This CL is a followup to https://webrtc-review.googlesource.com/c/src/+/38481.
With the new approach we can just use the generic RTP packetizer to pass frames
over the wire as the specific info is contained within the bitstream. This makes
the new codec more modular and reduces its footprint.
Bug: webrtc:7671
Change-Id: Ib07f72a9d338e3cbfdbf39cba9891a959b5f7552
Reviewed-on: https://webrtc-review.googlesource.com/43220
Reviewed-by: Niklas Enbom <niklas.enbom@webrtc.org>
Commit-Queue: Emircan Uysaler <emircan@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21753}
This provides an intermediate class for defining default, null
implementations of all the PeerConnectionInterface/
PeerConnectionInternal methods. Specific fake PeerConnections then can
inherit from this and only override the methods pertaining to the
scenarios it will be used in.
Bug: webrtc:8764
Change-Id: I7614303b5673747244053b54b839e58aada43d10
Reviewed-on: https://webrtc-review.googlesource.com/43245
Commit-Queue: Steve Anton <steveanton@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21752}
This reverts commit d756fd06fed1b6c65dcb263cbd8f00ca23d72f3b.
Original change's description:
> Revert "Wrap Alpha and YUV frame into one EncodedImage for transmission"
>
> This reverts commit 5670c86aeccc9bc1191725431de7998d21b73c07.
>
> Reason for revert: Breaks downstream build. Need to add "#include <cstring>" to stereo_encoder_adapter.cc to use std::memcpy.
>
> Original change's description:
> > Wrap Alpha and YUV frame into one EncodedImage for transmission
> >
> > With alpha channel, we observe the artifacts on the receiver side, and
> > the reason is that when YUV channel has a key frame, it gives frame_buffer2
> > a chance to drop some previous frames. Then it is possible that some alpha
> > frames got dropped, which break the alpha frame dependence chain.
> >
> > In this CL, we pack the YUV frame and alpha encoded frame together as one
> > entity to solve the issue.
> >
> > Bug: webrtc:8773
> > Change-Id: Ibe746a46cb41fd92b399a7069e1d89f02f292af7
> > Reviewed-on: https://webrtc-review.googlesource.com/38481
> > Commit-Queue: Qiang Chen <qiangchen@chromium.org>
> > Reviewed-by: Emircan Uysaler <emircan@webrtc.org>
> > Cr-Commit-Position: refs/heads/master@{#21737}
>
> TBR=qiangchen@chromium.org,emircan@webrtc.org
>
> Change-Id: I11eff814ce093bf6db327ebcd21b1b71a1929849
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Bug: webrtc:8773
> Reviewed-on: https://webrtc-review.googlesource.com/43260
> Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
> Commit-Queue: Taylor Brandstetter <deadbeef@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#21739}
TBR=deadbeef@webrtc.org,qiangchen@chromium.org,emircan@webrtc.org
Change-Id: I0d64b7e7a62e4f35aa012270d3826a23b3fb2337
Bug: webrtc:8773
Reviewed-on: https://webrtc-review.googlesource.com/43440
Commit-Queue: Qiang Chen <qiangchen@chromium.org>
Reviewed-by: Qiang Chen <qiangchen@chromium.org>
Cr-Commit-Position: refs/heads/master@{#21749}
PeerConnectionInternal is being introduced so that it can be mocked in
tests and so that a fake can be written for it to be used by stats
tests.
Bug: webrtc:8764
Change-Id: I375d12ce352523e8ac584402685a7870bc399fac
Reviewed-on: https://webrtc-review.googlesource.com/43202
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Commit-Queue: Steve Anton <steveanton@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21747}
This is required to figure out when we can deprecate and remove
SDES.
Bug: chromium:804275
Change-Id: Ie234e6b3c8f5de8e78dda8d755d955caa61b7aa7
Reviewed-on: https://webrtc-review.googlesource.com/43340
Reviewed-by: Fredrik Solenberg <solenberg@webrtc.org>
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21746}
This change handles a special case in NoiseSuppression. The special
case was found by the AudioProcessing fuzzer.
