This metric was always supposed to be the spec's answer to
googBucketDelay, and is defined as "The total number of seconds that
packets have spent buffered locally before being transmitted onto the
network." But our implementation measured the time between capture and
send, including encode time. This is incorrect and yields a much larger
value than expected.
This CL updated the metric to do what the spec says. Implementation-wise
we measure the time between pushing and popping each packet from the
queue (in modules/pacing/prioritized_packet_queue.cc).
The spec says to increment the delay counter at the same time as we
increment the packet counter in order for the app to be able to do
"delta totalPacketSendDelay / delta packetSent". For this reason,
`total_packet_delay` is added to RtpPacketCounter. (Previously, the
two counters were incremented on different threads and observers.)
Running Google Meet on a good network, I could observe a 2-3 ms average
send delay per packet with this implementation compared to 20-30 ms
with the old implementation. See b/137014977#comment170 for comparison
with googBucketDelay which is a little bit different by design -
totalPacketSendDelay is clearly better than googBucketDelay.
Since none of this depend on the media kind, we can wire up this metric
for audio as well in a follow-up:
https://webrtc-review.googlesource.com/c/src/+/280523
Bug: webrtc:14593
Change-Id: If8fcd82fee74030d0923ee5df2c2aea2264600d4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/280443
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#38480}
Add implementation of RTC_DCHECK_NOTREACHED equal to the RTC_NOTREACHED.
The new macros will replace the old one when old one's usage will be
removed. The idea of the renaming to provide a clear signal that this
is debug build only macros and will be stripped in the production build.
Bug: webrtc:9065
Change-Id: I4c35d8b03e74a4b3fd1ae75dba2f9c05643101db
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/237802
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35348}
Currently encode_frame_rate is updated (ComputeRate called) when a frame is encoded.
If a stream is stopped, encode_frame_rate will have an old value (the framerate at the time of the last encoded frame) instead of zero.
Bug: webrtc:13037
Change-Id: I1a2122df61e3e8187e57155dda71c0173cda4c5b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/228220
Commit-Queue: Åsa Persson <asapersson@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34695}
Update framerate for top spatial layer instead of per timestamp (to ensure all simulcast layers are updated).
Bug: webrtc:13037
Change-Id: I4fa423dee40d74aee22a87855207b885f0536e25
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/227344
Commit-Queue: Åsa Persson <asapersson@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34634}
This reverts commit 48a4d33719390b7bcaf8445a1581a00825f67bfb.
Reason for reland:
Relanding the original change but without the modification for
VideoSendStream::GetStats. Essentially there's a TODO there to fix
the downstream issue, which seems to be benign.
Original change's description:
> Revert "Remove Invoke from VideoChannel::FillBitrateInfo."
>
> This reverts commit 1a1795768e1bdb65054ebe15aa238c6edc78dd14.
>
> Reason for revert: Speculative revert (breaks downstream project).
>
> Original change's description:
> > Remove Invoke from VideoChannel::FillBitrateInfo.
> >
> > The method is relied upon by StatsCollector where it was called from the
> > signaling thread in a loop. Now there's at most one invoke (not N).
> >
> > Uncommenting thread checks and removing TODOs in SendStatisticsProxy,
> > VideoSendStream. Updating all related tests that fetched stats from
> > the wrong context.
> >
> > Bug: webrtc:12726
> > Change-Id: Ia7db1afd7e103ec4f9816f5647203c4e2495586e
> > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/216688
> > Commit-Queue: Tommi <tommi@webrtc.org>
> > Reviewed-by: Niels Moller <nisse@webrtc.org>
> > Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
> > Cr-Commit-Position: refs/heads/master@{#33894}
>
> TBR=ilnik@webrtc.org,nisse@webrtc.org,tommi@webrtc.org,webrtc-scoped@luci-project-accounts.iam.gserviceaccount.com
>
> Change-Id: I2520957cdb33492d187f04320c7416788fd0f820
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Bug: webrtc:12726
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217240
> Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
> Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#33898}
# Not skipping CQ checks because this is a reland.
Bug: webrtc:12726
Change-Id: I41cce3b11a29905cde982c22e82b9b1f5a98e654
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217222
Reviewed-by: Tommi <tommi@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33902}
This reverts commit 1a1795768e1bdb65054ebe15aa238c6edc78dd14.
Reason for revert: Speculative revert (breaks downstream project).
Original change's description:
> Remove Invoke from VideoChannel::FillBitrateInfo.
