, webrtc:5777
Review URL: https://codereview.webrtc.org/1886233003
Cr-Commit-Position: refs/heads/master@{#12371}
where the other audioprocessing unittests are located.
BUG=webrtc:5298
Review URL: https://codereview.webrtc.org/1846323002
Cr-Commit-Position: refs/heads/master@{#12343}
were added in https://codereview.webrtc.org/1773173002.
The reason for the revert is that for some scenarios
that CL causes problems in the coherence estimate used
in the AEC, which in turn causes echo leakage.
The reason for not reverting the actual CL is that
it would cause subsequent CLs to be reverted as well.
Therefore the choice was made to in this CK
instead revert the effects of that CL.
With the changes in this CL, the behavior is bitexact
to what it was before the CL mentioned above.
TBR=aluebs@webrtc.org
BUG=webrtc:5725
Review URL: https://codereview.webrtc.org/1867483003
Cr-Commit-Position: refs/heads/master@{#12259}
Revert reason: I unintentionally added a patch when rebasing that is breaking the bots.
This reverts commit 98c69a0ee785adeb9d95fffeb55cdb6cedbe82c6.
BUG=
Review URL: https://codereview.webrtc.org/1837313002 .
Cr-Commit-Position: refs/heads/master@{#12148}
This CL also extracts part of the functionality used
in the bitexactness test for the high-pass filter into
a separate file in order to be able to reuse that
functionality in bitexactness tests for the other
submodules in APM (including the bitexactness test for
the noise suppressor).
BUG=wertc:5336
Review URL: https://codereview.webrtc.org/1783203002
Cr-Commit-Position: refs/heads/master@{#12061}
It would be good to have a dedicated DebugDumpReplayer. There is one but it hides itself in DebugDumpTest.
This CL is to separate it out.
BUG=
Review URL: https://codereview.webrtc.org/1810463002
Cr-Commit-Position: refs/heads/master@{#12029}
Reason for revert:
Breaks Android it looks like.
See your own try jobs and
https://build.chromium.org/p/client.webrtc/builders/Android32%20Tests%20%28L%...
Original issue's description:
> Drop the 16kHz sample rate restriction on AECM and zero out higher bands
>
> The restriction has been removed completely and AECM now supports any
> number of higher bands. But this has been achieved by always zeroing out the
> higher bands, instead of applying a constant gain which is the average over half
> of the lower band (like it is done for the AEC), because that would be
> non-trivial to implement and we don't want to spend too much time on AECM, since
> we want to get rid of it in the long term anyway.
>
> R=peah@webrtc.org, solenberg@webrtc.org, tina.legrand@webrtc.org
>
> Committed: https://crrev.com/f687d53aabee0523ce6e9e0636163af8df120e41
> Cr-Commit-Position: refs/heads/master@{#11931}
TBR=peah@webrtc.org,turaj@webrtc.org,tina.legrand@webrtc.org,solenberg@webrtc.org,aluebs@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.webrtc.org/1781893002
Cr-Commit-Position: refs/heads/master@{#11932}
The restriction has been removed completely and AECM now supports any
number of higher bands. But this has been achieved by always zeroing out the
higher bands, instead of applying a constant gain which is the average over half
of the lower band (like it is done for the AEC), because that would be
non-trivial to implement and we don't want to spend too much time on AECM, since
we want to get rid of it in the long term anyway.
R=peah@webrtc.org, solenberg@webrtc.org, tina.legrand@webrtc.org
Review URL: https://codereview.webrtc.org/1774553002 .
Cr-Commit-Position: refs/heads/master@{#11931}
audio processing module.
The test also adds a new helper class called
VectorBasedAudioFrame that is intended to be
reused for the bitexactness tests for the other
submodules.
BUG=webrtc:1091
Review URL: https://codereview.webrtc.org/1510493004
Cr-Commit-Position: refs/heads/master@{#11864}
The audio level of the AEC's output level was calculated before overlapping add, and therefore, a compensation was needed. The compensation is multiplying the level by 2 since, before overlapping add, the level is roughly halved due to windowing.