A const copy of the capture audio stream is sent to
NoiseSuppression::AnalyzeCaptureAudio. Then audio undergoes processing
by e.g. the echo canceller. Then it's processed by
NoiseSuppression::ProcessCaptureAudio.
The special case is when the following conditions are all satisfied:
* All stream samples are constantly zero in the call to
AnalyzeCaptureAudio
* a processing component modifies it to be nonzero before the call to
ProcessCaptureAudio
* The array NoiseSuppressionC::magnPrevAnalyze is filled with
zeros. This holds after initialization.
In this case, there is a division by zero in WebRtcNs_ProcessCore. The
resulting NaN values pollute the output signal. They are only detected
several submodules later in the process chain. The NaN values cause
the EchoDetector to crash in debug mode.
There is special handling of the case when the signal is constant zero
in ProcessCore. This change avoids zero division by handling this
issue the same way.
Bug: chromium:803810 chromium:804634
Change-Id: I6d698dd0cd27e6d550b42085124300ce58533125
Reviewed-on: https://webrtc-review.googlesource.com/41282
Commit-Queue: Alex Loiko <aleloi@webrtc.org>
Reviewed-by: Henrik Lundin <henrik.lundin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21745}
Each simulcast stream requires dedicated decoder for decoding. SVC
can be decoded by single decoder. But in prod each receiver has its
decoder. We want to replicate this and also use one decoder per
spatial layer.
Also we create one frame writer per simulcast/spatial layer to dump
encoded/decoded frames of different layers to separate files.
Note that videoprocessor is still initialized with single
decoder/writer. It will be updated in next CL and start using
separate decoder/writer per layer.
Bug: webrtc:8524
Change-Id: I3bb3de77f97d51138b8b7675dd01bc281a078b2f
Reviewed-on: https://webrtc-review.googlesource.com/43280
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21744}
These parameters allow to configure number of simulcast/spatial layers
in video codec tests.
Bug: webrtc:8524
Change-Id: Iad1332732758a8297abcf740c24c483e5fccec9a
Reviewed-on: https://webrtc-review.googlesource.com/43020
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21741}
This reverts commit 5670c86aeccc9bc1191725431de7998d21b73c07.
Reason for revert: Breaks downstream build. Need to add "#include <cstring>" to stereo_encoder_adapter.cc to use std::memcpy.
Original change's description:
> Wrap Alpha and YUV frame into one EncodedImage for transmission
>
> With alpha channel, we observe the artifacts on the receiver side, and
> the reason is that when YUV channel has a key frame, it gives frame_buffer2
> a chance to drop some previous frames. Then it is possible that some alpha
> frames got dropped, which break the alpha frame dependence chain.
>
> In this CL, we pack the YUV frame and alpha encoded frame together as one
> entity to solve the issue.
>
> Bug: webrtc:8773
> Change-Id: Ibe746a46cb41fd92b399a7069e1d89f02f292af7
> Reviewed-on: https://webrtc-review.googlesource.com/38481
> Commit-Queue: Qiang Chen <qiangchen@chromium.org>
> Reviewed-by: Emircan Uysaler <emircan@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#21737}
TBR=qiangchen@chromium.org,emircan@webrtc.org
Change-Id: I11eff814ce093bf6db327ebcd21b1b71a1929849
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: webrtc:8773
Reviewed-on: https://webrtc-review.googlesource.com/43260
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Commit-Queue: Taylor Brandstetter <deadbeef@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21739}
On Chrome Remote Desktop for iOS, if all of these are true:
* Enable the PORTALLOCATOR_DISABLE_COSTLY_NETWORKS flag.
* Connect an iPhone to a Mac.
* Turn off WiFi on the iPhone and keep mobile data on.
* Connect to a host on a different network from iPhone.
Then the connection can never succeed. The reason is that iOS uses a
special network interface with link-local IP to communicate with the Mac.
BasicPortAllocator sees that interface and thinks it costs less than the
cellular networks, then it removes all cellular networks. However, that
link-local network cannot connect to a peer on external network.
This CL changes the behavior of the PORTALLOCATOR_DISABLE_COSTLY_NETWORKS
so that it ignores the cost of link-local networks when determining the
lowest network cost.
