This resolves an issue where when packets appear out of order at the
beginning of a stream, packet_buffer.cc might drop the entire packet
buffer because it detects a "large negative jump" even though the
difference in sequence numbers is very minor and is caused by network
congestion / packet re-ordering. Currently, when the issue occurs, this
can cause video corruption/artifacts. More details and reproduction is
available on the attached webrtc bug report 390329776.
Bug: webrtc:390329776
Change-Id: Idb56eb2e066d596d8afd7ec904359baf0cb3feef
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/374540
Reviewed-by: Sergey Silkin <ssilkin@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43753}
Matlab files extension is the same as ObjC, which is .m
This makes clang-format think that those files are ObjC and then it
wrongly formats them, leading to output that doesn't compile at all.
It's a known issue and the solution is to disable it in Matlab files.
I don't want to disable ObjC in whole folders, because of 2 reasons:
1) I want ObjC to be properly formatted if new files are added in the
future
2) C++ header files are interpreted as ObjC and it will disable their
formatting
According to clang documentation
(https://clang.llvm.org/docs/ClangFormatStyleOptions.html#disabling-formatting-on-a-piece-of-code), we can disable formatting inline.
However, comments in Matlab are prefixed with `%` and not `//`, so I
thought of a kinda hacky solution, which is `% // clang-format off`, and
it works perfectly.
No-Iwyu: Includes didn't change and it isn't related to formatting
Bug: webrtc:42225392
Change-Id: I281462fd1aecd3ff0428e6ee974514ebabc696ec
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/374060
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43700}
After applying this change, I reformatted the previously formatted files
and verified no changes were applied.
Bug: webrtc:42225392
Change-Id: I6079e1e85d94ae2bc892db1db81bc8223b3a08b7
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/374040
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43698}
Update unit tests to follow video_rtp_depacketizer_h264's behavior of
setting is_first_packet_in_frame. This flag might be used to determine
if a frame can be assembled.
Bug: webrtc:384391181
Change-Id: I6750c20056e426e12c1d4e21eea4c641def7cfbd
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/373168
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Sergey Silkin <ssilkin@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43669}
Follow H.265 spec section 7.4.2.4.4 to set is_first_packet_in_frame flag
in rtp header, to make sure we only set it to true when corresponding
NALU can be the first NALU in a frame.
Bug: chromium:384391181
Change-Id: I082c38513d9d213f8d354633539028b57777368f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/372742
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Reviewed-by: Sergey Silkin <ssilkin@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43651}
When echo controller factories are updated, it would be possible to pass Environment into EchoCanceller3 and thus rely on propagated field trials.
Bug: webrtc:369904700
Change-Id: Iba9c04edbaab23277874234bd289e2c37625b1c8
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/372040
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43614}
There is one RTCP receiver per receive stream. Therefore, only handle a
received CongestionControlFeedback in the RTCP receiver corresponding to
the first SSRC in the report.
Bug: webrtc:42225697
Change-Id: I9bc0009cb6840cddeaca25f39c597bc2c13a3604
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/372280
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43613}
With this CL, the decision if an RTP packet should be sent as ect(1) is made in RtpControllerSend depending on if RFC 8888 has been negotiated and if CCFB is received with ECN enabled.
Since webrtc does not yet adapt to ECN feedback, packets are sent as ECT(1) until the first feedback is received.
Change-Id: Iddf63849328afbe54a7c8f921f2e8db134aeff6a
Bug: webrtc:42225697
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/367388
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43609}
This is to fix build error when we set use_libcxx_modules=true in
chromium build.
Bug: chromium:40440396
Change-Id: Iad165a78a6920ccb858567d31fbe5e48d8a7b629
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/371620
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Henrik Lundin <henrik.lundin@webrtc.org>
Auto-Submit: Takuto Ikuta <tikuta@google.com>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43574}
This reverts commit 200fd82771ae29d23b2be40194be674b3437f0ab.
Reason for revert: breaks downstream
Original change's description:
> Validate frame consistency when writing DependencyDescriptor
>
> To write DependencyDescriptor frame properties should be consistent with
> the FrameDependencyStructure.
> Historically that was ensured by webrtc codec wrappers, but with with frame transform api interface there are now more ways to inject video frame for packetizing.
> Thus DependencyDescriptorWriter should be more protective to avoid crashes.
