Code is not used in production. Need this change for a local test.
Using TBR.
TBR: henrik.lundin
Bug: webrtc:9265
Change-Id: I9f0cb265a51507de59ef2d7fd151465133687525
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145330
Commit-Queue: Henrik Andreassson <henrika@webrtc.org>
Reviewed-by: Henrik Andreassson <henrika@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28557}
This partially reverts these 2 CLs:
1) Reland "Copy video frames metadata between encoded and plain frames in one place"
https://webrtc.googlesource.com/src/+/2ebf5239782bf6b46d4aa812f34fa9f9e5a02be9
2) Don't copy video frame metadata in each encoder/decoder
https://webrtc.googlesource.com/src/+/ab62b2ee51e622be6d0aade15e87e927fa60e6f2
The problem with them were that ColorSpace was made to always be copied from the
EncodedImage in the GenericDecoder, which overwrote ColorSpace information from
the decoder.
If decoder applied color space transition or bitstream color space information
was different from the WebRTC signaled one, the incorrect color space data were
passed to the renderer.
This CL removes introduced change regarding color space data: GenericDecoder
doesn't copy or store it and software decoders are restored to copy it.
Relevant tests are also removed.
Bug: chromium:982486
Change-Id: I989e01476ff7f7df376c05578ab8f540b95a1dd2
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145323
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28556}
This reverts commit 6f420e424885dab1d9f885365ea9abea5cc4a901.
Reason for revert: Speculative revert (some perf test are failing)
Original change's description:
> Reland "Add ability to set RTCP sender ssrc at construction time"
>
> This is a reland of 94c58fd815f0c7c6429aa53a79621ea9ef39c770
>
> Patch set 1 is the original CL.
> Patch set 2 introduced a trivial fix. In RtcpSender::SetSSRC(), check
> if either current SSRC is 0 or if the SSRC is identical to the current
> one. If so, don't schedule an early report.
> This prevents a regression in which audio jitter became too low(?)
>
> Original change's description:
> > Add ability to set RTCP sender ssrc at construction time
> >
> > Bug: webrtc:10774
> > Change-Id: Iaf5857e24359e9795434bcd0cdbe1658a2f9f5d3
> > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144632
> > Reviewed-by: Åsa Persson <asapersson@webrtc.org>
> > Commit-Queue: Erik Språng <sprang@webrtc.org>
> > Cr-Commit-Position: refs/heads/master@{#28506}
>
> Bug: webrtc:10774
> Change-Id: I103dfa48719aa41d6ab633cdac8b3a5c46b54843
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144565
> Commit-Queue: Erik Språng <sprang@webrtc.org>
> Reviewed-by: Åsa Persson <asapersson@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#28520}
TBR=asapersson@webrtc.org,sprang@webrtc.org,ilnik@webrtc.org
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: webrtc:10774
Change-Id: I39238d942b2bbe0a9c8ca752387a35ed9dd70650
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145327
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28555}
The GetImplementations function is similar to the GetSupportedFormats function, but instead of providing one SdpVideoFormat per codec it provides one per codec implementation. These SdpVideoFormats can then be tagged so that a certain implementation can be instantiated when CreateVideoEncoder is called.
Bug: webrtc:10795
Change-Id: I79f2380aa03d75d5f9f36138625abf3543c2339d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145215
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Philip Eliasson <philipel@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28553}
This reverts commit 6f129b3b7605dc69c8c188ca02d133250130570e.
Reason for revert: Speculative revert (some perf test are failing)
Original change's description:
> Optimize PacketRouter/RTPSender interactions.
>
> The legacy code-path uses a hashmap as cache in order to speed up
> finding the right rtp module to send on. The new path should use that
> as well.
> In addition, there are checks that verify if an RTP module can send
> padding, in some cases payload based. These result in a number of
> calls to methods in RTPSender requiring its lock to be taken. This CL
> introduces a combined SupportsPadding() check method which performs
> all those checks in one go.
