If rate controller is trusted, we disable the frame dropper in the
media optimization module.
Bug: webrtc:9722
Change-Id: I821f21fd74a400ee9d5aa3f6b42d4e569033acbe
Reviewed-on: https://webrtc-review.googlesource.com/c/105020
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#25107}
I don't think this has any impact, just wanted to have a first unit
test to play around with.
Bug: None
Change-Id: I892e2642f0243c5c9ee807cf71abcd96112b25f4
Reviewed-on: https://webrtc-review.googlesource.com/c/105000
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#25089}
Set number of decode threads equal to number of available cores and
limit the maximum value to the maximum number of tiles possible for
HD resolution.
Bug: webrtc:9829, b/117291409
Change-Id: Ib5ccd5cc412011d4438258491efc060cdd050fc7
Reviewed-on: https://webrtc-review.googlesource.com/c/104064
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#25059}
Add FakeVp8Decoder that parse width and height from the payload.
Add unit test for testing that width and height is set when decoding frames.
Bug: none
Change-Id: Ifbfff4f62f99625309ce0ef21cf89c76448769d8
Reviewed-on: https://webrtc-review.googlesource.com/c/103140
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Sebastian Jansson <srte@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#25038}
- Added field trial to force issuing of key frame on deactivation of
spatial layer. This fixes video corruptions in VP9 K-SVC tests where
layers can be activated/deactivated on-fly due to bandwidth change.
- Added 100ms network delay to the test with restricted link capacity.
This fixes rapid drop of available bandwidth which happens when
bandwidth overuse is detected in the very beginning of call and several
feedback packets arrive without any delay. Also, this makes the test
more realistic.
- Disabled filtering of spatial layer in the test with restricted
link capacity. 1) We don't really need filtering in this test.
2) It appeared that in video quality tests filtering is done before
sending packets to network simulator. Filtering of high layers causes
channel underuse which is compensated by increase of sent bitrate.
This is why we got sent/media bitrates about 2Mbps in test where link
was limited to 1Mbps.
Bug: chromium:889017
Change-Id: I33ffcee0274523f6183c3bbd27d3d29395417d52
Reviewed-on: https://webrtc-review.googlesource.com/c/103520
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24988}
This refactoring merged PopulateCodecSpecific and FrameEncoded into a
single callback method. It also removes the FrameConfig parameter and
instead relies on the temporal layer to remember it internally.
Bug: webrtc:9012
Change-Id: I489b76821b534398ad452643f1322f411d3455b1
Reviewed-on: https://webrtc-review.googlesource.com/95681
Reviewed-by: Sebastian Jansson <srte@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24957}
This CL makes the RtpGenericFrameDescriptor available in
RTPSenderVideo::SendVideo for encryption and in
RtpVideoStreamReceiver::OnReceivedFrame for decryption.
Bug: webrtc:9361
Change-Id: I5b6d10138c0874657862f103c8c9a2328e6d4a66
Reviewed-on: https://webrtc-review.googlesource.com/102720
Commit-Queue: Philip Eliasson <philipel@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24929}
Use in VideoQualityTest replaced by creating a wrapper for the decoder,
similarly to https://webrtc-review.googlesource.com/94152 which
deleted the corresponding method on VideoSendStream.
Bug: webrtc:9106
Change-Id: I0a7798bc44704af8b36017655b9ffa34fa1423e6
Reviewed-on: https://webrtc-review.googlesource.com/97580
Commit-Queue: Niels Moller <nisse@webrtc.org>
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24926}
This CL is a step towards making the TemporalLayers landable in api/ :
* It splits TemporalLayers from TemporalLayersChecker
* It initially renames temporal_layer.h to vp8_temporal_layers.h and
moved it into the include/ folder
* It removes the dependency on VideoCodec, which was essentially only
used to determine if screenshare_layers or default_temporal_layers
should be used, and the number of temporal temporal layers to use.
Subsequent CLs will make further cleanup before attempting a move to api
Bug: webrtc:9012
Change-Id: I87ea7aac66d39284eaebd86aa9d015aba2eaaaea
Reviewed-on: https://webrtc-review.googlesource.com/94156
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24920}
In the case a frame_object is kept for some time before it is deleted,
it may happend that a new frame is received with overlapping sequence
numbers. If the old frame_object is removed while receiving the new
frame there used to be a crash.
