Olga said she enabled the out-of-process field trial yesterday; that could explain this failure.

You should try building Chromium on Windows and verify it still works for WebRTC calls. If it does, perhaps just --use-file-for-fake-audio-capture is broken with the new audio process?

Field trial config landed in this build:
https://uberchromegw.corp.google.com/i/chromium.webrtc/builders/Win7%20Tester/builds/43383

#2: so it had been working fine up to some point at least.

Hmm, kRequireAudioHardwareForTesting isn't propagated to the service. The whitelist is here: https://cs.chromium.org/chromium/src/content/browser/utility_process_host.cc?l=277.

That flag isn't used in the service, so ignore comment 5.

Looks like --use-file-for-fake-audio-capture is propagated, so I guess that works on Win.

If I were you I'd check manually that audio is actually working on win, and then check that --use-file-for-fake-audio-capture works on Win.

 Issue 850974  has been merged into this issue.

I checked the latest Dev on my home computer - both YouTube and apprtc work with out of process audio.
Wonder if it's an issue with the service teardown in tests.

Marina, could you take a look?

Patrik - do these tests run on Linux/Mac bots?

They run on Linux, Mac and Windows, there was an #if for each OS to launch the corresponding audio recorder.

I tried it manually on Windows 10  and it works. I used this command:

.\out\Default\chrome.exe --enable-features=AudioServiceAudioStreams,AudioServiceOutOfProcess --use-fake-device-for-media-stream --use-file-for-fake-audio-capture="C:\Windows\Media\Alarm01.wav"

It's verified that audio works on Win; here are no regressions on Linux or Mac. Also, on Win regression happens several builds after the field trial is enabled.

So we conclude we can proceed with rolling out the audio service experiment to Beta.

Passing the bug to Patrik to take a look at the bot; he is disabling the failing tests now.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/7f205964e5298ab99035015b06b8ad8904979f53

commit 7f205964e5298ab99035015b06b8ad8904979f53
Author: Patrik Höglund <phoglund@chromium.org>
Date: Mon Jun 11 12:29:57 2018

Disable audio quality tests on Windows.

We have no idea why it's failing, so let's disable for now while
we investigate.

Bug:  chromium:850936 
Change-Id: I94643c513d06d51940b93b06ba98504b38b8e117
Reviewed-on: https://chromium-review.googlesource.com/1095100
Reviewed-by: Olga Sharonova <olka@chromium.org>
Commit-Queue: Olga Sharonova <olka@chromium.org>
Cr-Commit-Position: refs/heads/master@{#565977}
[modify] https://crrev.com/7f205964e5298ab99035015b06b8ad8904979f53/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc

Olga: see https://chromeperf.appspot.com/report?sid=cb074ba8fe8cdc9347489fd857d782fc30a37faa96ff06b99b05a5c4df89a01b

Looks like we have the same problem on mac and linux now after you landed https://chromium-review.googlesource.com/c/chromium/src/+/1105957.

First failing build: https://ci.chromium.org/buildbot/chromium.webrtc/Linux%20Tester/39835

 Issue 854575  has been merged into this issue.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/decb47544b12663a0c494b8798c0f3e82ad9f543

commit decb47544b12663a0c494b8798c0f3e82ad9f543
Author: Olga Sharonova <olka@chromium.org>
Date: Wed Jun 20 13:47:47 2018

Revert "Changing AudioService experiment field trial to test service launch on startup"

This reverts commit ec3716fdd0b71dfbd6c5534aff75dd250ff6ad32.

Reason for revert: It broke WebRTC performance tests which capture system audio loopback. The reason is unclear, but looks like tests-specific: audio works with manual testing; and the test only changed event timing at the process startup.
Reverting for now to have WebRTC audio quality test coverage.

Bug:850936

Original change's description:
> Changing AudioService experiment field trial to test service launch on startup
> 
> [1] introduced AudioServiceLaunchOnStartup feature which we want to roll out
> with a new experiment group. Updating field trial to run the config on bots
> first.
> 
> [1] https://chromium-review.googlesource.com/c/chromium/src/+/1104681
> 
> Bug: 851611,853781
> Change-Id: I3796a8ac205a6d37706a86e8647c933ee510f41a
> Reviewed-on: https://chromium-review.googlesource.com/1105957
> Reviewed-by: Robert Kaplow <rkaplow@chromium.org>
> Commit-Queue: Olga Sharonova <olka@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#568485}

