The |slice_qp_detla| reported by the hardware is not credible, which
causing the quality scaler cannot work properly,the resolution cannot
be adjusted correctly.
To fix this issue, this CL implements a bandwidth scaler which is used
for adjust resolution, this scaler will be used when QP based quality
scaler is not working due to untrusted QP reported by HW AVC encoder.
Bug: webrtc:12942
Change-Id: I2fc5f07a5400ec7e5ead2c2c502faee84d7f2a76
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/228860
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Evan Shrubsole <eshr@google.com>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35120}
min_pacing:8ms, to avoid the situation where bursts of frames are sent
to the decoder at once due to network jitter. The bursts of frames
caused the queues further down in the processing to be full and
therefore drop all frames.
max_decode_queue_size:8, in the event that too many frames have piled
up, do as before and send all frames to the decoder to avoid building
up any latency.
These setting only affect the low-latency video pipeline that is enabled
by setting the playout RTP header extension to min=0ms, max>0ms.
Bug: chromium:1138888
Change-Id: I8154bf3efe7450b770da8387f8fb6b23f6be26bd
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/233220
Commit-Queue: Johannes Kron <kron@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35119}
Instead of using two different headroom parameters, namely
`kHeadroomDbfs` and `kSaturationProtectorExtraHeadroomDb`, only use
the former that now also accounts for the deleted one - i.e., it equals
the sum of the two headrooms. In this way, tuning AGC2 will be easier.
This CL does *not* change the behavior of the AGC2 adaptive digital
controller - bitexactness verified with audioproc_f on a collection of
AEC dumps and Wav files (42 recordings in total).
The unit tests changes in agc2/saturation_protector_unittest.cc are
required since `extra_headroom_db` is removed and the changes in
agc2/adaptive_digital_gain_applier_unittest.cc are required because
`AdaptiveDigitalGainApplier` depends on `kHeadroomDbfs` which has been
updated as stated above.
Bug: webrtc:7494
Change-Id: I0a2a710bbede0caa53938090a004d185fdefaeb9
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/232905
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Alessio Bazzica <alessiob@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35109}
Currently the implementation of FrameTransformers uses distinct,
incompatible types for recevied vs about-to-be-sent frames. This adds a
flag in the interface so we can at least check that we are being given
the correct type. crbug.com/1250638 tracks removing the need for this.
Chrome will be updated after this to check the direction flag and provide
a javascript error if the wrong type of frame is written into the
encoded insertable streams writable stream, rather than crashing.
Bug: chromium:1247260
Change-Id: I9cbb66962ea0718ed47c5e5dba19a8ff9635b0b1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/232301
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Tony Herre <toprice@chromium.org>
Cr-Commit-Position: refs/heads/main@{#35100}
This CL improves `GainController2::CheckGainAdaptiveDigital`, namely:
- correctly initialize AGC2 with the correct number of channels
- attenuate the input signal in order to avoid that the target gain is
set to zero (which was the case before)
- run AG2 adaptive digital for a longer period to allow time to trigger
the adaptive behavior (namely, from 2s to 10s)
- minor code style improvements
Bug: webrtc:7494
Change-Id: Ib41de088b341bb30460238b83e306a507b2bc5af
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/233101
Commit-Queue: Alessio Bazzica <alessiob@webrtc.org>
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35099}
CreateSharedMemory is allowed to return nullptr if memory can't be
allocated but DesktopFrameWin didn't check to ensure was allocated
before accessing it. This CL just adds a null check, logs a
warning, and returns nullptr which is already done lower in the
function and the error is correctly handled in the screen and window
capturers which call DesktopFrameWin::Create().
Bug: chromium:1251651
Change-Id: Ie9231f03ba9c7a96823af986b9df38f97fcb682c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/232663
Reviewed-by: Jamie Walch <jamiewalch@chromium.org>
Commit-Queue: Joe Downing <joedow@chromium.org>
Cr-Commit-Position: refs/heads/main@{#35072}
This one is frequently accessed - Mainly by ::CreateReportBlocks and
is visible in performance profiles (although not very much).
