It may happen that if we have simulcast track with, let's say, 2 streams
A and B, we can receive frame X on A and then receive it again on B
when there is a switch from A to B. TO correctly handle it we need to
skip second receive of X. Later we need to add metric which will show
how many frames were in between when X was received twice.
Bug: webrtc:11557
Change-Id: I8c52a78674b62387f520a587f51e209ed7c0b0bc
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/176853
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31488}
Queue with multiple heads is planned to be used in
DefaultVideoQualityAnalyzer to store stream state. Stream state contains
ordered sequence of frame ids that were send for this video stream.
When frame is received by one receiver it should be removed from state
for that receiver and kept for others.
How it is used can be found in this CL:
https://webrtc-review.googlesource.com/c/src/+/176411
Bug: webrtc:11631
Change-Id: Ic7fabf4d77131805a91f08a2ccfffc73c08d3e2d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/176402
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31444}
To support multiple participants video quality analyzer may need to know
peer names in advance to simplify internal structures and metrics
reporting.
Bug: webrtc:11631
Change-Id: I4ffb1554ab7f0e015b8e937b7ffddd55aba9826f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/176364
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Reviewed-by: Andrey Logvin <landrey@webrtc.org>
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31415}
Add peer name to video quality analyzer interface to make it possible to
add multipeer support.
Change-Id: I2570cd4481503c8634bdd91208b3dd2fa1d62029
Bug: webrtc:11631
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/176329
Reviewed-by: Andrey Logvin <landrey@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31395}
Implemented an analogue of the cpu_usage metrics from third_party/webrtc/video/video_analyzer.h for third_party/webrtc/test/pc/e2e/analyzer/video/default_video_quality_analyzer.h
Bug: webrtc:11496
Change-Id: Ifdc9daa3351f1df5db98beb8f7dc7156fc7c2a16
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/174020
Reviewed-by: Artem Titov <titovartem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31141}
Move creation of video sinks for dumping and showing rendered video on
screen into video quality analyzer injection helper to eliminate need
to search for video config in on track callback, which makes this more
reliable and reusable.
Bug: webrtc:11479
Change-Id: I6bb5409688fd39268f9f97bde4c9b0833a64396b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/173820
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31128}
To migrate on new GetStats API and properly support target encode bitrate
for regular, simulcast and svc cases we need to calculate it inside video
quality analyzer getting values from SetRates in VideoEncoder.
Bug: webrtc:11381
Change-Id: Ia37acac764ed3c30f64cdbfda8906d543fa03ae2
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/171501
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Patrik Höglund <phoglund@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30881}
Use real video duration instead of test duration to calculate harmonic
frame rate in DefaultVideoQualityAnalyzer.
Bug: webrtc:11445
Change-Id: Ia5f96b2f87178419ec6ebe2ff5dbcb5a0c03c824
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/171104
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30854}
Extract collection of BWE stats from DefaultVideoQualityAnalyzer to
separate class to prepare for migration on new GetStats API and simplify
quality analyzer.
Bug: webrtc:11381
Change-Id: I0e7e2d7e40b467d7a42633a72a7ffc49ebcb0237
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/169444
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30650}
DefaultVideoQualityAnalyzer accumulates in flight frames in internal
queue to perform psnr/ssim computation. This queue can grow a lot if
test experience high frame loss. As a result of this, the analyzer
can use quite a lot of memory and cause OOM crashes.
This CL limits the size of the queue based on the assumption that after
a certain point a frame can be considered lost and so it is impossible
to calculate PSNR/SSIM.
Bug: webrtc:11373
Change-Id: Iaabcc8d1c3c9142dc58ea5f2f30f599864b088e8
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/168951
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30602}
Force copy video frame including video buffer in
DefaultVideoQualityAnalyzer to ensure that analyzer won't hold any
internal WebRTC buffers.
