Make it possible to switch from VP8 HW -> VP8 SW -> VP8 HW depending on bitrate and resolution.
BUG=webrtc:6634
Review-Url: https://codereview.webrtc.org/2988963002
Cr-Commit-Position: refs/heads/master@{#19362}
Reason for revert:
Speculative revet for breaking remoting_unittests in fyi bots.
https://build.chromium.org/p/chromium.webrtc.fyi/waterfall?builder=Win7%20Tester
Original issue's description:
> Add a flags field to video timing extension.
>
> The rtp header extension for video timing shuold have an additional
> field for signaling metadata, such as what triggered the extension for
> this particular frame. This will allow separating frames select because
> of outlier sizes from regular frames, for more accurate stats.
>
> This implementation is backwards compatible in that it can read video
> timing extensions without the new flag field, but it always sends with
> it included.
>
> BUG=webrtc:7594
>
> Review-Url: https://codereview.webrtc.org/3000753002
> Cr-Commit-Position: refs/heads/master@{#19353}
> Committed: cf5d485e14TBR=danilchap@webrtc.org,kthelgason@webrtc.org,stefan@webrtc.org,sprang@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:7594
Review-Url: https://codereview.webrtc.org/2995953002
Cr-Commit-Position: refs/heads/master@{#19360}
We're encountering a bug where audioRecord.read() can hang for long
enough that stopRecording() fails to join the recording thread (in two
seconds) and returns. In that case, JNI methods get unregistered and
when the recording thread calls nativeDataIsRecorded, it crashes when
it can't find the native method to call.
This version still isn't 100% safe, as the threading sequence still
technically allows for an ordering where (for some reason) the thread
fails to join after the final keepAlive check and long enough for all
the JNI methods to get unregistered, but that seems very unlikely.
BUG=b/64174142
Change-Id: Ie7432a70d0e53bace0885edf35e24bd3f6585399
Reviewed-on: https://chromium-review.googlesource.com/613501
Reviewed-by: Henrik Andreasson <henrika@webrtc.org>
Commit-Queue: Noah Richards <noahric@chromium.org>
Cr-Commit-Position: refs/heads/master@{#19358}
The rtp header extension for video timing shuold have an additional
field for signaling metadata, such as what triggered the extension for
this particular frame. This will allow separating frames select because
of outlier sizes from regular frames, for more accurate stats.
This implementation is backwards compatible in that it can read video
timing extensions without the new flag field, but it always sends with
it included.
BUG=webrtc:7594
Review-Url: https://codereview.webrtc.org/3000753002
Cr-Commit-Position: refs/heads/master@{#19353}
Given the current state of OpenSLES (disabled in many places), making
this a debug line makes more sense than an error.
BUG=none
Change-Id: I16d46d3f8234ebeffe820d92e7a6d7ed3eae11cd
Reviewed-on: https://chromium-review.googlesource.com/611491
Commit-Queue: Henrik Andreasson <henrika@webrtc.org>
Reviewed-by: Henrik Andreasson <henrika@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#19340}
Reason for revert:
Speculative revert to see if this caused regressions in android perf tests.
Original issue's description:
> Add functionality which limits the number of bytes on the network.
>
> The limit is based on the bandwidth delay product, but also adds some additional slack to compensate for the sawtooth-like BWE pattern and the slowness of the encoder rate control. The delay is estimated based on the time from sending a packet until an ack is received. Since acks are received in bursts (feedback is only sent periodically), a min filter is used to estimate the rtt.
>
> Whenever the in flight bytes reaches the congestion window, the pacer is paused, which in turn will result in send-side queues growing. Eventually the encoders will be paused as the pacer queue grows large (currently 2 seconds).
>
> BUG=webrtc:7926
>
> Review-Url: https://codereview.webrtc.org/2918323002
> Cr-Commit-Position: refs/heads/master@{#19289}
> Committed: 8497fdde43TBR=terelius@webrtc.org,philipel@webrtc.org,tschumim@webrtc.org,gnish@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:7926
Review-Url: https://codereview.webrtc.org/3001653002
Cr-Commit-Position: refs/heads/master@{#19339}
Rvalue reference arguments are generally banned by the style guide.
Bug: None
Change-Id: I4314859623ffcf056f53c42087b59696b5e71690
Reviewed-on: https://chromium-review.googlesource.com/531028
Commit-Queue: Minyue Li <minyue@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Reviewed-by: Michael T <tschumim@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#19338}
Reason for revert:
Speculative revert to see if this caused regressions in android perf tests.