TBR=rkaplow@chromium.org,olka@chromium.org

Change-Id: Ic50b37d434bc75fca78585ead8c964815bcd2cf0
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 851611, 853781
Reviewed-on: https://chromium-review.googlesource.com/1107957
Reviewed-by: Olga Sharonova <olka@chromium.org>
Commit-Queue: Olga Sharonova <olka@chromium.org>
Cr-Commit-Position: refs/heads/master@{#568822}
[modify] https://crrev.com/decb47544b12663a0c494b8798c0f3e82ad9f543/testing/variations/fieldtrial_testing_config.json

Not sure if at all related to this clusterfuzz generated issue titled "CHECK failure: request->state(device.type) == MEDIA_REQUEST_STATE_OPENING in media_stream_manag", but since time window seems to somewhat overlap and I did not land any fixes, I'd like to link this for visibility: https://bugs.chromium.org/p/chromium/issues/detail?id=853515#c1

Thanks to Oleh, we managed to run the tests locally on Linux with minimal effort.

And I guess I figured what is going on:
Browser tests are modifying command line parameters dynamically.
But they are propagated to the audio process only once, on the process startup.

I see TestAudioGainControl* failing when running out of process - and they play the wrong test audio sound (the beeping one).

The proper fix would be for the audio service to expose the test API which allows to set up the execution mode, and for the browser tests to use it.

Actually, the only problem is that ConfigureFakeDeviceToPlayFile() [1] is called outside of SetUpCommandLine() method: SetUpCommandLine() method is called before the browser main loop is constructed, so all the relevant flags set there will be propogated to the audio process. But ConfigureFakeDeviceToPlayFile() is called when the test starts => the audio process has already started and the flag won't be propagated.

So we can just fix tests to set up switches::kUseFileForFakeAudioCapture in SetUpCommandLine() for each of them.

[1] https://cs.chromium.org/chromium/src/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc?type=cs&g=0&l=166

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/c0d567278fa84771b439b83b7fbd5421107b84c5

commit c0d567278fa84771b439b83b7fbd5421107b84c5
Author: Olga Sharonova <olka@chromium.org>
Date: Tue Jul 03 08:16:33 2018

Fix fake audio setup in WebRTC audio perf tests.

Change them to set up switches::kUseFileForFakeAudioCapture in
SetUpCommandLine(), so that it can be propagated to the audio process.

Bug:  850936 
Change-Id: Ia6cccc01d2b5fa7545abe8cb1554d72c33146566
Reviewed-on: https://chromium-review.googlesource.com/1122857
Reviewed-by: Henrik Boström <hbos@chromium.org>
Commit-Queue: Olga Sharonova <olka@chromium.org>
Cr-Commit-Position: refs/heads/master@{#572142}
[modify] https://crrev.com/c0d567278fa84771b439b83b7fbd5421107b84c5/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/5b68db8cc139421cce89fde0e588e90236b9c49d

commit 5b68db8cc139421cce89fde0e588e90236b9c49d
Author: Olga Sharonova <olka@chromium.org>
Date: Tue Jul 03 10:27:46 2018

Revert "Fix fake audio setup in WebRTC audio perf tests."

This reverts commit c0d567278fa84771b439b83b7fbd5421107b84c5.

Reason for revert: Win tests are still broken :(
https://uberchromegw.corp.google.com/i/chromium.webrtc/builders/Win7%20Tester/builds/43887/steps/browser_tests

Original change's description:
> Fix fake audio setup in WebRTC audio perf tests.
> 
> Change them to set up switches::kUseFileForFakeAudioCapture in
> SetUpCommandLine(), so that it can be propagated to the audio process.
> 
> Bug:  850936 
> Change-Id: Ia6cccc01d2b5fa7545abe8cb1554d72c33146566
> Reviewed-on: https://chromium-review.googlesource.com/1122857
> Reviewed-by: Henrik Boström <hbos@chromium.org>
> Commit-Queue: Olga Sharonova <olka@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#572142}

TBR=hbos@chromium.org,olka@chromium.org

Change-Id: I0a9dffe2824ac3fff218193fabe5caa39cb2a5cf
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug:  850936 
Reviewed-on: https://chromium-review.googlesource.com/1124359
Reviewed-by: Olga Sharonova <olka@chromium.org>
Commit-Queue: Olga Sharonova <olka@chromium.org>
Cr-Commit-Position: refs/heads/master@{#572163}
[modify] https://crrev.com/5b68db8cc139421cce89fde0e588e90236b9c49d/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/d5df68e2b692ec48660df5ab64f93236a9676799

commit d5df68e2b692ec48660df5ab64f93236a9676799
Author: Olga Sharonova <olka@chromium.org>
Date: Tue Jul 03 13:31:00 2018

Reland "Fix fake audio setup in WebRTC audio perf tests."