This had to be that way because the level was calculated in frequency domain and the signal after overlapping add has only its time domain representation.
The level calculation has been updated to work on time domain signal and therefore the problem is not there any longer.
This CL is to put the calculation of the AEC output level after overlapping add and remove the compensation.
BUG=
R=peah@webrtc.org
Review URL: https://codereview.webrtc.org/1644133002 .
Cr-Commit-Position: refs/heads/master@{#11810}
This is needed when synthesizing a call based on
48 kHz audio files as otherwise an error is
generated about the wrong sample rate is generated.
That error is in turned caused by the sample rate
being changed from the default 16 kHz
at the first Capture API call event.
BUG=
Review URL: https://codereview.webrtc.org/1698243003
Cr-Commit-Position: refs/heads/master@{#11635}
This will make it align with protoc tools that use the relative
path from the project root to the files in the output path.
Having this, no hacks will need to be applied downstream.
TBR=henrik.lundin@webrtc.org
NOTRY=True
Review URL: https://codereview.webrtc.org/1673263002
Cr-Commit-Position: refs/heads/master@{#11540}
The level of the error signal after linear echo cancellation was based on non-buffered signal while that of the near-end and far-end signal based on buffered signal. This discrepancy made the comparison of them unfair.
This CL is to make calculating the error level rely on the same buffering.
BUG=
Review URL: https://codereview.webrtc.org/1510873004
Cr-Commit-Position: refs/heads/master@{#11408}
* Better param names
* Avoid using negative values for (bogus) placeholder channel counts (mostly in tests). Since channels will be changing to size_t, negative values will be illegal; it's sufficient to use 0 in these cases.
* Use arraysize()
* Use size_t for counting frames, samples, blocks, buffers, and bytes -- most of these are already size_t in most places, this just fixes some stragglers
* reinterpret_cast<int64_t>(void*) is not necessarily safe; use uintptr_t instead
* Remove unnecessary code, e.g. dead code, needlessly long/repetitive code, or function overrides that exactly match the base definition
* Fix indenting
* Use uint32_t for timestamps (matching how it's already a uint32_t in most places)
* Spelling
* RTC_CHECK_EQ(expected, actual)
* Rewrap
* Use .empty()
* Be more pedantic about matching int/int32_t/
* Remove pointless consts on input parameters to functions
* Add missing sanity checks
All this was found in the course of constructing https://codereview.webrtc.org/1316523002/ , and is being landed separately first.
BUG=none
TEST=none
Review URL: https://codereview.webrtc.org/1534193008
Cr-Commit-Position: refs/heads/master@{#11191}
Macro incorrectly displays DISABLED_ON_ANDROID in test names for
parameterized tests under --gtest_list_tests, causing tests to be
disabled on all platforms since they contain the DISABLED_ prefix rather
than their expanded variants.
This expands the macro variants to inline if they're disabled or not,
and removes building some tests under configurations where they should
fail, instead of building them but disabling them by default.
The change also removes gtest_disable.h as an unused include from many
other files.
BUG=webrtc:5387, webrtc:5400
R=kjellander@webrtc.org, phoglund@webrtc.orgTBR=henrik.lundin@webrtc.org
Review URL: https://codereview.webrtc.org/1547343002 .
Cr-Commit-Position: refs/heads/master@{#11150}
Reason for revert:
Compile error on Android needs to be fixed before relanding.
Original issue's description:
> Reland "Added option to specify a maximum file size when recording an AEC dump.", commit ae2c5ad12afc8cc29fe9c59dea432b697b871a87.
>
> The revert of the original CL was commit 36d4c545007129446e551c45c17b25377dce89a4.
> Original review: https://codereview.webrtc.org/1413483003/
>
> The original CL changes a function on audio_processing.h that is used by Chrome, this CL adds back the old function.