Bug: webrtc:10138
Change-Id: Ib195233f8b01c855220be1b9743c4f54fc62a22b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/168643
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30535}
The transport time metric was accidentially changed by the CL
https://webrtc-review.googlesource.com/c/src/+/153660
This CL restore the transport time metric to how it has been
measured before, that is, time from encoder output to decoder input.
Bug: webrtc:10975
Change-Id: I66f022f26976451d28c0374b22849f14f9c02378
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/153886
Reviewed-by: Artem Titov <titovartem@webrtc.org>
Commit-Queue: Johannes Kron <kron@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#29259}
Add ability to export internal state of SamplesStatsCounter to be able
then to plot that data.
Bug: webrtc:10138
Change-Id: I5aae5b7dea2989e9f82820933a9ab6f21db17556
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/152542
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Patrik Höglund <phoglund@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#29173}
Refactor it one more time to partly roll back previous change and unify
approach between capturer and renderer. Now we will be able to add single
screen shower listener to display video during the test on the screen.
Bug: webrtc:10138
Change-Id: Ib19117b0943e7c6dfc14630faca1f0e4ee2d038f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/151649
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#29093}
Separate renderer part into steps and make it easier to add more steps
as separate interceptors.
Bug: webrtc:10138
Change-Id: I667fc85d0da4fb59090e69caa4c32bd4afc3bd05
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/151645
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#29082}
Add frames_in_flight metric into PC framework to catch frames that were
captured but weren't delivered to the other side. Right now they won't
be reported as dropped, because it's unclear were they dropped or will
they be delivered. So the new metric is introduced. The lower value is
better for it.
Bug: webrtc:10138
Change-Id: Ide26b362a6b862bd961793cb53293becd92cfaa8
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/148599
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28834}
This reverts commit d978cb43c238ca24b2320acd7b656f446b906101.
Reason for revert: It breaks perf tests: https://ci.chromium.org/p/webrtc/builders/perf/Perf%20Android32%20(L%20Nexus4)/1561
Original change's description:
> Record audio/video bytes sent in analyzer stream stats.
>
> For each SSRC report, record the number of bytes sent for that stream
> and expose them in analyzer stats. These numbers can be used to
> determine useful metrics such as total media throughput (by adding the
> bytes sent for all streams) and overhead (by subtracting that amount
> from the total bytes sent to the network).
>
> Bug: webrtc:9719
> Change-Id: I977bbd40acdd0a1ec64763ddd55a642b9a50f309
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/146240
> Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
> Reviewed-by: Artem Titov <titovartem@webrtc.org>
> Commit-Queue: Bjorn Mellem <mellem@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#28637}
TBR=mbonadei@webrtc.org,mellem@webrtc.org,titovartem@webrtc.org
Change-Id: I3e46307dd6ef121b9377b93fc8d9fa788245ea5f
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: webrtc:9719
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/146605
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28646}
For each SSRC report, record the number of bytes sent for that stream
and expose them in analyzer stats. These numbers can be used to
determine useful metrics such as total media throughput (by adding the
bytes sent for all streams) and overhead (by subtracting that amount
from the total bytes sent to the network).
Bug: webrtc:9719
Change-Id: I977bbd40acdd0a1ec64763ddd55a642b9a50f309
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/146240
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Artem Titov <titovartem@webrtc.org>
Commit-Queue: Bjorn Mellem <mellem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28637}
SVC support is limited:
During SVC testing there is no SFU, so framework will try to emulate SFU
behavior in regular p2p call. Because of it there are such limitations:
* if |target_spatial_index| is not equal to the highest spatial layer
then no packet/frame drops are allowed.
If there will be any drops, that will affect requested layer, then
WebRTC SVC implementation will continue decoding only the highest
available layer and won't restore lower layers, so analyzer won't
receive required data which will cause wrong results or test failures.
Bug: webrtc:10138
Change-Id: I079566260ca9f1815935bce365d1bca10766663a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144882
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28612}