This is a reland of c0d567278fa84771b439b83b7fbd5421107b84c5

The fix helps only on Mac and Linux.

Original change's description:
> Fix fake audio setup in WebRTC audio perf tests.
>
> Change them to set up switches::kUseFileForFakeAudioCapture in
> SetUpCommandLine(), so that it can be propagated to the audio process.
>
> Bug:  850936 
> Change-Id: Ia6cccc01d2b5fa7545abe8cb1554d72c33146566
> Reviewed-on: https://chromium-review.googlesource.com/1122857
> Reviewed-by: Henrik Boström <hbos@chromium.org>
> Commit-Queue: Olga Sharonova <olka@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#572142}

Bug:  850936 
Change-Id: Ida4f3b888bc0332b68050d6b1cb8a0d5fef3cf32
Reviewed-on: https://chromium-review.googlesource.com/1124519
Reviewed-by: Henrik Boström <hbos@chromium.org>
Commit-Queue: Henrik Boström <hbos@chromium.org>
Cr-Commit-Position: refs/heads/master@{#572189}
[modify] https://crrev.com/d5df68e2b692ec48660df5ab64f93236a9676799/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc

#24 fixed failures on Linux and Mac from #16: the bots are green after updated field trial in #25 landed.

Windows problem seem to be different, and may still be related to how command line flags are propagated to to the audio process.

Marina, could you repro it locally and figure out what is going on?

As we discussed, since Stereo Mix is unavailable in newer versions of Windows, we may try to use some virtual audio driver to to do a system audio loopback to the default input.

+Oleh - could you help Marina with this if needed?

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/a3ec8289ffb15b064fb185f50e48746ce07d408a

commit a3ec8289ffb15b064fb185f50e48746ce07d408a
Author: Marina Ciocea <marinaciocea@chromium.org>
Date: Tue Jul 17 14:42:34 2018

Re-enable webrtc audio quality browsertests on windows.

Tests were previously disabled when they started failing (more context in linked bug). Re-enabling because tests pass locally on win7 machine.

Bug:  850936 
Change-Id: Ie1ab4e82daf2c0265e932237d3edf6b3399ee803
Reviewed-on: https://chromium-review.googlesource.com/1139066
Reviewed-by: Henrik Boström <hbos@chromium.org>
Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
Cr-Commit-Position: refs/heads/master@{#575638}
[modify] https://crrev.com/a3ec8289ffb15b064fb185f50e48746ce07d408a/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/21305961c9778e5425c6a515eddea53d7f38d5b3

commit 21305961c9778e5425c6a515eddea53d7f38d5b3
Author: Marina Ciocea <marinaciocea@chromium.org>
Date: Tue Jul 17 21:04:16 2018

Revert "Re-enable webrtc audio quality browsertests on windows."

This reverts commit a3ec8289ffb15b064fb185f50e48746ce07d408a.

Reason for revert: win tester bot failure

Original change's description:
> Re-enable webrtc audio quality browsertests on windows.
> 
> Tests were previously disabled when they started failing (more context in linked bug). Re-enabling because tests pass locally on win7 machine.
> 
> Bug:  850936 
> Change-Id: Ie1ab4e82daf2c0265e932237d3edf6b3399ee803
> Reviewed-on: https://chromium-review.googlesource.com/1139066
> Reviewed-by: Henrik Boström <hbos@chromium.org>
> Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#575638}

TBR=hbos@chromium.org,marinaciocea@chromium.org

Change-Id: I1cf598f783dd33be1f8834c5d7baf01a20479ffd
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug:  850936 
Reviewed-on: https://chromium-review.googlesource.com/1140854
Reviewed-by: Marina Ciocea <marinaciocea@chromium.org>
Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
Cr-Commit-Position: refs/heads/master@{#575769}
[modify] https://crrev.com/21305961c9778e5425c6a515eddea53d7f38d5b3/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc

I ran the tests locally on win7 laptop with virtual audio cable (workaround for devices that don't support stereo mix) and they pass. They still fail on the bot though (re-enabled them and reverted). 
There might be some difference between the stereo mix and VAC.
Also, the bot runs win server 2008.

I will try to get a win server 2008 and continue investigating.

After more discussions with @henrika we concluded that we should not be running these tests on win server 2008. @Oleh and I agreed that this bot should be reconfigured to run win 7 instead. No further investigations until the bot runs win 7. Reducing priority of the bug until then.

Oleh, could you suggest what is the status of work in #32?