>
> NOTRY=true
> TBR=glaznev@webrtc.org, henrik.lundin@webrtc.org, solenberg@google.com, henrikg@webrtc.org, perkj@webrtc.org
> BUG=webrtc:4741
>
> Committed: https://crrev.com/f4f5cb09277d5ef6aeac8341e5f54a055867803a
> Cr-Commit-Position: refs/heads/master@{#11093}
TBR=glaznev@webrtc.org,henrik.lundin@webrtc.org,solenberg@google.com,henrikg@webrtc.org,perkj@webrtc.org,kwiberg@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:4741
Review URL: https://codereview.webrtc.org/1537213002
Cr-Commit-Position: refs/heads/master@{#11094}
Reason for revert:
Breaks Chrome-FYI bots because of a change in the StartDebugRecording function in audio_processing.h, that is called from Chrome.
Original issue's description:
> Added option to specify a maximum file size when recording an AEC dump.
>
> For applications with a strict filesize limit for debug files,
> I added an option to specify a maximum filesize for AEC dumps. An
> existing unit test is extended to check that the feature works as
> advertised.
>
> BUG=webrtc:4741
> TBR=glaznev@webrtc.org
>
> Committed: https://crrev.com/ae2c5ad12afc8cc29fe9c59dea432b697b871a87
> Cr-Commit-Position: refs/heads/master@{#11081}
TBR=pthatcher@webrtc.org,henrik.lundin@webrtc.org,henrikg@webrtc.org,solenberg@webrtc.org,andrew@webrtc.org,kwiberg@webrtc.org,perkj@webrtc.org,glaznev@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:4741
Review URL: https://codereview.webrtc.org/1533913004
Cr-Commit-Position: refs/heads/master@{#11087}
For applications with a strict filesize limit for debug files,
I added an option to specify a maximum filesize for AEC dumps. An
existing unit test is extended to check that the feature works as
advertised.
BUG=webrtc:4741
TBR=glaznev@webrtc.org
Review URL: https://codereview.webrtc.org/1413483003
Cr-Commit-Position: refs/heads/master@{#11081}
We can now use std::move instead!
This CL leaves the Pass methods in place; a follow-up CL will add deprecation annotations to them.
Review URL: https://codereview.webrtc.org/1460043002
Cr-Commit-Position: refs/heads/master@{#11064}
This is the second revert. The first attempt in https://codereview.webrtc.org/1423693008/
was missing a subtle curly brace caused by a merge conflict.
I'm going to let this one go through the CQ.
Reason for revert:
This breaks iOS GYP generation as described on http://www.webrtc.org/native-code/ios
I'm going to drive getting the build_with_libjingle=1 setting removed from the bots to match the official instructions.
See https://code.google.com/p/webrtc/issues/detail?id=4653 for more context, as this is exactly what that issue tries to solve.
Original issue's description:
> Add aecdump support to audioproc_f.
>
> Add a new interface to abstract away file operations. This CL temporarily
> removes support for dumping the output of reverse streams. It will be easy to
> restore in the new framework, although we may decide to only allow it with
> the aecdump format.
>
> We also now require the user to specify the output format, rather than
> defaulting to the input format.
>
> TEST=Bit-exact output to the previous audioproc_f version using an input wav
> file, and to the legacy audioproc using an aecdump file.
>
> Committed: https://crrev.com/bdafe31b86e9819b0adb9041f87e6194b7422b08
> Cr-Commit-Position: refs/heads/master@{#10460}
TBR=aluebs@webrtc.org,peah@webrtc.org,andrew@webrtc.org
BUG=
Review URL: https://codereview.webrtc.org/1412963007
Cr-Commit-Position: refs/heads/master@{#10532}
Reason for revert:
Oh dear, this broke compilation.
I guess more was built on top of this CL before I reverted it.
Reverting now for futher investigation (and re-land using CQ)
Original issue's description:
> Revert of Add aecdump support to audioproc_f. (patchset #8 id:200001 of https://codereview.webrtc.org/1409943002/ )
>
> Reason for revert:
> This breaks iOS GYP generation as described on http://www.webrtc.org/native-code/ios
> I'm going to drive getting the build_with_libjingle=1 setting removed from the bots to match the official instructions.