By using webrtc::flat_map, better data cache locality is expected.
Bug: webrtc:12689
Change-Id: Ic2ebcad806788074b2b4cb244a25395a48df1852
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/232541
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35054}
This defaults the calculation landed in cl 196502. The less readable legacy calculation method will be deleted in a future CL.
Bug: none
Change-Id: Ida02a5208e354835b964c69355ad1e9d5bba18aa
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/231956
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Commit-Queue: Christoffer Rodbro <crodbro@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35027}
Thanks to the elimination of `ExperimentalNs`, there is no need anymore
to pass `webrtc::Config` to build APM.
Hence, `AudioProcessingBuilder::Create(const webrtc::Config&)` is also
removed.
Bug: webrtc:5298
Change-Id: I0a3482376a7753434486fe564681f7b9f83939c5
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/232128
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Alessio Bazzica <alessiob@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35025}
This CL ensures that each DesktopFrame's updated-region is expressed in
the frame's own coordinates, where the top-left is always (0, 0).
For example, DesktopFrame::GetFrameDataAtPos() and its callers use
this coordinate system.
Previously, whenever a RANDR monitor with a non-zero offset was
selected, ScreenCapturerX11 would hit some DCHECKs when trying to
copy pixels from previous frames, or when capturing new pixels into
them from XDAMAGE regions.
Bug: None
Change-Id: I7b2e8d0449359ee7b263ad60af193e2bf89aa1f4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/232085
Reviewed-by: Joe Downing <joedow@chromium.org>
Commit-Queue: Joe Downing <joedow@chromium.org>
Cr-Commit-Position: refs/heads/main@{#35017}
The nack threshold feature is unlikely to provide any value, since
reordered packets are rare. This CL also removes the factory method
from the NackTracker class, since it did not add much value.
Bug: webrtc:10178
Change-Id: Ib6bece4a2d9f95bd4298799aaa15627f5c014b61
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/231953
Reviewed-by: Jakob Ivarsson <jakobi@webrtc.org>
Commit-Queue: Ivo Creusen <ivoc@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34993}
This is tested by a simple unit test and a new fuzzer that verify that all that can be parsed also can be written.
Bug: webrtc:12000
Change-Id: I461aedf97d3dec6e8916e72110fa097c3b31c27f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/231642
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34986}
This is done by not adding missing packets to the NACK list if the number of samples per packet is too large.
Bug: webrtc:10178
Change-Id: If46398d6d05ea35f30d7028040d3b808559e950b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/231841
Reviewed-by: Ivo Creusen <ivoc@webrtc.org>
Commit-Queue: Jakob Ivarsson <jakobi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34984}
The documentation for `OpenPipeWireRemote()` says:
> Open a file descriptor to the PipeWire remote where the camera nodes
> are available. The file descriptor should be used to create a
> pw_core object, by using pw_context_connect_fd.
In `InitPipeWire()` we already successfully requested the FD, but then
went on and used the unrestricted default socket.
This does not matter in non-sandboxed environments, as the stream we
want to use is available from both FDs. In flatpak sandboxes, however,
this requires to give full Pipewire access to the application.
Fix this by simply using the right, restricted FD, and while on it,
also make sure to not leak it.
This change has already landed in downstream in Firefox, see
https://phabricator.services.mozilla.com/D122904https://phabricator.services.mozilla.com/D124508
Bug: webrtc:13152
Change-Id: I3f8995c54c797e1a90a980f231e496a13cbe65b4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/230803
Reviewed-by: Joe Downing <joedow@chromium.org>
Commit-Queue: Joe Downing <joedow@chromium.org>
Cr-Commit-Position: refs/heads/main@{#34983}
The VP9 encoder may drop a frame internally which will not advance the
frame pattern. Consider the following scenario where only spatial layer
0 and temporal layer 0 is active:
1. Key frame encoded
2. Spatial layer 1 is activated
3. Delta T0 dropped
4. Delta T0 encoded
No S1T0 frame is encoded in (1) since it's not active. When
NextFrameConfig is called in (3) it will say that future frames may
reference T0 on both S0 and S1, but it's then dropped.