Mirko, do you still consider this as blocking Stable sign-off for Issue 832666?

No progress there yet, sorry.

Re #33: No, it is not a blocker for Issue 832666.

The machine has been reinstalled to Windows 7 in bug 866565

Ok, re-enabling tests. Just double checking, the win 7 version has the needed driver to support stereo mix device needed by the tests?

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/c36d213a25008b6fd7edda75ccf181b5044594d5

commit c36d213a25008b6fd7edda75ccf181b5044594d5
Author: Marina Ciocea <marinaciocea@chromium.org>
Date: Tue Jul 31 08:16:23 2018

Reland "Re-enable webrtc audio quality browsertests on windows."

This is a reland of a3ec8289ffb15b064fb185f50e48746ce07d408a

Original change's description:
> Re-enable webrtc audio quality browsertests on windows.
> 
> Tests were previously disabled when they started failing (more context in linked bug). Re-enabling because tests pass locally on win7 machine.
> 
> Bug:  850936 
> Change-Id: Ie1ab4e82daf2c0265e932237d3edf6b3399ee803
> Reviewed-on: https://chromium-review.googlesource.com/1139066
> Reviewed-by: Henrik Boström <hbos@chromium.org>
> Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#575638}

Bug:  850936 
Change-Id: I44be2227ffed90d25549669de4dd3f354a98d92a
Reviewed-on: https://chromium-review.googlesource.com/1156224
Reviewed-by: Henrik Boström <hbos@chromium.org>
Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
Cr-Commit-Position: refs/heads/master@{#579344}
[modify] https://crrev.com/c36d213a25008b6fd7edda75ccf181b5044594d5/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc

Sorry, seems like it does not have any audio drivers :(
No stereo mix.

Can audio drivers be installed?

I'll see if I can get it to work

Ah ok, then I'm disabling the tests again until the bot driver is installed.

Once the bot is ready, please just reland this CL: https://crrev.com/c/1139066

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/50f29a60d4deb1d0924921010e824d17f8337ba0

commit 50f29a60d4deb1d0924921010e824d17f8337ba0
Author: Marina Ciocea <marinaciocea@chromium.org>
Date: Tue Jul 31 08:48:59 2018

Revert "Reland "Re-enable webrtc audio quality browsertests on windows.""

This reverts commit c36d213a25008b6fd7edda75ccf181b5044594d5.

Reason for revert: The bot does not have the needed audio driver to run this test.

Original change's description:
> Reland "Re-enable webrtc audio quality browsertests on windows."
> 
> This is a reland of a3ec8289ffb15b064fb185f50e48746ce07d408a
> 
> Original change's description:
> > Re-enable webrtc audio quality browsertests on windows.
> > 
> > Tests were previously disabled when they started failing (more context in linked bug). Re-enabling because tests pass locally on win7 machine.
> > 
> > Bug:  850936 
> > Change-Id: Ie1ab4e82daf2c0265e932237d3edf6b3399ee803
> > Reviewed-on: https://chromium-review.googlesource.com/1139066
> > Reviewed-by: Henrik Boström <hbos@chromium.org>
> > Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#575638}
> 
> Bug:  850936 
> Change-Id: I44be2227ffed90d25549669de4dd3f354a98d92a
> Reviewed-on: https://chromium-review.googlesource.com/1156224
> Reviewed-by: Henrik Boström <hbos@chromium.org>
> Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#579344}

TBR=hbos@chromium.org,marinaciocea@chromium.org

Change-Id: I30ce4fe93a7787cb5cca85fcf969c254e829dd74
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug:  850936 
Reviewed-on: https://chromium-review.googlesource.com/1156344
Reviewed-by: Marina Ciocea <marinaciocea@chromium.org>
Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
Cr-Commit-Position: refs/heads/master@{#579350}
[modify] https://crrev.com/50f29a60d4deb1d0924921010e824d17f8337ba0/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc

I have configured stereo mix now. I should've mentioned that there was no need to disable the test. Please enable it.

Great, thanks Oleh.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/0ef48c7e1c26e89d1f41d9a6e8a5243ec03746dd

commit 0ef48c7e1c26e89d1f41d9a6e8a5243ec03746dd
Author: Marina Ciocea <marinaciocea@chromium.org>
Date: Tue Jul 31 11:14:49 2018

Reland "Re-enable webrtc audio quality browsertests on windows."