Original issue's description:
> Make the acceptable queue in the cwnd experiment configurable.
>
> BUG=webrtc:7926
>
> Review-Url: https://codereview.webrtc.org/2998753002
> Cr-Commit-Position: refs/heads/master@{#19320}
> Committed: 7c83c56b6dTBR=philipel@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:7926
Review-Url: https://codereview.webrtc.org/2999893002
Cr-Commit-Position: refs/heads/master@{#19337}
This CL completely removes the methods
AudioProcessing::{Start,Stop}DebugDumpRecording. These methods have
been replaced with AudioProcessing::{Attach,Detach}AecDump. Their
implementation was removed in the parent CL
https://chromium-review.googlesource.com/c/589147
Bug: webrtc:7404
Change-Id: Ia3d5314985af9c74f79c94c514ded1f8afc78fb5
Reviewed-on: https://chromium-review.googlesource.com/589152
Commit-Queue: Alex Loiko <aleloi@webrtc.org>
Reviewed-by: Alessio Bazzica <alessiob@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#19334}
AudioProcessingModule has a feature to make a recording of its
configuration, inputs and outputs over a period of time. In the past
CLs, this feature has been rewritten to move file IO away from
real-time audio threads. The interface has changed from
{Start,Stop}DebugDumpRecording to {Attach,Detach}AecDump.
This CL removes the previous implementation of the old interface
StartDebugRecording. The public interface is left to not cause
problems to downstream projects. It will be removed in the dependent
CL https://chromium-review.googlesource.com/c/589152/
With this CL, usage of WEBRTC_AUDIOPROC_DEBUG_DUMP and ~300 LOC of
logging code is removed from AudioProcessingImpl.
Bug: webrtc:7404
Change-Id: I16e7b557774e4bc997e1f5de4f97ed2c31d63879
Reviewed-on: https://chromium-review.googlesource.com/589147
Reviewed-by: Alessio Bazzica <alessiob@webrtc.org>
Commit-Queue: Alex Loiko <aleloi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#19332}
Also remove |key_frame_interval| from argument list, since it is always
set to -1.
BUG=webrtc:6634
Review-Url: https://codereview.webrtc.org/2999643002
Cr-Commit-Position: refs/heads/master@{#19331}
* Don't loop over fps, but do loop over codec implementation type.
* Order codec settings as they are defined in the test.
BUG=webrtc:6634
Review-Url: https://codereview.webrtc.org/3000613002
Cr-Commit-Position: refs/heads/master@{#19330}
Reason for revert:
Create fix CL.
Original issue's description:
> Revert of Request keyframes more frequently on stream start/decoding error. (patchset #1 id:1 of https://codereview.webrtc.org/2993793002/ )
>
> Reason for revert:
> Broke downstream test that was waiting for 5 keyframes to be received within 10 seconds. Maybe the issue is that "stats_callback_->OnCompleteFrame(frame->num_references == 0, ..." was changed to "frame->is_keyframe()"?
>
> Original issue's description:
> > Request keyframes more frequently on stream start/decoding error.
> >
> > In this CL:
> > - Added FrameObject::is_keyframe() convinience function.
> > - Moved logic to request keyframes on decoding error from VideoReceived to
> > VideoReceiveStream.
> > - Added keyframe_required as a parameter to FrameBuffer::NextFrame.
> >
> > BUG=webrtc:8074
> >
> > Review-Url: https://codereview.webrtc.org/2993793002
> > Cr-Commit-Position: refs/heads/master@{#19280}
> > Committed: 26b4804358
>
> TBR=terelius@webrtc.org,stefan@webrtc.org,noahric@chromium.org,philipel@webrtc.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=webrtc:8074
>
> Review-Url: https://codereview.webrtc.org/2994043002
> Cr-Commit-Position: refs/heads/master@{#19295}
> Committed: 77a983185fTBR=terelius@webrtc.org,stefan@webrtc.org,noahric@chromium.org,deadbeef@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
BUG=webrtc:8074
Review-Url: https://codereview.webrtc.org/2996823002
Cr-Commit-Position: refs/heads/master@{#19324}
This change implements GetWindowList() on X11. WindowCapturerLinux and
GetWindowUnderPoint() can share the logic of this function.
