If the encoder takes a long time to start up and emit frames the polling
interval of the quality scaler would get out of sync. This causes it to
sometimes make scaling decisions based on only a handful of frames.
This CL ensures that we have observed some minimum number of frames
before deciding to scale up or down.
BUG=b/36734056
Review-Url: https://codereview.webrtc.org/2789483002
Cr-Commit-Position: refs/heads/master@{#17523}
These tests were disabled due to flakiness when running on the bots.
Hopefully synchronizing all operations that run on Task Queue will
fix this.
BUG=webrtc:6799
Review-Url: https://codereview.webrtc.org/2774643002
Cr-Commit-Position: refs/heads/master@{#17463}
Use AdaptDown/AdaptUp instead of ScaleDown/ScaleUp, since we may want to
adapt using other means than resolution.
Also, extend vie_encoder with unit test that actually uses frames scaled
to resolution as determined by VideoAdapter, since that seems to be the
default implementation.
BUG=webrtc:4172
Review-Url: https://codereview.webrtc.org/2652893015
Cr-Commit-Position: refs/heads/master@{#16402}
These tests have been a little flaky on the bots.
Hopefully increasing the timeout by 200% will help.
BUG=webrtc:6799
Review-Url: https://codereview.webrtc.org/2541743006
Cr-Commit-Position: refs/heads/master@{#15355}
This brings QualityScaler much more in line with OveruseFrameDetector.
The two classes are conceptually similar, and should be used in the
same way. The biggest changes in this CL are:
- Quality scaling is now only done in ViEEncoder and not in each
encoder implementation separately.
- QualityScaler now checks the average QP asynchronously, instead of
having to be polled on each frame.
- QualityScaler is no longer responsible for actually scaling the frames,
but has a callback to ViEEncoder that it uses to express it's desire
for lower resolution.
BUG=webrtc:6495
Review-Url: https://codereview.webrtc.org/2398963003
Cr-Commit-Position: refs/heads/master@{#15286}
The MovingAverage class was very specific to the QualityScaler. This
commit generalizes the MovingAverage class to be useful in other
situations as well, and adapts the QualityScaler to use the new
MovingAverage.
BUG=webrtc:6304
Review-Url: https://codereview.webrtc.org/2310853002
Cr-Commit-Position: refs/heads/master@{#14207}
Don't use VideoFrameBuffer::MutableDataY and friends, instead, use
I420Buffer::SetToBlack.
Also introduce static method I420Buffer::Create, to create an object and
return a scoped_refptr.
TBR=marpan@webrtc.org # Trivial change to video_denoiser.cc
BUG=webrtc:5921
Review-Url: https://codereview.webrtc.org/2078943002
Cr-Commit-Position: refs/heads/master@{#13212}
Introduce a new method I420Buffer::CropAndScale, and a static
convenience helper I420Buffer::CenterCropAndScale. Use them for almost
all scaling needs.
Delete the Scaler class and the cricket::VideoFrame::Stretch* methods.
BUG=webrtc:5682
R=pbos@webrtc.org, perkj@webrtc.org, stefan@webrtc.org
Review URL: https://codereview.webrtc.org/2020593002 .
Cr-Commit-Position: refs/heads/master@{#13110}
Framerate-reduction code is disabled on all platforms, and this code
adds complexity. It's necessary to react fast, especially on mobile
platforms or other bad network conditions and framerate reduction adds
another step between HD and QVGA.
BUG=webrtc:5678, webrtc:5830
R=jackychen@webrtc.org, mflodman@webrtc.org
Review URL: https://codereview.webrtc.org/1885893002 .
Cr-Commit-Position: refs/heads/master@{#12503}
Makes QualityScaler start at QVGA for <250k initial bitrates. Useful in
combination with overriding max bitrates to a max lower than that for
connections where we know that the max bitrate is capped below where VGA
is useful.
BUG=webrtc:5678
R=glaznev@webrtc.org
Review URL: https://codereview.webrtc.org/1900483004 .
Cr-Commit-Position: refs/heads/master@{#12416}
Increases measure time for downscale back to 5 seconds, this is required
to not over-react on hand-waving or quick device rotations.
Also increase max thresholds for QP a bit to not overreact when quality
still looks somewhat OK. Min thresholds for H264 seemed very low and are
increased to be sure that we can go back up again. The window is still
quite big with the increased max QP.
Also changes libvpx thresholds to use the same thresholds as the
encoder, they were excessively low before and wouldn't adapt on bad QPs
at all before (but rely on >60% framedropping based on bitrates to go
down).
BUG=webrtc:5678
R=stefan@webrtc.orgTBR=glaznev@webrtc.org
Review URL: https://codereview.webrtc.org/1894083002 .
Cr-Commit-Position: refs/heads/master@{#12403}
This denominator doesn't make any semantic sense, it's better to use
real thresholds for when things look "good" or "bad" rather than
fractions of a max QP.
BUG=webrtc:5678
R=danilchap@webrtc.org
Review URL: https://codereview.webrtc.org/1855393005 .
Cr-Commit-Position: refs/heads/master@{#12363}
Permits going from HD to QVGA in 6 seconds instead of 10. Also adds
windows for going up quickly in the beginning of a call (before any
downscaling happens due to bad quality).
BUG=webrtc:5678
R=glaznev@webrtc.org, stefan@webrtc.org
Review URL: https://codereview.webrtc.org/1830593003 .
Cr-Commit-Position: refs/heads/master@{#12219}
This makes it clearer this code not meant to be used as an API.
I could not find any use of this in downstream code.
BUG=webrtc:5095
TESTED=git cl try -c --bot=android_compile_rel --bot=linux_compile_rel --bot=win_compile_rel --bot=mac_compile_rel --bot=ios_rel --bot=linux_gn_rel --bot=win_x64_gn_rel --bot=mac_x64_gn_rel --bot=android_gn_rel -m tryserver.webrtc
R=stefan@webrtc.orgTBR=magjed@webrtc.org
Review URL: https://codereview.webrtc.org/1440873005 .
Cr-Commit-Position: refs/heads/master@{#10699}
Android hardware H.264 seems to keep a steady high-QP flow instead of
dropping frames, so framedrops aren't sufficient to detect a bad state
where downscaling would be beneficial.
BUG=webrtc:4968
R=magjed@webrtc.org, stefan@webrtc.org
Review URL: https://codereview.webrtc.org/1364253002 .
Cr-Commit-Position: refs/heads/master@{#10078}
