Behind a flag, the new behavior changes how the "media rate" utilization
is calculated:
* Instead of per spatial & temporal layer, it's per spatial layer only.
* Overshoot is compared to real target vs adjusted target.
* Window takes quite periods/frame drops more into consideration.
This should lead to less push-back when not network constrained and
complex content is used causing bursty behavior.
Bug: b/349561566
Change-Id: I402e6531183493c963fec48ae363ce0b859b396a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/356480
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42593}
Bitrate limits should have been applied to the spatial layers in ApplySpatialLayerBitrateLimits (and usage is restricted to a single active stream/layer).
Bug: b/299588022
Change-Id: Iaae4ece28b8a95eea7d4bacba292847ba5b4000b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/355841
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Åsa Persson <asapersson@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42588}
Convert most field trials used in PCLF tests.
Change-Id: I26c0c4b1164bb0870aae1a488942cde888cb459d
Bug: webrtc:10335
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/322703
Commit-Queue: Jeremy Leconte <jleconte@google.com>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Jonas Oreland <jonaso@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#40909}
Change-Id: I118817fe9fc4d4e674268743ac7d6d2773d366de
Bug: webrtc:15496
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/320260
Reviewed-by: Michael Horowitz <mhoro@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Dan Tan <dwtan@google.com>
Cr-Commit-Position: refs/heads/main@{#40756}
This class is used for both simulcast and SVC use cases. Update variable
names and code comments to reflect this fact.
Also add TODOs that we'll need to address for VP9 simulcast.
Bug: webrtc:14884
Change-Id: I814c8fa0097306b16d552f55ca391ac8f716348a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/294383
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#39367}
At the point where an EncodedImage is reported to the
EncoderBitrateAdjuster (in order to estimate utilization), the image
data has been cleared so the size is 0 - meaning the esimtated
utilization is 0 so pushback is in effect only applied at the
beginning before an estimate is available.
This CL fixes that by explicitly using spatial/temporal id and size in
bytes, rather than passing along the EncodedImage proxy.
It is unclear when this broke, but the regression seems rather old.
This CL will affect the encoded bitrate (and thus indirectly BWE
ramp-up rate), but should avoid exessive delay at low bitrates.
Perf bots will likely trigger alerts, this is expected.
In case there are undesired side-effects, we can entirely disable the
adjuster using existing field-trials.
Bug: webrtc:12606
Change-Id: I936c2045f554696d8b4bb518eee6871ffc12c47d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/212900
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33550}
This CL adds a field trial that enables the EncoderBitrateAdjuster to
allow higher target bitrate if we are not network constrained. We still
don't allow the bitrate to go higher than the average target media rate
though.
Bug: webrtc:10155
Change-Id: Id5995070aa0cbe84b9305a422279141b38664bb1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/132717
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27627}
The overshoot detector uses a simple pacer model to determine an
estimate of how much the encoder is overusing the target bitrate.
This utilization factor can then be adjuster for when configuring the
actual target bitrate.
Spatial layers (simulcast streams) are adjusted separately.
Temporal layers are measured separately, but are combined into a single
utilization factor per spatial layer.
Bug: webrtc:10155
Change-Id: I8ea58dc6c4871e880553d7c22202f11cb2feb216
Reviewed-on: https://webrtc-review.googlesource.com/c/114886
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26573}