>
> Bug: None
> Change-Id: I2d18d0d6e7d8cfe92c81d08cef248a4daa7dcd4b
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144780
> Reviewed-by: Åsa Persson <asapersson@webrtc.org>
> Reviewed-by: Sebastian Jansson <srte@webrtc.org>
> Commit-Queue: Erik Språng <sprang@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#28535}
TBR=asapersson@webrtc.org,sprang@webrtc.org,srte@webrtc.org
Change-Id: I8499dc0fd6e6d0b9fa7a0886c8754655e5589780
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: None
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145326
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28552}
This reverts commit bb7727211c535f8a9dce27891941b52b6ea8e750.
Reason for revert: Speculative revert (some perf test are failing)
Original change's description:
> Make new pacer padding more like old one
>
> The (currently unused) new pacer code path was implemented with what
> was intended as a more careful padding strategy.
> Unfortunately this doesn't work as well as expected due to the fact
> that the padding budget cannot build up underuse.
>
> I made another CL that could fix that, but I think it adds complexity
> for dubious gains. It also will make it more difficult to find any
> potential regression when switching to the new path, should one occur.
> See https://webrtc-review.googlesource.com/c/src/+/144563
>
> Therefore, this CL makes the new code path choose RTX payload in the
> same way as is currently done.
>
> Bug: webrtc:10633
> Change-Id: If2115d4fa7463add959faa77c63101286c27e0f5
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145202
> Reviewed-by: Stefan Holmer <stefan@webrtc.org>
> Commit-Queue: Erik Språng <sprang@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#28537}
TBR=sprang@webrtc.org,stefan@webrtc.org
Change-Id: I99b72858414e0a245da596d94694449da88fd626
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: webrtc:10633
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145324
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28550}
default connection address is a hostname candidate.
Using a FQDN in the c= line has caused an inter-op issue with Firefox
when hostname candidates are the only candidates gathered when forming
the media sections. To address this issue, we use 0.0.0.0:9 when a
hostname candidate would be used to populate the c= and the m= lines.
The SDP grammar related to ICE candidates has been moved out of RFC8445,
and is currently defined in draft-ietf-mmusic-ice-sip-sdp. A FQDN
address must not be used in the connection address attribute per the
latest draft, if the ICE agent generates local candidates. Also, the
wildcard addresses (0.0.0.0 or ::) with port 9 are given the exception
as the connection address that will not result in an ICE mismatch. We
thus adopt the aforementioned solution after combining these
considerations.
Bug: chromium:927309, chromium:982108
Change-Id: I3df2db0f154276da39f99650289cf81baa677e74
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145280
Commit-Queue: Qingsi Wang <qingsi@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28547}
This reverts commit 7325bc3917e6dd4c92e7a18fd879ba91f0b2851f.
Reason for revert: FecTest.UlpfecTest is consistently failing.
Original change's description:
> Refactor FEC code to use COW buffers
>
> This refactoring helps to reduce unnecessary memcpy calls on the receive
> side.
>
> This CL is the first stage of refactoring: it only replaces
> |uint8 data[IP_PACKET_SIZE]| with |rtc::CopyOnWriteBuffer data| and does
> necessary changes.
>
> A follow-up CL will remove length field of the Packet class.
>
>
> Bug: webrtc:10750
> Change-Id: Ie233da83ff33f6370f511955e4c65d59522389a7
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144881
> Reviewed-by: Artem Titov <titovartem@webrtc.org>
> Reviewed-by: Stefan Holmer <stefan@webrtc.org>
> Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
> Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#28539}
TBR=brandtr@webrtc.org,ilnik@webrtc.org,asapersson@webrtc.org,stefan@webrtc.org,titovartem@webrtc.org
Change-Id: I07c34256a76174f09a0d27eacbae6488e66f4b43
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: webrtc:10750
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145340
Reviewed-by: Qingsi Wang <qingsi@webrtc.org>
Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28545}
As per the spec, you should be able to use simulcast with screenshare.
We remove the field trial for it and keep the old behavior only for
screenshare sources with conference flag on.
Bug: webrtc:8785
Change-Id: I1d6d4e18256fb5cfe0195620706de068f25b8d9b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144785
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28543}
This change removes RTCP Feedback loop if we are using datagram transport by removing transport sequence numbers from RTP packets and recreating RTCP Feedback from Datagram ACKs and Timestamps.
