Follow-up CL to VP8 and VP9 encoders taking care of mapping.
Context again:
This CL is part of Optimized Scaling efforts. In Chromium, the native
frame buffer is getting an optimized CropAndScale() implementation. To
support HW accelerated scaling, returning pre-scaled images and skipping
unnecessary intermediate downscales, WebRTC needs to 1) use CropAndScale
instead of libyuv::XXXXScale and 2) only map buffers it actually intends
to encode.
In this CL, VideoStreamEncoder no longer calls GetMappedFrameBuffer() on
behalf of the encoders, since the encoders are now able to either do the
mapping or performs ToI420() anyway.
- Tests for old VSE behaviors are updated to test the new behavior (i.e.
that native frames are pretty much always forwarded).
- The "having to call ToI420() twice" workaround to Android bug
https://crbug.com/webrtc/12602 is added to H264 and AV1 encoders.
Bug: webrtc:12469
Change-Id: Ibdc2e138d4782a140f433c8330950e61b9829f43
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/211940
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Evan Shrubsole <eshr@google.com>
Cr-Commit-Position: refs/heads/master@{#33548}
`RTCInboundRtpStreamStats.lastPacketReceivedTimestamp` must be a time
value in milliseconds with Unix epoch as time origin (see
bugs.webrtc.org/12605#c4).
This change fixes both audio and video `RTCInboundRtpStreamStats` stats.
Tested: verified from chrome://webrtc-internals during an appr.tc call
Bug: webrtc:12605
Change-Id: I68157fcf01a5933f3d4e5d3918b4a9d3fbd64f16
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/212865
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Alessio Bazzica <alessiob@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33547}
This patch adds support for sending zero video layer allocations
header extensions. This can be used to signal that a stream is
turned off.
Bug: webrtc:12000
Change-Id: Id18fbbff2216ca23179c58ef7bbe2ebea5e242af
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/212743
Commit-Queue: Jonas Oreland <jonaso@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33541}
This is a follow-up to the VP9, fixing VP8 this time. Context again:
This CL is part of Optimized Scaling efforts. In Chromium, the native
frame buffer is getting an optimized CropAndScale() implementation. To
support HW accelerated scaling, returning pre-scaled images and skipping
unnecessary intermediate downscales, WebRTC needs to 1) use CropAndScale
instead of libyuv::XXXXScale and 2) only map buffers it actually intends
to encode.
- To achieve this, WebRTC encoders are updated to map kNative video
buffers so that in a follow-up CL VideoStreamEncoder can stop mapping
intermediate buffer sizes.
Bug: webrtc:12469, chromium:1157072
Change-Id: I026527ae77e36f66d02e149ad6fe304f6a8ccb05
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/212600
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Evan Shrubsole <eshr@google.com>
Cr-Commit-Position: refs/heads/master@{#33537}
Namespace used because of copy-pasting an old pattern, should never have been used in the first place. Removing it now to make followup refactoring prettier.
Bug: webrtc:12579
Change-Id: I00a80958401cfa368769dc0a1d8bbdd76aaa4ef5
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/212603
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Philip Eliasson <philipel@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33536}
WgcCaptureSession would crash when copying the frame data for an image
from a portrait oriented monitor. This is because we were using the
height of the image multiplied by the rowpitch of the buffer to
determine the size of the data to be copied. However, in portrait
mode the height measures the same dimension as the rowpitch, leading
to us overrunning the frame buffer.
The fix is to use the height and width of the image multiplied by
the number of bytes per pixel to determine how much data to copy
out of the buffer, and only use the rowpitch to advance the pointer
in the source data buffer. This has the added benefit of giving us
contiguous data, reducing the size of the DesktopFrame that we output.
Bug: webrtc:12490
Change-Id: I4c26f8864cb57ac566a742af70fea1da504b9706
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/209501
Reviewed-by: Joe Downing <joedow@chromium.org>
Commit-Queue: Austin Orion <auorion@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#33532}
This CL is part of Optimized Scaling efforts. In Chromium, the native
frame buffer is getting an optimized CropAndScale() implementation. To
support HW accelerated scaling, returning pre-scaled images and skipping
unnecessary intermediate downscales, WebRTC needs to 1) use CropAndScale
instead of libyuv::XXXXScale and 2) only map buffers it actually intends
to encode.
- To achieve this, WebRTC encoders are updated to map kNative video
buffers so that in a follow-up CL VideoStreamEncoder can stop mapping
intermediate buffer sizes.
In this CL LibvpxVp9Encoder is updated to map kNative buffers of pixel
formats it supports and convert ToI420() if the kNative buffer is
something else. A fake native buffer that keeps track of which
resolutions were mapped, MappableNativeBuffer, is added.
Because VP9 is currently an SVC encoder and not a simulcast encoder, it
does not need to invoke CropAndScale.
