This patch is a follow up to https://webrtc-review.googlesource.com/c/src/+/143177.
That patch modified the updating of CSRCS until "publishing" the frame
to the renderer, however the update was added to just after
calling renderer->OnFrame(video_frame).
This patch reverses the calls of renderer->OnFrame(video_frame)
and source_tracker_.OnFrameDelivered(video_frame.packet_infos())
so that the CSRCS are available when the frame is available.
This fixes the the flakes described in webrtc:10827 that has a
test that checks the CSRCs directly after a frame is available.
Note: an optimal/correct solution would be to update the renderer
and the source tracker in the same critical section so that they
would be available at the same time.
Bug: webrtc:10827
Change-Id: Ibf6efa832d8f2f2bcce0a9b0b946188bb67d48b1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/149171
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Chen Xing <chxg@google.com>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Jonas Oreland <jonaso@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28885}
Old way to produce this histogram was based on RtcpStatisticsCallback
reporting sent RTCP messages, with some additional processing by the
ReportBlockStats class. After this cl, to grand average fraction loss
is computed by StreamStatistician, queried by VideoReceiveStream when
the stream is closed down, and passed to ReceiveStatisticsProxy which
produces histograms.
This is a preparation for deleting the RtcpStatisticsCallback from
ReceiveStatistics.
Bug: webrtc:10679
Change-Id: Ie37062c1ae590fd92d3bd0f94c510e135ab93e8d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/147722
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28747}
[1/2] - Make new version pure-virtual, and deprecated version non-pure.
This will allow deleting the deprecated version from downstream
projects.
[2/2] - Remove deprecated version.
TBR=stefan@webrtc.org
Bug: webrtc:10336
Change-Id: Ia132ef071b1f379fc74834178e75e981ca908125
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144042
Commit-Queue: Elad Alon <eladalon@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28413}
Only remaining user is WavReader. Demote its constructor
accepting a PlatformFile to private, to refactor implementation
in a later cl.
Bug: webrtc:6463
Change-Id: I7b950be6f02073cb135dd0fab1190b9dc0de1fba
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144025
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Commit-Queue: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28410}
This change replaces the `ContributingSources`-implementation of `GetContributingSources()` and `GetSynchronizationSources()` on the video side with the spec-compliant `SourceTracker`-implementation.
The most noticeable impact is that the per-frame dictionaries are now updated when frames are delivered to the RTCRtpReceiver's MediaStreamTrack rather than when RTP packets are received on the network.
Bug: webrtc:10545
Change-Id: I895b5790280ac94c1501801d226c643633c67349
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/143177
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Commit-Queue: Chen Xing <chxg@google.com>
Cr-Commit-Position: refs/heads/master@{#28386}
Currently, if LNTF and NACK messages are both created, they will
be sent out in separate RTCP messages. This is wasteful.
This CL is the first of in a series of CLs that will ensure that
these feedback messages can be buffered together, without introducing
more of a delay than the CPU time required to process both messages.
Bug: webrtc:10336
Change-Id: I950324112ee346695a12a17d025483ea5e99c732
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/139112
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Commit-Queue: Elad Alon <eladalon@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28136}
This is a partial revert of
https://webrtc-review.googlesource.com/c/src/+/130101.
The KeyFrameRequestSender argument is added back to the constructor of
RtpVideoStreamReceiver. It is optional; if a null pointer is passed,
key frame requests are sent via the internal RtpRtcp module, and this is
how the class is used by VideoReceiveStream. An injectable
KeyFrameRequestSender is useful for tests, for downstream applications
that want to route key frame requests elsewhere, and should also aid
later migration to RtcpTransciever.
Bug: None
Change-Id: Idf9baeed21570625ad74e9afbe38f7ea5bf79feb
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/139107
Commit-Queue: Niels Moller <nisse@webrtc.org>
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28102}
Currently used ':' is bad because it prevents from specifying absolute
path on windows (e.g. "C:\directory").
Now to specify path on windows, it can be passed unchanged in field trial:
"WebRTC-DecoderDataDumpDirectory/C:\path\on\windows/"
On linux ';' has to be used instead of '/':
"WebRTC-DecoderDataDumpDirectory/;path;on;linux/"
Bug: none
Change-Id: Ia46c94bdfab95385618dde4fd3f2857e9ddf2d1c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/138832
Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Elad Alon <eladalon@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28085}
Currently we pass media_transport from PeerConnection to media layers. The goal of this change is to replace media_transport with struct MediaTransportCondif, which will enable adding different transports (i.e. we plan to add DatagramTransport) as well as other media-transport related settings without changing 100s of files.
TODO: In the future we should consider also adding rtp_transport in the same config, but it will require a bit more work, so I did not include it in the same change.
Bug: webrtc:9719
Change-Id: Ie31e1faa3ed9e6beefe30a3da208130509ce00cd
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/137181
Commit-Queue: Anton Sukhanov <sukhanov@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Fredrik Solenberg <solenberg@webrtc.org>
Reviewed-by: Steve Anton <steveanton@webrtc.org>
Reviewed-by: Bjorn Mellem <mellem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28016}
They are called only from VideoReceiveStream, which can access
VCMTiming directly.
Bug: webrtc:7408
Change-Id: Ibf5799b1441c00b41143342ca1d99024cb68ba17
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/133569
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Commit-Queue: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27700}
There are unexpected changed reported from the perf bots,
disabling task queue mode while investigating.
Bug: webrtc:950335
Change-Id: I280605a8fc8a0f97f7a4b90f53a08e087b61cdfe
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/132710
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Sebastian Jansson <srte@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27595}
This is a reland of 13943b7b7f6d00568912b9969db2c7871d18e21f
Original change's description:
> Running FrameBuffer on task queue.