>
> See https://code.google.com/p/webrtc/issues/detail?id=4653 for more context, as this is exactly what that issue tries to solve.
>
> Original issue's description:
> > Add aecdump support to audioproc_f.
> >
> > Add a new interface to abstract away file operations. This CL temporarily
> > removes support for dumping the output of reverse streams. It will be easy to
> > restore in the new framework, although we may decide to only allow it with
> > the aecdump format.
> >
> > We also now require the user to specify the output format, rather than
> > defaulting to the input format.
> >
> > TEST=Bit-exact output to the previous audioproc_f version using an input wav
> > file, and to the legacy audioproc using an aecdump file.
> >
> > Committed: https://crrev.com/bdafe31b86e9819b0adb9041f87e6194b7422b08
> > Cr-Commit-Position: refs/heads/master@{#10460}
>
> TBR=aluebs@webrtc.org,peah@webrtc.org,andrew@webrtc.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Committed: https://crrev.com/d279941bb54bfdc6e7324bf36cac76581474b96d
> Cr-Commit-Position: refs/heads/master@{#10523}
TBR=aluebs@webrtc.org,peah@webrtc.org,andrew@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.webrtc.org/1419953010
Cr-Commit-Position: refs/heads/master@{#10524}
Reason for revert:
This breaks iOS GYP generation as described on http://www.webrtc.org/native-code/ios
I'm going to drive getting the build_with_libjingle=1 setting removed from the bots to match the official instructions.
See https://code.google.com/p/webrtc/issues/detail?id=4653 for more context, as this is exactly what that issue tries to solve.
Original issue's description:
> Add aecdump support to audioproc_f.
>
> Add a new interface to abstract away file operations. This CL temporarily
> removes support for dumping the output of reverse streams. It will be easy to
> restore in the new framework, although we may decide to only allow it with
> the aecdump format.
>
> We also now require the user to specify the output format, rather than
> defaulting to the input format.
>
> TEST=Bit-exact output to the previous audioproc_f version using an input wav
> file, and to the legacy audioproc using an aecdump file.
>
> Committed: https://crrev.com/bdafe31b86e9819b0adb9041f87e6194b7422b08
> Cr-Commit-Position: refs/heads/master@{#10460}
TBR=aluebs@webrtc.org,peah@webrtc.org,andrew@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.webrtc.org/1423693008
Cr-Commit-Position: refs/heads/master@{#10523}
This test is to verify that the debug dump can perfectly reproduce APM states if the recording is made from the first input sample.
BUG=
Review URL: https://codereview.webrtc.org/1393353003
Cr-Commit-Position: refs/heads/master@{#10506}
This changes the following module directories:
* webrtc/modules/audio_conference_mixer/interface
* webrtc/modules/interface
* webrtc/modules/media_file/interface
* webrtc/modules/rtp_rtcp/interface
* webrtc/modules/utility/interface
To avoid breaking downstream, I followed this recipe:
1. Copy the interface dir to a new sibling directory: include
2. Update the header guards in the include directory to match the style guide.
3. Update the header guards in the interface directory to match the ones in include. This is required to avoid getting redefinitions in the not-yet-updated downstream code.
4. Add a pragma warning in the header files in the interface dir. Example:
#pragma message("WARNING: webrtc/modules/interface is DEPRECATED; "
"use webrtc/modules/include")
5. Search for all source references to webrtc/modules/interface and update them to webrtc/modules/include (*.c*,*.h,*.mm,*.S)
6. Update all GYP+GN files. This required manual inspection since many subdirectories of webrtc/modules referenced the interface dir using ../interface etc(*.gyp*,*.gn*)
BUG=5095
TESTED=Passing compile-trybots with --clobber flag:
git cl try --clobber --bot=win_compile_rel --bot=linux_compile_rel --bot=android_compile_rel --bot=mac_compile_rel --bot=ios_rel -m tryserver.webrtc
R=stefan@webrtc.org, tommi@webrtc.org
Review URL: https://codereview.webrtc.org/1417683006 .
Cr-Commit-Position: refs/heads/master@{#10500}