Bug: webrtc:8780
Change-Id: I9bde50426b356fdbf89d4b826fc7b28c8c523c10
Reviewed-on: https://webrtc-review.googlesource.com/41460
Commit-Queue: Yuwei Huang <yuweih@google.com>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21738}
With alpha channel, we observe the artifacts on the receiver side, and
the reason is that when YUV channel has a key frame, it gives frame_buffer2
a chance to drop some previous frames. Then it is possible that some alpha
frames got dropped, which break the alpha frame dependence chain.
In this CL, we pack the YUV frame and alpha encoded frame together as one
entity to solve the issue.
Bug: webrtc:8773
Change-Id: Ibe746a46cb41fd92b399a7069e1d89f02f292af7
Reviewed-on: https://webrtc-review.googlesource.com/38481
Commit-Queue: Qiang Chen <qiangchen@chromium.org>
Reviewed-by: Emircan Uysaler <emircan@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21737}
This prevents us from prematurely overwriting the packets in the history
if the RTT is underestimated.
Bug: webrtc:8766
Change-Id: I042e8ce74cdce2a0451596f4217779fc856b51f4
Reviewed-on: https://webrtc-review.googlesource.com/42960
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21735}
This will generate an all-zeroes track stat when the sender
has not yet been connected (SSRC has not been assigned).
Bug: webrtc:8673
Change-Id: Id59e6941bc87eba6bb33b4d2a8fd808d985052c7
Reviewed-on: https://webrtc-review.googlesource.com/43080
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Fredrik Solenberg <solenberg@webrtc.org>
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21734}
This is a reland of 1175ecd09599194caa05ecafdc4503e940809fc9.
Original change's description:
> Make it possible to run video_quality_loopback_test in swarming.
>
> I made two changes to video_quality_loopback_test to make it possible
> to run it on swarming.
>
> 1. We need to know the path to frame_analyzer when we're generating
> the build files, and it must be already present.
> I made frame_analyzer a resource, so it's downloaded to a known path
> before generating the build files.
> 2. The .zip files for apprtc and golang are downloaded and isolated.
> The script now extracts them and installs AppRTC.
>
> Passing task:
> https://chromium-swarm.appspot.com/task?id=3b230bcc04128210
>
> Bug: chromium:755660
> Change-Id: I34090897402421d5b7e29f21fbed354551197f92
> Reviewed-on: https://webrtc-review.googlesource.com/40920
> Reviewed-by: Patrik Höglund <phoglund@webrtc.org>
> Commit-Queue: Edward Lemur <ehmaldonado@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#21716}
Bug: chromium:755660
Change-Id: Ie3ca62d64b4fe856485287b7d9d3d9e3f75dc091
Reviewed-on: https://webrtc-review.googlesource.com/42860
Reviewed-by: Patrik Höglund <phoglund@webrtc.org>
Commit-Queue: Edward Lemur <ehmaldonado@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21733}
To eliminate circular dependencies, we need to eliminate the include
of media/base/mediachannel.h from api/peerconnectioninterface.h.
MediaConfig is one of the types the PeerConnection api depends on,
since it's part of PeerConnectionInterface::RTCConfiguration. It's
formally a public member, but the intention is that applications should use
accessor mehtods on RTCConfiguration and never access the contents of
MediaConfig directly.
Bug: webrtc:7504
Change-Id: Idfab6f69132d6b90d1628fa4543a393e22db79ac
Reviewed-on: https://webrtc-review.googlesource.com/41260
Commit-Queue: Niels Moller <nisse@webrtc.org>
Reviewed-by: Patrik Höglund <phoglund@google.com>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Reviewed-by: Patrik Höglund <phoglund@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21731}
In crrev.com/531028, the JNI generator starts to add heap profiler
events to JNI generated functions.
This will cause a ~80KiB regression and at the moment it is breaking
the Chromium Roll into WebRTC.
This CL defines a void macro to re-enable the Chromium Roll avoiding
the size regression.
Bug: chromium:801260
Change-Id: I9543299199c4e14b6b9b235c5cb98c0d53cf29ea
Reviewed-on: https://webrtc-review.googlesource.com/43021
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21730}
Try to use frame timestamps first if they look reasonable, otherwise
use realtime clock.