>
> This prepares for running WebRTC in simulated time where event::Wait
> based timing doesn't work.
>
> Bug: webrtc:10365
> Change-Id: Ia0f9b1cc8e3c8c27a38e45b40487050a4699d8cf
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/129962
> Reviewed-by: Philip Eliasson <philipel@webrtc.org>
> Reviewed-by: Erik Språng <sprang@webrtc.org>
> Commit-Queue: Sebastian Jansson <srte@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#27422}
Bug: webrtc:10365
Change-Id: I412d3e0fe06c6dd57cdb42974f09e03f3a6ad038
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/131124
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Sebastian Jansson <srte@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27572}
profile-level-id for H.264 comes in through the SdpVideoFormat,
rather than through these members.
Bug: None
Change-Id: I9c4ea8873346ca16174aecf5f90a649cbaf913dd
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/132545
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27571}
This change introduces new logic to allow the injection of the FrameDecryptor
into an arbitrary already running VideoReceiveStream without resetting it. It
does this by taking advantage of the BufferedFrameDecryptor which will
forcefully be created regardless of whether a FrameDecryptor is passed in
during construction of the VideoReceiver if the
crypto_option.require_frame_encryption is true. By allowing the
BufferedFrameDecryptor to swap out which FrameDecryptor it uses this allows the
Receiver to switch decryptors without resetting the stream.
This is intended to mostly be used when you set your FrameDecryptor at a point
post creation for the first time.
Bug: webrtc:10416
Change-Id: If656b2acc447e2e77537cfa394729e5c3a8b660a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/130361
Commit-Queue: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Jonas Oreland <jonaso@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27458}
Extracting the work that's thread dependent from the work that will
also be done when using task queue.
Bug: webrtc:10365
Change-Id: I648796fe016c966c731c9b7f85d2a871c1f2a349
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/131241
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Commit-Queue: Sebastian Jansson <srte@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27454}
This prepares for running WebRTC in simulated time where event::Wait
based timing doesn't work.
Bug: webrtc:10365
Change-Id: Ia0f9b1cc8e3c8c27a38e45b40487050a4699d8cf
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/129962
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Sebastian Jansson <srte@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27422}
And delete corresponding dependencies on :webrtc_common. After this
change, common_types.h is included directly only from code in the
following directories:
api/
api/video/
api/video_codecs/
common_video/libyuv/include/
media/base/
modules/remote_bitrate_estimator/
modules/rtp_rtcp/source/
modules/video_coding/codecs/vp9/
There remains plenty of indirect dependencies on the types declared in
common_types.h, but the fewer direct dependencies should make it
easier to find the proper place for each type.
Bug: webrtc:5876
Change-Id: I93e8f214025ecb613c19fdec2015bd3f96c59aae
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/130501
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Commit-Queue: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27376}
VCMReceiveStatisticsCallback originates in the old jitter buffer, and
is no longer used.
VCMFrameTypeCallback originates in VideoReceiver::RequestKeyFrame,
which is called from OncomingPacket, Process, Decode(uint16_t
maxWaitTimeMs), all of which are unused by VideoReceiveStream.
So delete the code to wire them up via VideoStreamDecoder.
Bug: webrtc:7408
Change-Id: I173bc94eb32f2641f943c125083db038c3bcaeb1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/128870
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27277}
It's used for driving the old jitter buffer, which is used only when
vcm::VideoReceiver is used via the legacy VideoCodingModule api.
Bug: webrtc:7408
Change-Id: I179d5b26e112d9f94615d2e1b410b51a657aa05b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/127294
Commit-Queue: Niels Moller <nisse@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27147}
The video decoder thread is the pilot user.
For now this is an Android-only feature, since that's the only
platform we can print stack traces on.
Bug: webrtc:9987
Change-Id: Ie638c619673b5f159d91a32683fd787baf46479a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/126222
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27127}
Specifying relative path like "~" somehow doesn't work in some cases.
To pass an absolute path some parameter rewriting is required.
Also, as the stream gets recreated several times at the beginning of the
call, append the time to the filename to make them unique.
Bug: none
Change-Id: I1c914f9081adb4ac5c34584d96e542e8a863547b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/126700
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27101}
The current Key Frame request system doesn't take into account failed
decryptions and this can lead to WebRTC spamming new key frame requests when
the issue is actually in the decryptor layer. To prevent this if frame
decryption is required for the PeerConnection key frame requests will not be
sent at 200ms intervals but will wait until the stream is decryptable before
utilizing this logic.
Bug: webrtc:10330
Change-Id: I188a21dfd142dec6175d9def95f39a2bc517017c
Reviewed-on: https://webrtc-review.googlesource.com/c/123414
Commit-Queue: Benjamin Wright <benwright@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26931}
On api level two methods were added to api/media_stream_interface.cc on VideoSourceInterface,
GetLatency and SetLatency. Latency is measured in seconds, delay in milliseconds but both describes
the same concept.
Bug: webrtc:10287
Change-Id: Ib8dc62a4d73f63fab7e10b82c716096ee6199957
Reviewed-on: https://webrtc-review.googlesource.com/c/123482
Commit-Queue: Ruslan Burakov <kuddai@google.com>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Reviewed-by: Steve Anton <steveanton@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26877}
Further, strictly require VideoReceiveStream::Config::rendererer
to be non-null when the VideoReceiveStream is started. This is
already true by construction in the production code.
Bug: None
Change-Id: Ia0a41cfafa44215efc195a9eb6204194930c3dde
Reviewed-on: https://webrtc-review.googlesource.com/c/115040
Commit-Queue: Rasmus Brandt <brandtr@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26384}