On step (4), the SVC controller essentially thinks it's encoding a new
picture and will happily reference the T0 on what it thinks is the first
delta frame. However, this is actually still the key frame and since
there was no S1T0 frame produced it will reference an invalid buffer.
To fix this, only say it's possible to reference a T0 frame after it has
been successfully encoded.
Bug: webrtc:11999, webrtc:13142, chromium:1178444
Change-Id: Iab3d2042ce0b3fa7d952b2831d1a36b1a6613a86
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/231695
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Emil Lundmark <lndmrk@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34982}
Unlike libvpx, the VideoBitrateAllocation expects that the bitrate
allocation is separate for each temporal layer. In this instance, if the
bitrates are not separated it will fool the SVC controller into thinking
that all temporal layers are always active.
Bug: webrtc:11999
Change-Id: Ibc33ac00b8b7716c011b94e1ec9c640cedb5274e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/231693
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Emil Lundmark <lndmrk@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34980}
Requesting nacks in more cases allows the delay adaptation to better
predict if it's worth it to increase the jitter buffer delay to wait for
the nacked packets to arrive.
Bug: webrtc:10178
Change-Id: I46adb76c013dbb8df0b99eb3f7a398f8f21c2bf6
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/231648
Commit-Queue: Ivo Creusen <ivoc@webrtc.org>
Reviewed-by: Jakob Ivarsson <jakobi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34970}
This cover scenario where target bitrate is changed in a middle of
of group of frame after spatial upswitch.
This change should avoid wasting encoder resources to produce those
frames, reduce number of errors
"Encoder produced a frame for layer that wasn't requested"
Bug: webrtc:11999
Change-Id: I06045259b1cee2c21bfdabbafff3892b57c82a84
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/230543
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34969}
This is done by adding a reorder optimizer that estimates the probability of receiving reordered packets.
The optimal delay is decided by balancing the cost of increasing the delay against the probability of missing a reordered packet, resulting in a loss. This balance is decided using the `ms_per_loss_percent` parameter.
The usage and parameters can be controlled via field trial.
Bug: webrtc:10178
Change-Id: Ic484df0412af35610e74b3a6070f2bac7a926a63
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/231541
Reviewed-by: Ivo Creusen <ivoc@webrtc.org>
Commit-Queue: Jakob Ivarsson <jakobi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34954}
Setting the rtp header extensions on the packet delivery thread
(currently worker, soon to be network), is now possible without
taking the hit of deleting and recreating the receive stream (and
rtp receiver and related state).
Bug: webrtc:11993
Change-Id: I9bbe306844a25d85d79cd216092ead66eaf68960
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/223741
Reviewed-by: Niels Moller <nisse@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34953}
The HandleXr method has output arguments that are not set when an RTCP
report cannot be parsed. We should give these a sensible default value
to avoid accessing uninitialized memory
Bug: chromium:1247182
Change-Id: I6c54260aef3834643c41b96c0709489522d82533
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/231237
Commit-Queue: Ivo Creusen <ivoc@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34943}
To focus on the ability to predict clipping, the clipping predictor
evaluator doesn't increment the true positive count anymore when a
prediction is simultaneously observed with a detection.
Note that `WebRTC.Audio.Agc.ClippingPredictor.F1Score` is still used
to log the F1 score - i.e., the histogram hasn't been renamed.
Bug: webrtc:12774
Change-Id: Ia987e568a6df2a3ddba7fa1b5697d6feda22d20c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/231233
Commit-Queue: Alessio Bazzica <alessiob@webrtc.org>
Reviewed-by: Hanna Silen <silen@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34942}