This is a reland of a3ec8289ffb15b064fb185f50e48746ce07d408a

Original change's description:
> Re-enable webrtc audio quality browsertests on windows.
>
> Tests were previously disabled when they started failing (more context in linked bug). Re-enabling because tests pass locally on win7 machine.
>
> Bug:  850936 
> Change-Id: Ie1ab4e82daf2c0265e932237d3edf6b3399ee803
> Reviewed-on: https://chromium-review.googlesource.com/1139066
> Reviewed-by: Henrik Boström <hbos@chromium.org>
> Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#575638}

NOTRY=True

Bug:  850936 
Change-Id: I53579cd78304a845c6ec45500980c1a5cc6f96eb
Reviewed-on: https://chromium-review.googlesource.com/1156444
Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
Reviewed-by: Henrik Boström <hbos@chromium.org>
Cr-Commit-Position: refs/heads/master@{#579366}
[modify] https://crrev.com/0ef48c7e1c26e89d1f41d9a6e8a5243ec03746dd/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc

After re-enabling the tests for Windows, 2/4 pass:

Tests that pass:
- WebRtcAudioQualityBrowserTest.MANUAL_TestCallQualityWithAudioFromFakeDevice
- WebRtcAudioQualityBrowserTest.MANUAL_TestCallQualityWithAudioFromWebAudio

Tests that fail:
- WebRtcAudioQualityBrowserTest.MANUAL_TestAutoGainIsOffWithAudioProcessingOff (fails on all 3 tries)
- WebRtcAudioQualityBrowserTest.MANUAL_TestAutoGainControlOnLowAudio (passes on 3rd try)

Is this the same behavior as when the bug was reported?

Log: https://logs.chromium.org/v/?s=chromium%2Fbb%2Fchromium.webrtc%2FWin7_Tester%2F44475%2F%2B%2Frecipes%2Fsteps%2Fbrowser_tests%2F0%2Fstdout

Could these failures be caused by audio volume? Should it be set to 50% or 100% after all?

That could be the case, based on the description of the CL introducing AGC tests[1]. 

AFAIK the volume should be 50%.

Might help us to take a look at the files used in the test. Is it possible to grab the recording and actual files from the bot? Looks like they are not cleaned up in case of test failure [2].

[1] https://chromium.googlesource.com/chromium/src/+/fd442ed0421ff475ed3db26ec6d5ab61de7021f0
[2] https://cs.chromium.org/chromium/src/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc?rcl=b0c4e26bb5fe004283ad47673565f4f82b880fa9&l=822

But there is "gsutil upload" > "recordings from test"
https://ci.chromium.org/buildbot/chromium.webrtc/Win7%20Tester/44476Z

The audio volume was at 100%, I've set it to 50%.

Volume to 50% didn't change anything, the 2 tests still fail.
I'll disable the tests until I have more time to investigate.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/3e0fc5722f6a8510db3cc8b9f006300968c91f9b

commit 3e0fc5722f6a8510db3cc8b9f006300968c91f9b
Author: Marina Ciocea <marinaciocea@chromium.org>
Date: Wed Aug 01 15:31:42 2018

Disable failing webrtc audio quality tests on Windows.

This CL leaves enabled the tests that pass:
- WebRtcAudioQualityBrowserTest.MANUAL_TestCallQualityWithAudioFromFakeDevice
- WebRtcAudioQualityBrowserTest.MANUAL_TestCallQualityWithAudioFromWebAudio

and disables failing tests:
- WebRtcAudioQualityBrowserTest.MANUAL_TestAutoGainIsOffWithAudioProcessingOff
- WebRtcAudioQualityBrowserTest.MANUAL_TestAutoGainControlOnLowAudio

Bug:  850936 
Change-Id: Ib19073f01f0880cda81bd9900d2d07c354c71ae3


- NOTRY because these tests don't run by default on bots
- TBR because owners are OOO.

NOTRY=true
TBR=hbos@chromium.org

Change-Id: Ib19073f01f0880cda81bd9900d2d07c354c71ae3
Reviewed-on: https://chromium-review.googlesource.com/1158699
Reviewed-by: Marina Ciocea <marinaciocea@chromium.org>
Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
Cr-Commit-Position: refs/heads/master@{#579814}
[modify] https://crrev.com/3e0fc5722f6a8510db3cc8b9f006300968c91f9b/chrome/browser/media/webrtc/webrtc_audio_quality_browsertest.cc

Me and Oleh talked about these tests and it appears they have bad bot-workstation fidelity nowadays. We can talk about this on the next audio team meeting, but I think the tests need an overhaul so they're simpler to run, or maybe that we should delete them because they are difficult to maintain and don't catch many bugs anyway.

Is this finished? If not, please bump to M71.

Issue 876282 has been merged into this issue.

We'll need to follow up on this for M71.