Bug: webrtc:7950
Change-Id: Ida746840d6f51d31e0470e5ae4955b6f5a4cfaf2
Reviewed-on: https://chromium-review.googlesource.com/606560
Reviewed-by: Jamie Walch <jamiewalch@chromium.org>
Commit-Queue: Zijie He <zijiehe@chromium.org>
Cr-Commit-Position: refs/heads/master@{#19314}
This CL brings us one step closer to removing CodecDatabase and
GenericEncoder, by removing the static VCM::Codec(). Codec specific
methods are moved to video_encoder.cc (they already belonged to this
class) and getting default generic codec settings has been moved to a
test specific file.
This CL also makes video_encoder.h pass style guide and lint checks,
since these checks are triggered with the new video_encoder.cc file.
BUG=webrtc:8064
Review-Url: https://codereview.webrtc.org/2993923002
Cr-Commit-Position: refs/heads/master@{#19303}
That way, the debug printout will tell us which of x and y that was false.
BUG=none
Review-Url: https://codereview.webrtc.org/2988153003
Cr-Commit-Position: refs/heads/master@{#19297}
Reason for revert:
Broke downstream test that was waiting for 5 keyframes to be received within 10 seconds. Maybe the issue is that "stats_callback_->OnCompleteFrame(frame->num_references == 0, ..." was changed to "frame->is_keyframe()"?
Original issue's description:
> Request keyframes more frequently on stream start/decoding error.
>
> In this CL:
> - Added FrameObject::is_keyframe() convinience function.
> - Moved logic to request keyframes on decoding error from VideoReceived to
> VideoReceiveStream.
> - Added keyframe_required as a parameter to FrameBuffer::NextFrame.
>
> BUG=webrtc:8074
>
> Review-Url: https://codereview.webrtc.org/2993793002
> Cr-Commit-Position: refs/heads/master@{#19280}
> Committed: 26b4804358TBR=terelius@webrtc.org,stefan@webrtc.org,noahric@chromium.org,philipel@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:8074
Review-Url: https://codereview.webrtc.org/2994043002
Cr-Commit-Position: refs/heads/master@{#19295}
When compiling webrtc's call.cc with VC++ 2017 (is_clang = false) the
following compile error occurs:
sequence_number_util.h(90): error C2672: 'rtc::SafeLt': no matching
overloaded function found
note: see reference to class template instantiation
'webrtc::SeqNumUnwrapper<T,M>' being compiled
This error is not associated with any particular instantiation of
SeqNumUnwrapper (there isn't one) and this undefined nature of 'T' seems
to be what confuses the compiler. When it tries to locate SafeLt for an
undefined type 'T' it gets confused.
SafeLt is unnecessary in this context and changing it to use the '<'
operator directly avoids the problem.
The bug has been reported to Microsoft.
BUG=chromium:753488
Review-Url: https://codereview.webrtc.org/2997623002
Cr-Commit-Position: refs/heads/master@{#19292}
The limit is based on the bandwidth delay product, but also adds some additional slack to compensate for the sawtooth-like BWE pattern and the slowness of the encoder rate control. The delay is estimated based on the time from sending a packet until an ack is received. Since acks are received in bursts (feedback is only sent periodically), a min filter is used to estimate the rtt.
Whenever the in flight bytes reaches the congestion window, the pacer is paused, which in turn will result in send-side queues growing. Eventually the encoders will be paused as the pacer queue grows large (currently 2 seconds).
BUG=webrtc:7926
Review-Url: https://codereview.webrtc.org/2918323002
Cr-Commit-Position: refs/heads/master@{#19289}
In this CL:
- Added FrameObject::is_keyframe() convinience function.
- Moved logic to request keyframes on decoding error from VideoReceived to
VideoReceiveStream.
- Added keyframe_required as a parameter to FrameBuffer::NextFrame.
BUG=webrtc:8074
Review-Url: https://codereview.webrtc.org/2993793002
Cr-Commit-Position: refs/heads/master@{#19280}
During window capture on Windows 10, if the selected window is minimized,
ShouldUseScreenCapturer() still thinks it's on top and continue to do a
screencapture which is meaningless.
This cl will set |.is_top_window| with false to minimized window,then we
can skip doing any capture to it.