This CL also fixes MultiplexEncoderAdapter, but because it simply
forwards frames it only cares about the pixel format when
|supports_augmented_data_| is true so this is the only time we map it.
Because this encoder is not used with kNative in practise, we don't care
to make this path optimal.
Bug: webrtc:12469, chromium:1157072
Change-Id: I74edf85b18eccd0d250776bbade7a6444478efce
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/212580
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Evan Shrubsole <eshr@google.com>
Cr-Commit-Position: refs/heads/master@{#33526}
This is a reland of aa6adffba325f4b698a1e94aeab020bfdc47adec
What was changed in the reland is that the merging of the bands is
excluded from the code that is not run when the output is not used.
I.e., the merging is always done.
This is important to have since some clients may apply muting before APM,
and still flag to APM that the signal is muted. If the merging is not
always done, those clients will get nonzero output from APM during muting.
Original change's description:
> Reduce complexity in the APM pipeline when the output is not used
>
> This CL selectively turns off parts of the audio processing when
> the output of APM is not used. The parts turned off are such that
> don't need to continuously need to be trained, but rather can be
> temporarily deactivated.
>
> The purpose of this CL is to allow CPU to be reduced when the
> client is muted.
>
> The CL will be follow by additional CLs, adding similar functionality
> in the echo canceller and the noiser suppressor
>
> Bug: b/177830919
> Change-Id: I72d24505197a53872562c0955f3e7b670c43df6b
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/209703
> Commit-Queue: Per Åhgren <peah@webrtc.org>
> Reviewed-by: Sam Zackrisson <saza@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#33431}
Bug: b/177830919
Change-Id: Ib74dd1cefa173d45101e26c4f2b931860abc6d08
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/211760
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33478}
This CL adds functionality in the noise suppressor that allows the
computational complexity to be reduced when the output of APM is not used.
Bug: b/177830919
Change-Id: I849351ba9559fae770e4667d78e38abde5230eed
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/211342
Reviewed-by: Gustaf Ullberg <gustaf@webrtc.org>
Commit-Queue: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33477}
This CL adds functionality in AEC3 that allows the computational
complexity to be reduced when the output of APM is not used.
Bug: b/177830919
Change-Id: I08121364bf966f34311f54ffa5affbfd8b4db1e2
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/211341
Reviewed-by: Gustaf Ullberg <gustaf@webrtc.org>
Commit-Queue: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33476}
This will speed up key frame encoding (together with libaom changes)
3x-4x times with ~13% BDRate loss on key frames only
Bug: None
Change-Id: I24332f4f7285811cdc6619ba29844fe564cae95e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/212040
Reviewed-by: Marco Paniconi <marpan@webrtc.org>
Commit-Queue: Fyodor Kyslov <kyslov@google.com>
Cr-Commit-Position: refs/heads/master@{#33468}
This CL adds functionality that allows adjusting the audio levels
internally in APM. The main purpose of the functionality is to allow
APM to optionally be moved to an integration that does not provide an
analog gain to control, and the implementation of this has been
tailored specifically to meet the requirements for that.
More specifically, this CL does
-Add a new variant of the pre-amplifier gain that is intended to replace
the pre-amplifier gain (but at the moment can coexist with that). The
main differences with the pre-amplifier gain is that an attenuating
gain is allowed, the gain is applied jointly with any emulated analog
gain, and that its packaging fits better with the post gain.
-Add an emulation of an analog microphone gain. The emulation is
designed to match the analog mic gain functionality in Chrome OS (which
is digital) but should be usable also on other platforms.
-Add a post-gain which is applied after all processing has been applied.
The purpose of this gain is for it to work well with the integration
in ChromeOS, and be used to compensate for the offset that there is
applied on some USB audio devices.
Bug: b/177830918
Change-Id: I0f312996e4088c9bd242a713a703eaaeb17f188a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/209707
Commit-Queue: Per Åhgren <peah@webrtc.org>
Reviewed-by: Gustaf Ullberg <gustaf@webrtc.org>
Reviewed-by: Alessio Bazzica <alessiob@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33466}
This CL adds one frame (10 ms) of silence in APM output after unmuting to mask
audio resulting from the turning on the processing that was deactivated
during the muting.
Bug: b/177830919
Change-Id: If44cfb0ef270dde839dcd3f0b98d1c91e81668dd
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/211343
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33454}
Add missing members needed to surface `RTCRemoteOutboundRtpStreamStats`
via `ChannelReceive::GetRTCPStatistics()` - i.e., audio streams.
`GetSenderReportStats()` is added to both `ModuleRtpRtcpImpl` and
`ModuleRtpRtcpImpl2` and used by `ChannelReceive::GetRTCPStatistics()`.