>
> The method is relied upon by StatsCollector where it was called from the
> signaling thread in a loop. Now there's at most one invoke (not N).
>
> Uncommenting thread checks and removing TODOs in SendStatisticsProxy,
> VideoSendStream. Updating all related tests that fetched stats from
> the wrong context.
>
> Bug: webrtc:12726
> Change-Id: Ia7db1afd7e103ec4f9816f5647203c4e2495586e
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/216688
> Commit-Queue: Tommi <tommi@webrtc.org>
> Reviewed-by: Niels Moller <nisse@webrtc.org>
> Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#33894}
TBR=ilnik@webrtc.org,nisse@webrtc.org,tommi@webrtc.org,webrtc-scoped@luci-project-accounts.iam.gserviceaccount.com
Change-Id: I2520957cdb33492d187f04320c7416788fd0f820
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: webrtc:12726
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217240
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33898}
The method is relied upon by StatsCollector where it was called from the
signaling thread in a loop. Now there's at most one invoke (not N).
Uncommenting thread checks and removing TODOs in SendStatisticsProxy,
VideoSendStream. Updating all related tests that fetched stats from
the wrong context.
Bug: webrtc:12726
Change-Id: Ia7db1afd7e103ec4f9816f5647203c4e2495586e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/216688
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33894}
In particular move end_of_picture flag out of vp9 specific information
since VP9 is not the only codec that can use spatial scalability and
thus need to distinguish layer frame and picture (aka temporal unit).
Bug: webrtc:12167
Change-Id: I0d046d8785fbea55281209ad099738c03ea7db96
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/192542
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#32588}
A more detailed explaination is in the bug, but this changes
the way that adaptation happens when multiple resources are
limited. Only the one that is most limited can trigger an
adaptation up. If multiple resources are most limited both
need to underuse to adapt up.
Some of the changes in this patch to make it all work:
* VideoStreamEncoder unittests that did not reflect this
new behaviour have been changed.
* PeekNextRestrictions returns the adaptation counters as
well as the restrictions.
* Adaptation statstics have changed so that when adapting
up all resources are tagged as triggering the adaptation.
Additionally the statistics for the current adaptation is
now the total number of adaptations per reason, rather then
the number of adaptations due to that reason.
* PreventAdaptUpDueToActiveCounts is removed as most limited
resource is a strong implementation of that.
Bug: webrtc:11553
Change-Id: If1545a201c8e019598edf82657a1befde8b05268
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/176128
Commit-Queue: Evan Shrubsole <eshr@google.com>
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31497}
Replaces this with 2 methods instead, adding clarity.
ClearAdaptationStats
- Resets the adaptations statistics to 0. This is done,
when the degredation is reset, for example when the preference
is changed to/from BALANCED.
UpdateAdaptationMaskingSettings
- Updates the settings for adaptation statistics reporting.
This way we don't report quality adaptations if quality scaling
is not enabled (same for resolution/fps scaling).
The adaptation counting inside the SendStatisticsProxy is
now done in a struct that counts the totals, and then masks
out these counts based on the adaptation settings. The
MaskedAdaptationSteps uses optionals to hide the values we
shoudn't report, while the AdaptationSteps always hold the real
totals.
All tests have been updated to use the Reset/Clear method as needed.
Now that AdaptationCounters and AdaptSteps use the same structure,
AdaptationCounters was moved to api/video and replaces AdaptSteps.
The AdaptReason enum is also redundant now, and will be removed
in a follow-up CL.
R=hbos@webrtc.org
Bug: webrtc:11392
Change-Id: Iaed6488581325d341a056b5bbf76a01c19d6c282
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/171685
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Evan Shrubsole <eshr@google.com>
Cr-Commit-Position: refs/heads/master@{#31083}
--- Background ---
The webrtc::VideoSendStream::StreamStats are converted into
VideoSenderInfo objects which turn into "outbound-rtp" stats objects in
getStats() (or "ssrc" objects in legacy getStats()).
StreamStats are created for each type of substream: RTP media streams,
RTX streams and FlexFEC streams - each with individual packet counters.
The RTX stream is responsible for retransmissions of a referenced media
stream and the FlexFEC stream is responsible for FEC of a referenced
media stream. RTX/FEC streams do not show up as separate objects in
getStats(). Only the media streams become "outbound-rtp" objects, but
their packet and byte counters have to include the RTX and FEC counters.
--- Overview of this CL ---
This CL adds MergeInfoAboutOutboundRtpSubstreams(). It takes
StreamStats of all kinds as input, and outputs media-only StreamStats
- incorporating the RTX and FEC counters into the relevant media
StreamStats.