Bug: webrtc:9629
Change-Id: I270a8caa2b58b73c000542aa504c0ebe277d49c4
Reviewed-on: https://webrtc-review.googlesource.com/102683
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Commit-Queue: Johannes Kron <kron@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24914}
This CL removes some deprecated build targets (and their headers)
from system_wrappers:
- field_trial_api
- field_trial_default
- metrics_api
- metrics_default
It also refreshes all the dependencies on field_trial.h and metrics.h.
A nice side effect is that it is finally possible to remove 'nogncheck'
from the following files (when it was used with field_trial_default
and metrics_default):
- sdk/objc/api/peerconnection/RTCMetricsSampleInfo+Private.h
- sdk/android/src/jni/pc/peerconnectionfactory.cc
- sdk/objc/api/peerconnection/RTCFieldTrials.mm
Bug: webrtc:9631
Change-Id: Ib621f41ef8ad0aba4fe1c1d7e749c044afc956c3
No-Try: True
Reviewed-on: https://webrtc-review.googlesource.com/100524
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24878}
The RTC_DCHECK is hit sometimes. This happens when there is no overlap
between the nack_list and frames in keyframes. The existing code
correctly handles this situation.
Bug: webrtc:9629
Change-Id: I7e3eed1b04781cd69974c5d3eb86e382e9587268
Reviewed-on: https://webrtc-review.googlesource.com/102340
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Commit-Queue: Johannes Kron <kron@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24860}
Mark all low spatial layer frames as references (not just frames of
active low spatial layers) if inter-layer prediction is enabled since
these frames are indirect references of high spatial layer, which can
later be enabled without key frame.
Bug: webrtc:9782
Change-Id: Iffa5039fab2673a5582e7cdc9be4a36d9e8deb63
Reviewed-on: https://webrtc-review.googlesource.com/102063
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24849}
Replace calls to .str() which copies with .Release which moves in cases where that's safe.
This CL was generated by this command:
git grep -l 'StringBuilder' |
xargs perl -i -0 -pe "s/(rtc::StringBuilder (\S+);.*?return )\\g2.str\(\)/\$1\$2.Release\(\)/sg"
Bug: webrtc:8982
Change-Id: If4dadbeb039df010aaaa9e58da81c1971a84fe8f
Reviewed-on: https://webrtc-review.googlesource.com/100307
Commit-Queue: Jonas Olsson <jonasolsson@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24790}
VP9 frame rate controller is supposed to be used in screen mode only
but it was partially enabled in normal video mode. This restricts use
of VP9 frame rate controller to screen mode.
Bug: chromium:884164
Change-Id: Ie2eaa31f3364a8abccbc4171007708cf7040fc38
Reviewed-on: https://webrtc-review.googlesource.com/100424
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24769}
Uninline RTPFragmentaion functions
fix RTPFragmentation move constructor and assign operators (was recursive for win)
replace assert with rtc::dchecked_cast
Remove unused includes and dependencies.
Fix other targets that used those includes transitively instead of directly
Bug: None
Change-Id: I647cb1eda107dc7d87d25234095545bc2842fa40
Reviewed-on: https://webrtc-review.googlesource.com/100500
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24759}
If the bandwidth is just on the edge of being able to enable a new
stream, the keyframe generated when it is enabled might be large enough
to trigger an overuse and force the stream off again.
To avoid toggling, this CL adds hysteresis so that the available
bandwidth needs to be above X% to start bitrate in order to enable the
stream. It will be shut down once available bitrate falls below the
original enabling bitrate.
For screen content, X defaults to 35.
For realtime content, X defaults to 0.
Both can be individually modified via field trials.
Bug: webrtc:9734
Change-Id: I941332d7be7f2a801d13d9202b2076d330e7df32
Reviewed-on: https://webrtc-review.googlesource.com/100308
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24745}
Today we use |is_first_packet_in_frame| to know when a frame begins and the
|markerBit| to know when it ends, but the markerbit does not actually mark the
end of a frame, it marks the end of a picture.