Also, lower limit from 90% of target to 85%.
Bug: webrtc:4172, chromium:802290
Change-Id: Iad489be7c7cf637345be4795e5089936ab9fab07
Reviewed-on: https://webrtc-review.googlesource.com/41041
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21729}
This CL increases the speech of the initial alignment in AEC3 by
loosening the requirements on the accuracy of the initial estimates.
Bug: webrtc:8784, chromium:804270
Change-Id: I86e2d97830843524090a1cf877965739f66dc058
Reviewed-on: https://webrtc-review.googlesource.com/40660
Commit-Queue: Per Åhgren <peah@webrtc.org>
Reviewed-by: Gustaf Ullberg <gustaf@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21728}
This reverts commit 1175ecd09599194caa05ecafdc4503e940809fc9.
Reason for revert: Breaks the bots.
Original change's description:
> Make it possible to run video_quality_loopback_test in swarming.
>
> I made two changes to video_quality_loopback_test to make it possible
> to run it on swarming.
>
> 1. We need to know the path to frame_analyzer when we're generating
> the build files, and it must be already present.
> I made frame_analyzer a resource, so it's downloaded to a known path
> before generating the build files.
> 2. The .zip files for apprtc and golang are downloaded and isolated.
> The script now extracts them and installs AppRTC.
>
> Passing task:
> https://chromium-swarm.appspot.com/task?id=3b230bcc04128210
>
> Bug: chromium:755660
> Change-Id: I34090897402421d5b7e29f21fbed354551197f92
> Reviewed-on: https://webrtc-review.googlesource.com/40920
> Reviewed-by: Patrik Höglund <phoglund@webrtc.org>
> Commit-Queue: Edward Lemur <ehmaldonado@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#21716}
TBR=phoglund@webrtc.org,ehmaldonado@webrtc.org
Change-Id: Id25d26adc547ff6f9ab178601e37527459c8b5ef
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:755660
Reviewed-on: https://webrtc-review.googlesource.com/42800
Reviewed-by: Edward Lemur <ehmaldonado@webrtc.org>
Commit-Queue: Edward Lemur <ehmaldonado@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21723}
This feature is not needed in video codec testing framework. In WebRTC
video codecs never deal with packet loss. Packet loss is handled by
jitter buffer which prevents passing of incomplete frames to decoder.
Bug: webrtc:8768
Change-Id: I211cf51d913bec6a1f935e30691661d428ebd3b6
Reviewed-on: https://webrtc-review.googlesource.com/40740
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21722}
The custom template seems to be broken. The necessary flag
"use_unprocessed_jars" has been added to dist_jar template and the
custom template is not needed anymore.
Bug: None
Change-Id: I0ca7a91ee47c8de659bcaffa5661bff55af50375
Reviewed-on: https://webrtc-review.googlesource.com/42680
Reviewed-by: Anders Carlsson <andersc@webrtc.org>
Commit-Queue: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21721}
Prior to this change, in certain circumstances the RTP header length
used when creating a RedPacket was incorrect. This was due to an
assumption that a new media packet would _always_ be added to the
UlpfecGenerator's internal media packet buffer. This is not correct,
and the fix is to keep track of whatever RTP header length that is
currently correct.
Bug: webrtc:8767
Change-Id: I6d61429a19d4693dde9330f0469d13c5dfbeac52
Reviewed-on: https://webrtc-review.googlesource.com/40600
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21720}
This allows the script to be reliably run on the bots.
Bug: webrtc:8365
Change-Id: Ie4f216403c1f4fabe8db07f261fe06d1e64be0ae
Reviewed-on: https://webrtc-review.googlesource.com/42720
Reviewed-by: Edward Lemur <ehmaldonado@webrtc.org>
Commit-Queue: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21717}
I made two changes to video_quality_loopback_test to make it possible
to run it on swarming.
1. We need to know the path to frame_analyzer when we're generating
the build files, and it must be already present.
I made frame_analyzer a resource, so it's downloaded to a known path
before generating the build files.
2. The .zip files for apprtc and golang are downloaded and isolated.
The script now extracts them and installs AppRTC.