- For outgoing RTP packets, remove transport sequence number and store it with datagram_id. Note that removing transport sequence numbers does not only save 4-8 bytes per packet, but also prevents generation of feedback packets on the receiver side.
- When datagram ACKs, we re-created RTCP feedback with timestamp.
- Replacing previous assumption that datagram_id was the same as packet_id by storing incremental counter of datagram ids (I noticed some packets come without packet_id, which is a bit strange, but easy to support and it's also good not to rely on packet_ids being unique across multiple ssrcs).
Bug: webrtc:9719
Change-Id: Iecfe938ecea1a74e7c9e1484f0e985d72643d4a1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145269
Commit-Queue: Anton Sukhanov <sukhanov@webrtc.org>
Reviewed-by: Steve Anton <steveanton@webrtc.org>
Reviewed-by: Bjorn Mellem <mellem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28542}
Problem: Any 'Reformat WebRTC code base' commit pollutes blame history.
Solution: Add it in ignore list.
Caveats: (from man git-hyper-blame)
When a line skips over an ignored commit, a guess is made as to which
commit previously modified that line, but it is not always clear where
the line came from.
If the ignored commit makes lots of changes in close proximity,
in particular adding/removing/reordering lines, then the wrong
authors may be blamed for nearby edits.
Bug: webrtc:9340
Change-Id: I0fd86e65edd6f63685c20da929a6b750fb0fb600
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145218
Commit-Queue: Yves Gerey <yvesg@google.com>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Henrik Andreassson <henrika@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28541}
This refactoring helps to reduce unnecessary memcpy calls on the receive
side.
This CL is the first stage of refactoring: it only replaces
|uint8 data[IP_PACKET_SIZE]| with |rtc::CopyOnWriteBuffer data| and does
necessary changes.
A follow-up CL will remove length field of the Packet class.
Bug: webrtc:10750
Change-Id: Ie233da83ff33f6370f511955e4c65d59522389a7
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144881
Reviewed-by: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28539}
The (currently unused) new pacer code path was implemented with what
was intended as a more careful padding strategy.
Unfortunately this doesn't work as well as expected due to the fact
that the padding budget cannot build up underuse.
I made another CL that could fix that, but I think it adds complexity
for dubious gains. It also will make it more difficult to find any
potential regression when switching to the new path, should one occur.
See https://webrtc-review.googlesource.com/c/src/+/144563
Therefore, this CL makes the new code path choose RTX payload in the
same way as is currently done.
Bug: webrtc:10633
Change-Id: If2115d4fa7463add959faa77c63101286c27e0f5
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145202
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28537}
The legacy code-path uses a hashmap as cache in order to speed up
finding the right rtp module to send on. The new path should use that
as well.
In addition, there are checks that verify if an RTP module can send
padding, in some cases payload based. These result in a number of
calls to methods in RTPSender requiring its lock to be taken. This CL
introduces a combined SupportsPadding() check method which performs
all those checks in one go.
Bug: None
Change-Id: I2d18d0d6e7d8cfe92c81d08cef248a4daa7dcd4b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144780
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Reviewed-by: Sebastian Jansson <srte@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28535}
Datagram_dtls_adaptor needs access to rtp_rtcp modules and this moves helps to keep p2p/base/ without dependency on rtp_rtcp.
Bug: webrtc:9719
Change-Id: Ic337be3fb9f68106187a84efa815eefbe5b0fcd7
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145267
Reviewed-by: Steve Anton <steveanton@webrtc.org>
Commit-Queue: Anton Sukhanov <sukhanov@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28533}
Removing extension will be used in DatagramDtlsAdaptor to remove transport sequence number to avoid having both datagram and RTP feedback loops. The sequence number will be stored in temporary map and used to re-create RTCP fdeedback packed when we receive datagram ACK. It would enable integration of Datagram transport without any changes in the upper layers of RTP stack. Note that Datagram adaptor changes will be implemented in a separate changelist.
In this change:
- Implement method to remove extension by rebuilding entire packet without given extension type.
- Fails if extension is not registered or not set.