BUG=chromium:568835
Review-Url: https://codereview.webrtc.org/2997493002
Cr-Commit-Position: refs/heads/master@{#19276}
Found via supersize query:
size_info.symbols.WhereFullNameMatches(r'\bk[A-Z]').WhereInSection('d')
This moves 90 symbols from .data -> .data.rel.ro (5.50kb)
BUG=chromium:747064
Review-Url: https://codereview.webrtc.org/2986163002
Cr-Commit-Position: refs/heads/master@{#19274}
A crash has been randomly detected across different versions. The NSImage
crashes the binary in its lockFocusFlipped() function. The suspicious issue is
that NSCursor::image() returns an invalid NSImage.
BUG=chromium:752036
Review-Url: https://codereview.webrtc.org/2993173003
Cr-Commit-Position: refs/heads/master@{#19273}
Earlier the pid/tl0 was incorrectly reinitialized upon encoder reconfiguration,
and this fix was implemented to mitigate that. This fix can however guess wrong
and cause a valid stream to be interupted.
BUG=webrtc:7920
Review-Url: https://codereview.webrtc.org/2969043002
Cr-Commit-Position: refs/heads/master@{#19268}
WindowUnderPoint() is a platform independent function to return the id of the
first window in z-order under a certain DesktopVector. It equals to
GetAncestor(WindowFromPoint(point), GA_ROOT)
on Windows.
This CL includes the change to Windows / Mac OSX only to control the size in a
reasonable range. Implementation for Linux will be added in a coming change.
Bug: webrtc:7950
Change-Id: I57e423294fc8aeaa12d05cb626a1912240b2d4d0
Reviewed-on: https://chromium-review.googlesource.com/595022
Commit-Queue: Zijie He <zijiehe@chromium.org>
Reviewed-by: Jamie Walch <jamiewalch@chromium.org>
Cr-Commit-Position: refs/heads/master@{#19263}
It serves a very limited purpose: converting from the input YUV
file to an output Y4M file. The experimenter can do this manually,
if this is of interest. (It is generally not.)
BUG=webrtc:6634
Review-Url: https://codereview.webrtc.org/2993063002
Cr-Commit-Position: refs/heads/master@{#19257}
The current ObjC HW encoder is implemented as a C++
webrtc::VideoEncoder. We then wrap it two times in the following way:
webrtc::VideoEncoder -> RTCVideoEncoder -> webrtc::VideoEncoder.
This was originally done to minimize the code diff when landing the
injectable encoder.
This CL removes the first wrapping and implements the ObjC HW encoder
as a RTCVideoEncoder directly. Similarly, the decoder is implemented
as a RTCVideoDecoder directly.
Based on andersc@ CL: https://codereview.webrtc.org/2978623002/.
BUG=webrtc:7924
Review-Url: https://codereview.webrtc.org/2987413002
Cr-Commit-Position: refs/heads/master@{#19255}
- Make all overridden methods of VideoProcessorImpl public,
in preparation of the removal of the VideoProcessor interface.
- Place corresponding method definitions in correct order
in .cc file.
- Harmonize the stdout printing.
- Make timestamp calculations adhere to set frame rate.
Except for the last bullet, these changes should not lead to
different functionality.
BUG=webrtc:6634
Review-Url: https://codereview.webrtc.org/2995513002
Cr-Commit-Position: refs/heads/master@{#19254}
Reason for revert:
Revert to create fix CL.
Original issue's description:
> Revert of Fix off-by-one bugs in video_coding::PacketBuffer when the buffer is filled with a single frame. (patchset #5 id:80001 of https://codereview.chromium.org/2993513002/ )
>
> Reason for revert:
> Break performance bots.
>
> Original issue's description:
> > Fix off-by-one bugs in video_coding::PacketBuffer when the buffer is filled with a single frame.
> >
> > BUG=webrtc:8028
> >
> > Review-Url: https://codereview.webrtc.org/2993513002
> > Cr-Commit-Position: refs/heads/master@{#19209}
> > Committed: ee13e8919c
>
> TBR=stefan@webrtc.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=webrtc:8028
>
> Review-Url: https://codereview.webrtc.org/2990183002
> Cr-Commit-Position: refs/heads/master@{#19211}
> Committed: c18f1d7c94TBR=stefan@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:8028
TBR=stefan@webrtc.org
Review-Url: https://codereview.webrtc.org/2989313003
Cr-Commit-Position: refs/heads/master@{#19249}