>
> Bug: chromium:379282549
> Change-Id: I98f226ff09c32154e18888c8e811e7981567ad45
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/371301
> Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
> Reviewed-by: Åsa Persson <asapersson@webrtc.org>
> Cr-Commit-Position: refs/heads/main@{#43551}
Bug: chromium:379282549
Change-Id: I7711756f774648cbb85c51b61424bb950c1d3775
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/371420
Commit-Queue: Jeremy Leconte <jleconte@webrtc.org>
Owners-Override: Jeremy Leconte <jleconte@webrtc.org>
Bot-Commit: rubber-stamper@appspot.gserviceaccount.com <rubber-stamper@appspot.gserviceaccount.com>
Cr-Commit-Position: refs/heads/main@{#43556}
BuiltinAudioProcessingBuilder should be used instead.
This would allow AudioProcessingImpl to have Environment construction parameter and thus use propagated rather than global field trials.
Bug: webrtc:369904700
Change-Id: I4fcc299bb9e65c109a3fe476c755a81c2aea551c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/368480
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43553}
To write DependencyDescriptor frame properties should be consistent with
the FrameDependencyStructure.
Historically that was ensured by webrtc codec wrappers, but with with frame transform api interface there are now more ways to inject video frame for packetizing.
Thus DependencyDescriptorWriter should be more protective to avoid crashes.
Bug: chromium:379282549
Change-Id: I98f226ff09c32154e18888c8e811e7981567ad45
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/371301
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43551}
The `FrameToRender` method is deprecated and has been replaced by
`OnFrameToRender`.
Bug: webrtc:358039777
Change-Id: Ibe56bd43cf045d814137ba8c4374bc9b9ce8ef6c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/371302
Commit-Queue: Fanny Linderborg <linderborg@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43547}
This is to fix build error when we set use_libcxx_modules=true in
chromium build.
Bug: chromium:40440396
Change-Id: I5ab1cfcc0d060021892aae0e5ff3f0b647ae4266
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/370860
Commit-Queue: Takuto Ikuta <tikuta@google.com>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Tomas Gunnarsson <tommi@webrtc.org>
Auto-Submit: Takuto Ikuta <tikuta@google.com>
Cr-Commit-Position: refs/heads/main@{#43541}
Based on the description, this dependency have no meaningful upstream,
and is maintained inside webrtc.
Marking this dependency's URL to indicate the webrtc's repo is the
canonical repo.
Fixed: chromium:362397270
Change-Id: If6e16a6e34e0083be31d4436fcdfa7c83cd9179a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/370980
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Tomas Gunnarsson <tommi@webrtc.org>
Auto-Submit: Jiewei Qian <qjw@google.com>
Reviewed-by: Tomas Gunnarsson <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43535}
Fix todo to ensure TransportSequence numbers are generated if CCFB according to RFC 8888 is used. Transport sequence numbers are used in BWE algorithms regardless of feedback format.
Bug: webrtc:42225697
Change-Id: I6eab95c0241d590f6e7a90d19c82d13ab8692f2b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/370341
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43515}
Further, add use of it in libvpx_vp8_encoder and with tuning for keyframes and lower bound of std_dev = 1.25 to work around some edge cases. Plus some minor cleanup.
Bug: webrtc:358039777
Change-Id: I6f624a6a8c7ccfe2fe656e4c089c225296f0264f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/370061
Reviewed-by: Sergey Silkin <ssilkin@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43513}
Set use_default_launcher=false in rtc_test on android
Bug: webrtc:42223878
Change-Id: If05da40b420d5da8f9e0f39560eb07380ebada14
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/368921
Owners-Override: Jeremy Leconte <jleconte@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Jeremy Leconte <jleconte@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Jeremy Leconte <jleconte@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43505}
Previous computation assumed that local clock is UTC. It isn't.
Adding integration test for abs-capture stats.
Bug: webrtc:380712819
Change-Id: I054d61984cbd017b7ad04ab13e5a687eab89db69
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/369421
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Johannes Kron <kron@webrtc.org>
Reviewed-by: Tomas Gunnarsson <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43465}
0 is natural default value for types that can be accumulated
Having default constructor simplify usage of these types in templated code.
Bug: None
Change-Id: If005c69018a2a11011bc789502fdbc600cad3278
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/369440
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43460}
This is a reland of commit 38ddea5ee3320bf3441aeb3654e099b3695c9789
Original change's description:
> Add InsertPacket method that takes RtpPacketInfo.
>
> The version which only passes receive_time will be removed (once migrated).
> Keeping the version that only passes header and payload for convenience.
>
> This will allow us to attach more metadata on the worker thread before InsertPacket, instead of on the playout thread after GetAudio. Eventually, the plan is to split the RTP handling on the worker thread into a separate class.