- Unit test
Bug: webrtc:9719
Change-Id: I9d3811d5d97fadde5a294d5da643b2ebc6a8196e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145100
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Bjorn Mellem <mellem@webrtc.org>
Commit-Queue: Anton Sukhanov <sukhanov@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28530}
This is a reland of 2d0880b56954f57548deea51dfa678b80dbf618f
To fix perf bot issue reading of perf results file was updated. Now perf
results file will be generated by each test and then returned via output to
the python script, which will get it and put into final file.
Original change's description:
> Fix collection of audio metrics from PC test framework for audio test
>
> Bug: webrtc:10138
> Change-Id: I18a8509a0cdc4ed1db6894c7540d5c0a155d6233
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144784
> Reviewed-by: Oleh Prypin <oprypin@webrtc.org>
> Reviewed-by: Oskar Sundbom <ossu@webrtc.org>
> Commit-Queue: Artem Titov <titovartem@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#28514}
Bug: webrtc:10138
Change-Id: I1347f09427736362a2d550612b48e09c06cfb1d1
No-Presubmit: True
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145201
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Oskar Sundbom <ossu@webrtc.org>
Reviewed-by: Oleh Prypin <oprypin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28527}
This is a trivial CL, updating rtp_generator.cc according to changes in
APIs in other places.
Bug: webrtc:10807
Change-Id: Ie85c6283f2d78dcf742979378db0b4fb0914c96c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/145209
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28526}
Users of webrtc generally should be able to choose own task queue implementation.
Poison avoids accidental dependency of a low level component on the default implementation
Android and ios apis are still de-facto forced to use the default implementation.
Bug: webrtc:10284
Change-Id: I67ecf2317f43ee32b0c9e8a6e69f1e0987cf1914
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144786
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28524}
This is a reland of 94c58fd815f0c7c6429aa53a79621ea9ef39c770
Patch set 1 is the original CL.
Patch set 2 introduced a trivial fix. In RtcpSender::SetSSRC(), check
if either current SSRC is 0 or if the SSRC is identical to the current
one. If so, don't schedule an early report.
This prevents a regression in which audio jitter became too low(?)
Original change's description:
> Add ability to set RTCP sender ssrc at construction time
>
> Bug: webrtc:10774
> Change-Id: Iaf5857e24359e9795434bcd0cdbe1658a2f9f5d3
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144632
> Reviewed-by: Åsa Persson <asapersson@webrtc.org>
> Commit-Queue: Erik Språng <sprang@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#28506}
Bug: webrtc:10774
Change-Id: I103dfa48719aa41d6ab633cdac8b3a5c46b54843
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144565
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Åsa Persson <asapersson@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28520}
Ensures that newly added ChannelMixer is utilized when number of channels
is larger than two in the output mixer.
Decided to land with henrik.lundin as TBR since he has reviewed all other
changes in WebRTC related to channel mixing for multi-channel cases.
All this CL does is to ensure that the new channel mixing scheme can be used
in Chrome. The old scheme is still used for mono and stereo combinations.
TBR: henrik.lundin
Bug: webrtc:10783
Change-Id: I11c02f1b4ef60e847095efbcd5e5f5faf27a5cdd
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/140290
Commit-Queue: Henrik Andreassson <henrika@webrtc.org>
Reviewed-by: Henrik Andreassson <henrika@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28517}
If VideoEncoderConfig::max_bitrate_bps is unset then max bitrate of
video stream is set equal to max bitrate value recommended by encoder
for given resolution via encoder capabilities (if available).
Bug: webrtc:10796
Change-Id: I7fce9afc476b794a16956e694e891faee110048e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/144526
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28515}
A cloaked window is composited but not visible to the user.
When Win10 feature 'Cortana' is enabled it creates a window
that is always invisible and its z-order is top most. Because
of that the cropping capturer detects occlusion everywhere
preventing it from capturing anything.
The solution is to ignore all cloaked windows like if
::IsWindowVisible would return false.
Bug: chromium:978885
Change-Id: Id5aa8dc81dcf4979ffb30dd808fa2a553934c6e6
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/143980
Commit-Queue: Julien Isorce <julien.isorce@chromium.org>
Reviewed-by: Sergey Ulanov <sergeyu@chromium.org>
Reviewed-by: Guido Urdaneta <guidou@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28510}