Bug: webrtc:9361
Change-Id: Icc70e6075590cdc31e875a4eb9d489868adbb67c
Reviewed-on: https://webrtc-review.googlesource.com/100160
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Philip Eliasson <philipel@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24722}
Set target frame rate of spatial layer equal to minimum of two: maximum
frame rate of layer (SpatialLayer::maxFramerate) and maximum frame rate
of codec (VideoCodec::maxFramerate).
Bug: webrtc:9740, webrtc:9739, chromium:882358
Change-Id: I34f36e7fd2889f0417474347abab5327fa2d9d7c
Reviewed-on: https://webrtc-review.googlesource.com/99501
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24686}
These attributes were moved from CodecSpecificInfo to EncodedImage in
cl https://webrtc-review.googlesource.com/c/src/+/96780. This followup
deletes the old member variables, which were left temporarily to
transition downstream code.
Bug: webrtc:9378
Change-Id: I1b38ce404a005aec9d48916b73233cfbd7523cfe
Reviewed-on: https://webrtc-review.googlesource.com/97021
Reviewed-by: Sergey Silkin <ssilkin@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Commit-Queue: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24670}
This is a reland of ae9e188e67
Original change's description:
> Frame rate controller per spatial layer.
>
> This allows VP9 encoder wrapper to control frame rate of each spatial
> layer. The wrapper configures encoder to skip encoding spatial layer
> when actual frame rate exceeds the target frame rate of that layer.
> Target frame rate of high spatial layer is expected to be equal or
> higher then that of low spatial layer. For now frame rate controller
> is only enabled in screen sharing mode.
>
> Added unit test which configures encoder to produce 3 spatial layers
> with frame rates 10, 20 and 30fps and verifies that absolute delta of
> final and target rate doesn't exceed 10%.
>
> Bug: webrtc:9682
> Change-Id: I7a7833f63927dd475e7b42d43e4d29061613e64e
> Reviewed-on: https://webrtc-review.googlesource.com/96640
> Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
> Reviewed-by: Erik Språng <sprang@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#24593}
TBR=sprang@webrtc.org
Bug: webrtc:9682
Change-Id: Idcce315890c79301da532f9ba4997e9606f73fb0
Reviewed-on: https://webrtc-review.googlesource.com/99340
Reviewed-by: Sergey Silkin <ssilkin@webrtc.org>
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24669}
The declaration in common_types.h is probably a left-over from a
previous cleanup.
Bug: None
Change-Id: I3ee1bad2494ede0022c6aa8fdd106035471d50e2
Reviewed-on: https://webrtc-review.googlesource.com/99220
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24666}
This reverts commit ae9e188e67.
Reason for revert: Verify if this causes chromium:882358.
Original change's description:
> Frame rate controller per spatial layer.
>
> This allows VP9 encoder wrapper to control frame rate of each spatial
> layer. The wrapper configures encoder to skip encoding spatial layer
> when actual frame rate exceeds the target frame rate of that layer.
> Target frame rate of high spatial layer is expected to be equal or
> higher then that of low spatial layer. For now frame rate controller
> is only enabled in screen sharing mode.
>
> Added unit test which configures encoder to produce 3 spatial layers
> with frame rates 10, 20 and 30fps and verifies that absolute delta of
> final and target rate doesn't exceed 10%.
>
> Bug: webrtc:9682
> Change-Id: I7a7833f63927dd475e7b42d43e4d29061613e64e
> Reviewed-on: https://webrtc-review.googlesource.com/96640
> Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
> Reviewed-by: Erik Språng <sprang@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#24593}
TBR=sprang@webrtc.org,ssilkin@webrtc.org
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: webrtc:9682, chromium:882358
Change-Id: Idc4051eef72104823038ed9139bb9c75018f7d86
Reviewed-on: https://webrtc-review.googlesource.com/99082
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Reviewed-by: Sergey Silkin <ssilkin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24646}
Today, the internal frame dropper in libvpx vp8 encoder is enabled or
disabled based on video or screen content. This is then expected to
match up with screenshare vs default temporal layers implementation.
This cl makes libvpx query the temporal layers implementation as well,
breaking this implicit dependency and allows frames to be dropped if
default temporal layers is used with screen content.
Bug: webrtc:9734
Change-Id: If2523a211f4929f16e65a02fa7a6b4edf7328571
Reviewed-on: https://webrtc-review.googlesource.com/99062
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#24645}