>
> Bug: webrtc:42223109
> Change-Id: I5399b53b9fc5c2f1c996e109054b1b0877ecca05
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/369000
> Reviewed-by: Henrik Lundin <henrik.lundin@webrtc.org>
> Commit-Queue: Jakob Ivarsson <jakobi@webrtc.org>
> Cr-Commit-Position: refs/heads/main@{#43445}
Bug: webrtc:42223109
Change-Id: I97d1d3d390e6d3de8bf9355b895ec336339d079f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/369260
Reviewed-by: Henrik Lundin <henrik.lundin@webrtc.org>
Commit-Queue: Jakob Ivarsson <jakobi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43454}
This reverts commit 38ddea5ee3320bf3441aeb3654e099b3695c9789.
Reason for revert: not backwards compatible
Original change's description:
> Add InsertPacket method that takes RtpPacketInfo.
>
> The version which only passes receive_time will be removed (once migrated).
> Keeping the version that only passes header and payload for convenience.
>
> This will allow us to attach more metadata on the worker thread before InsertPacket, instead of on the playout thread after GetAudio. Eventually, the plan is to split the RTP handling on the worker thread into a separate class.
>
> Bug: webrtc:42223109
> Change-Id: I5399b53b9fc5c2f1c996e109054b1b0877ecca05
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/369000
> Reviewed-by: Henrik Lundin <henrik.lundin@webrtc.org>
> Commit-Queue: Jakob Ivarsson <jakobi@webrtc.org>
> Cr-Commit-Position: refs/heads/main@{#43445}
Bug: webrtc:42223109
Change-Id: Ie7cf397cfbe5dedca009f16e5e9e3af40adbe99b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/369200
Reviewed-by: Henrik Lundin <henrik.lundin@webrtc.org>
Bot-Commit: rubber-stamper@appspot.gserviceaccount.com <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Jakob Ivarsson <jakobi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43450}
If max allocated bitrate change, default max limit probe to 2x current
BWE.
Bug: webrtc:369044000, b/370883514
Change-Id: Ibaf79fff94157186002728828d6574bea21afd24
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/368820
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43448}
The version which only passes receive_time will be removed (once migrated).
Keeping the version that only passes header and payload for convenience.
This will allow us to attach more metadata on the worker thread before InsertPacket, instead of on the playout thread after GetAudio. Eventually, the plan is to split the RTP handling on the worker thread into a separate class.
Bug: webrtc:42223109
Change-Id: I5399b53b9fc5c2f1c996e109054b1b0877ecca05
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/369000
Reviewed-by: Henrik Lundin <henrik.lundin@webrtc.org>
Commit-Queue: Jakob Ivarsson <jakobi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43445}
- Add `WebRTC-Audio-OpusDecodeStereoByDefault` field trial
- Behind that field trial, `AudioDecoderOpus::SdpToConfig` uses 2
instead of 1 as default number of channels when the `stereo` codec
param is unspecified
- Instead of wiring up `FieldTrialsView` to `SdpToConfig`, which
requires API changes that break downstream projects, a change in
`AudioDecoderOpus::Config` is made to signal when the number of
channels is forced via SDP config
Bug: webrtc:379996136
Change-Id: If70eb19bc7e3bc74dd0423610cb04ae33ea602fe
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/368860
Commit-Queue: Alessio Bazzica <alessiob@webrtc.org>
Reviewed-by: Tomas Gunnarsson <tommi@webrtc.org>
Reviewed-by: Jakob Ivarsson <jakobi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43440}
Too big median window will cause errors with large clock drifts, since we'll end up using old values for estimated clock drift.
If the window is too small, the remote clock offset estimation could be noisy or we could even end up using outliers as the offset estimation.
I will not claim that I choose the correct value, and I'm not sure how to measure the quality of the remote clock offset estimations.
Bug: webrtc:379809147
Change-Id: Ib317548d3eec74105d468ef53830e12eb114df7d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/368580
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Olov Brändström <brandstrom@google.com>
Cr-Commit-Position: refs/heads/main@{#43439}
This updates test code that tests interleaved audio frames to use
some of the same properties and types as AudioFrame (rather than copy).
The CL also moves code from audio_processing_unittest.cc that modifies
the buffer owned by Int16FrameData, into Int16FrameData.
Bug: none
Change-Id: Iab37227deb302bf4fc832633d312262e5249caad
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/355960
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Tomas Gunnarsson <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43424}