Bug: webrtc:12529
Change-Id: Ia8f5dfe2e4cfc43e3ddd28f2f1149f5c00f9269d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/211041
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Reviewed-by: Gustaf Ullberg <gustaf@webrtc.org>
Commit-Queue: Alessio Bazzica <alessiob@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33452}
A refactoring (https://webrtc-review.googlesource.com/c/src/+/196520)
of decoder metadata handling introduced a bug which causes us to log an
info-level entry for every frame decoded if the implementation changes
during runtime (e.g. due to software fallback).
This CL fixes that to avoid spamming the logs.
Bug: webrtc:12271
Change-Id: I89016351b8752b259299c4cf56c6feddcca43460
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/211664
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33451}
This reverts commit aa6adffba325f4b698a1e94aeab020bfdc47adec.
Reason for revert: breaks webrtc-importer
Original change's description:
> Reduce complexity in the APM pipeline when the output is not used
>
> This CL selectively turns off parts of the audio processing when
> the output of APM is not used. The parts turned off are such that
> don't need to continuously need to be trained, but rather can be
> temporarily deactivated.
>
> The purpose of this CL is to allow CPU to be reduced when the
> client is muted.
>
> The CL will be follow by additional CLs, adding similar functionality
> in the echo canceller and the noiser suppressor
>
> Bug: b/177830919
> Change-Id: I72d24505197a53872562c0955f3e7b670c43df6b
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/209703
> Commit-Queue: Per Åhgren <peah@webrtc.org>
> Reviewed-by: Sam Zackrisson <saza@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#33431}
Bug: b/177830919
Change-Id: I937cd61dedcd43150933eb1b9d65aebe68401e91
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/211348
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33433}
This CL selectively turns off parts of the audio processing when
the output of APM is not used. The parts turned off are such that
don't need to continuously need to be trained, but rather can be
temporarily deactivated.
The purpose of this CL is to allow CPU to be reduced when the
client is muted.
The CL will be follow by additional CLs, adding similar functionality
in the echo canceller and the noiser suppressor
Bug: b/177830919
Change-Id: I72d24505197a53872562c0955f3e7b670c43df6b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/209703
Commit-Queue: Per Åhgren <peah@webrtc.org>
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33431}
`LastReceivedNTP()` does not need to be part of the public members of
`ModuleRtpRtcpImpl` and `ModuleRtpRtcpImpl2` since it is used only
once in the same class.
This change is requried by the child CL [1] which adds a public getter
needed to add remote-outbound stats.
[1] https://webrtc-review.googlesource.com/c/src/+/211041
Bug: webrtc:12529
Change-Id: I82cfea5ee795de37fffa3d759ce9f581ca775d55
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/211043
Commit-Queue: Alessio Bazzica <alessiob@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33420}
The mutex is removed from the old existing implementation and instead a wrapper is implemented that ensure thread-safety.
Both the thread-safe and unsafe version share the same implementation of the logic.
There are two ways of construction:
webrtc::ReceiveStatistics::Create - thread-safe version.
webrtc::ReceiveStatistics::CreateUnLocked -thread-unsafe
Bug: none
Change-Id: Ica375919fda70180335c8f9ea666497811daf866
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/211240
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33419}
The access to |_timestampMap| was guarded by a lock but
not the access to the data pointer stored in |_timestampMap|.
There was a potential race condition if new data was added
in VCMGenericDecoder::Decode() while the data pointer
retrieved from _timestampMap.Pop() was being used in
VCMDecodedFrameCallback::Decoded().
This CL moves the storage of data to within |_timestampMap|,
instead of being a pointer so that it's guarded by the same
lock.
Bug: webrtc:11229
Change-Id: I3f2afb568ed724db5719d508a73de402c4531dec
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/209361
Commit-Queue: Johannes Kron <kron@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33374}
Check if codec was successfully created and exit from RunTest if not
before creating VideoProcessor.
Bug: none
Change-Id: Ia6d7171650dbc9824fb78f4a8e2851f755cfd63b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/209362
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33372}
According to our previous data from trace_event with using direct memcpy
and libyuv::CopyPlane on chromebook atlas, the average cpu duration is
0.624ms and 0.541ms, so using libyuv::CopyPlane is 13.3% faster than
direct memcpy.
Bug: webrtc:12496
Change-Id: I1c41424b402a7eec34052c67933f2e88eaf0a8f4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/196485
Reviewed-by: Tommi <tommi@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33344}
The low-latency renderer is activated by the RTP header extension
playout-delay if the min value is set to 0 and the max value is
set to something greater than 0.
According to the specification of the playout-delay header
extension it doesn't have to be set for every frame but only if
it is changed. The bug that this CL fixes occured if a playout
delay had been set previously but some frames without any specified
playout-delay were received. In this case max composition delay
would not be set and the low-latency renderer algorithm would be
disabled for the rest of the session.
Bug: chromium:1138888
Change-Id: I12d10715fd5ec29f6ee78296ddfe975d7edab8a9
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/208581
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Johannes Kron <kron@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33330}