Passing task:
https://chromium-swarm.appspot.com/task?id=3b230bcc04128210
Bug: chromium:755660
Change-Id: I34090897402421d5b7e29f21fbed354551197f92
Reviewed-on: https://webrtc-review.googlesource.com/40920
Reviewed-by: Patrik Höglund <phoglund@webrtc.org>
Commit-Queue: Edward Lemur <ehmaldonado@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21716}
This makes it possible to only inject 1 or 0 video codec factories when
consuming the API using the PeerConnectionFactory+Native header.
Bug: webrtc:7925
Change-Id: I671d8dcdbdf2198a31f3890ff6b416441bd32d48
Reviewed-on: https://webrtc-review.googlesource.com/42661
Reviewed-by: Kári Helgason <kthelgason@webrtc.org>
Commit-Queue: Anders Carlsson <andersc@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21715}
The conversational_speech_generator tool now adjusts the level of
different speech segments.
Implementation:
The Turn and MultiEndCall::SpeakingTurn structs have an extra 'gain'
member. It's read and parsed in timing.cc and put in a Turn
struct. It's put in a SpeakingTurn struct in multiend_call.cc and read
and applied to the signal in simulator.cc
Bug: webrtc:7494
Change-Id: I9b82a896eb616c8b5ef14d41dfdfd085ef1d3fbb
Reviewed-on: https://webrtc-review.googlesource.com/26280
Commit-Queue: Alex Loiko <aleloi@webrtc.org>
Reviewed-by: Alessio Bazzica <alessiob@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21714}
This minor issue was found by the UBSAN fuzzer.
We have used the Godbolt compiler explorer to check that similar
changes produce identical compiled code.
Bug: chromium:803078
Change-Id: Ib3fa38c101d7bda53d8d39062cb2c0a55144305f
Reviewed-on: https://webrtc-review.googlesource.com/42580
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Alex Loiko <aleloi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21713}
AudioFrameOperations was moved from modules/utility to audio/utility
over 1 year ago. This CL finally removes the
modules/utility/include/audio_frame_operations.h header and closes the
bug.
TBR=perkj@webrtc.org
Bug: webrtc:6548
Change-Id: I36230eadef2f8edb81008562fdc9208cf29845a4
Reviewed-on: https://webrtc-review.googlesource.com/36140
Reviewed-by: Alex Loiko <aleloi@webrtc.org>
Commit-Queue: Alex Loiko <aleloi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21712}
WebRTC is still targeting C++11, this C++14 only feature sneaked in
because the Android toolchain used to build WebRTC on the trybots
uses C++14 features.
Bug: None
Change-Id: I095fb76134dff729c72b7660cdb3d6abc4de2e0c
Reviewed-on: https://webrtc-review.googlesource.com/42501
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21711}
This reverts commit ce804ddd72f768781654a996b0f6a9551d5f2efa.
Reason for revert: chromium:800732 has been fixed.
Original change's description:
> Remove gradle from DEPS.
>
> The gradle git repo seems to be broken, so remove it from
> DEPS until it is fixed.
>
> Bug: webrtc:8724
> Change-Id: I718d3faadf9c636df8e840b0f8d32c52a73d7da4
> Reviewed-on: https://webrtc-review.googlesource.com/38600
> Reviewed-by: Patrik Höglund <phoglund@webrtc.org>
> Commit-Queue: Edward Lemur <ehmaldonado@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#21549}
TBR=phoglund@webrtc.org,ehmaldonado@webrtc.org
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: webrtc:8724
Change-Id: Iad03a9e00320831833c589aa52643225df4e32a2
No-Try: True
Reviewed-on: https://webrtc-review.googlesource.com/42480
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21707}
Various places are using "memset(ptr, 0, size)" which might get optimized
away by the compiler if "ptr" is not used afterwards. The new functions
can be used to securely clear memory instead.
Bug: None
Change-Id: I067a51d17ff84d95dc4934d46a24027fbcb4825d
Reviewed-on: https://webrtc-review.googlesource.com/35500
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Commit-Queue: Joachim Bauch <jbauch@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21706}