The merged StreamStats objects is a smaller set of objects than the
non-merged counterparts, but when aggregating all packet counters
together we end up with exact same packet and count as before.
Because WebRtcVideoSendStream::GetVideoSenderInfo() currently aggregates
the StreamStats into a single VideoSenderInfo (single "outbound-rtp"),
this CL should not have any observable side-effects. Prior to this CL:
aggregate StreamStats. After this CL: merge StreamStats and then
aggregate them.
However, when simulcast stats are implemented (WIP CL:
https://webrtc-review.googlesource.com/c/src/+/168120) each RTP media
stream should turn into an individual "outbound-rtp" object. We will
then no longer aggregate all StreamStats into a single "info". This CL
unblocks simulcast stats by providing StreamStats objects that could be
turned into individual VideoSenderInfos.
--- The Changes ---
1. Methods added to RtpConfig to be able to easily tell the relationship
between RTP, RTX and FEC ssrcs.
2. StreamStats gets a StreamType (kMedia, kRtx or kFlexfec) that
replaces the booleans (is_rtx, is_flexfec).
3. "referenced_media_ssrc" is added to StreamStats, making it possible
to tell which kRtx/kFlexFec stream stats need to be merged with which
kMedia StreamStats.
4. MergeInfoAboutOutboundRtpSubstreams() added and used.
Bug: webrtc:11439
Change-Id: Iaf9002041169a054ddfd32c7ea06bd1dc36c6bca
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/170826
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30869}
With current congestion window pushback, when congestion window is filling up, it will reduce bitrate directly and encoder may reduce encode quality, resolution, or framerate to adapt to the allocated bitrate, the behavior is depending on the degradation preference.
This change enable congestion window to only drop frames to reduce bitrate (when needed) instead of reduce general bitrate allocation.
Bug: webrtc:11334
Change-Id: I9cf5c20a0858c4d07d006942abe72aa5e1f7cb38
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/168059
Commit-Queue: Ying Wang <yinwa@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30483}
Removed old code, which took care of VP8 case, but actually interferes with
VP9 case.
Now regardless of codec bw_limited_resolution is calculated based on signals
from the bitrate allocator.
Bug: webrtc:11015
Change-Id: Ic99dbb504ab2d1a4b5f15ca93193a1af05ae5924
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/160651
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#29937}
This is a standardized metric:
https://w3c.github.io/webrtc-stats/#dom-rtcoutboundrtpstreamstats-totalencodedbytestarget
We estimate the target frame size in bytes from the current encoder
target bitrate and encoder framerate.
We would expect that the average bytes produced by the encoder would
over time match the average target, which is calculated by polling
getStats() twice and dividing the delta totalEncodedBytesTarget with
the delta framesEncoded. This is meant to make googTargetEncBitrate
obsolete.
Bug: webrtc:10446
Change-Id: Ib10ce236476a2f965582d5c536f419952926d4e6
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/137200
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28022}
Use the C++-style stdlib headers, add `std::` prefix, in order to avoid implicit casts to double.
Bug: None
Change-Id: I78d9caaee715be341d2480c6d5e769068966d577
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/133625
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Commit-Queue: Oleh Prypin <oprypin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27905}
Use size() accessor function. Also replace most nearby uses of _buffer
with data().
Bug: webrtc:9378
Change-Id: I1ac3459612f7c6151bd057d05448da1c4e1c6e3d
Reviewed-on: https://webrtc-review.googlesource.com/c/116783
Commit-Queue: Niels Moller <nisse@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26273}
This deprecates the following methods in VideoEncoder:
virtual ScalingSettings GetScalingSettings() const;
virtual bool SupportsNativeHandle() const;
virtual const char* ImplementationName() const;
Though they are not marked RTC_DEPRECATED since we still want to call
them from within the default GetEncoderInfo() until downstream
projects have been updated.
Furthmore, implementation name is changed from const char* to
std:string, which prevents some lifetime issues with dynamic encoder
names, and CodecSpecificInfo.codec_name is removed in favor of getting
the implementation name via GetEncoderInfo().
This CL removes calls to these deprecated methods, follow-ups will also
remove implementations of the methods and replace them with new
GetEncoderInfo() substitutions.
Bug: webrtc:9890
Change-Id: I6fd6e531480c0b952f53dbd5105e0b0adc3e3b0c
Reviewed-on: https://webrtc-review.googlesource.com/c/106905
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#25351}
