Commit graph

120971 commits

Author SHA1 Message Date
Nicolas Dufresne
ab70aa60e2 xvimagesink: Fix crash in pool on error
The set_config() virtual function is not support to free the config. As a side
effect, when there is protocol error of some sort, we endup with a crash.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7285>
2024-08-01 16:42:35 +00:00
Nicolas Dufresne
5df658cfdd qt6: glwindow: Don't leak previously rendered buffer
If the consumer reads the buffers too slowily, simply unref the
previously rendered buffer instead of leaking it.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7286>
2024-08-01 12:44:06 +00:00
Víctor Manuel Jáquez Leal
28e16f897e vkimagebufferpool: fix documentation grammar
Original-patch-by: Matthew Waters <matthew@centricular.com>
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7288>
2024-08-01 10:09:34 +00:00
Carlos Bentzen
48ae40f477 webrtcbin: create and associate transceivers earlier in negotation
According to https://w3c.github.io/webrtc-pc/#set-the-session-description
(steps in 4.6.10.), we should be creating and associating transceivers when
setting session descriptions.

Before this commit, webrtcbin deviated from the spec:
1. Transceivers from sink pads where created when the sink pad was
   requested, but not associated after setting local description, only
   when signaling is STABLE.
2. Transceivers from remote offers were not created after applying the
   the remote description, only when the answer is created, and were then
   only associated once signaling is STABLE.

This commit makes webrtcbin follow the spec more closely with regards to
timing of transceivers creation and association.

A unit test is added, checking that the transceivers are created and
associated after every session description is set.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7156>
2024-08-01 07:38:46 +00:00
Víctor Manuel Jáquez Leal
ef9875640e vulkanupload: honor downstream pool allocation parameters
If a downstream buffer pool is offered, vulkanupload checks its allocation
parameters to honor them. Only adds to usage the TRANSFER bits, which are
required to upload buffers.

Also, fail if the buffer pool cannot be configured with the current parameters.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7219>
2024-07-31 12:15:43 +00:00
Víctor Manuel Jáquez Leal
baac191d13 vkimagebufferpool: expose config_get_allocation_params()
Also enhanced the documentation and added a config parameter check for
gst_vulkan_image_buffer_pool_config_set_allocation_params()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7219>
2024-07-31 12:15:43 +00:00
Shengqi Yu
7576d14762 v4l2object: append non colorimetry structure to probed caps
If the stream has a special colorimetry that is not in the colorimetry
list, it will cause negotiation to fail. We should allow passing any
colorimetry, so add an extra structure without the colorimetry field.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7029>
2024-07-31 09:28:18 +00:00
Hou Qi
5dffbd492c v4l2: Fix colorimetry mismatch for encoded format with RGB color-matrix
video-info supports encoded format to have RGB color-matrix, while
v4l2object just leave the v4l2 matrix to default when mapping
GST_VIDEO_COLOR_MATRIX_RGB. It causes gst matrix changed to be
GST_VIDEO_COLOR_MATRIX_BT601 when mapping v4l2 colorimetry.

So add support for encoded format with RGB color-matrix in v4l2object.
Note that for M2M encoders, we should in theory assume that that we can
transfer this value from OUTPUT to CAPTURE queues, though its only true
if the drivers does not do CSC. For now, we don't support any RGB
codecs, but leaving a note for the future.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3952>
2024-07-30 20:26:06 +00:00
Nicolas Dufresne
1ddb8797b5 v4l2object: SRGB colorspace is documented limited-range
Split JPEG and SRGB so that we can follow the specified difference. The
SRGB definition in V4L2 does not follow the standard, and is document
so. This is also why JPEG colorspace exists.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3952>
2024-07-30 20:26:06 +00:00
Nicolas Dufresne
20eb14b85b v4l2object: Fix size of plane_size array calculation
Due to missing parenthesys, only the first element of the array was
being cleared. As it is a staticly sized array in the object, this
code could also be simplified.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3952>
2024-07-30 20:26:06 +00:00
Nicolas Dufresne
152df21644 v4l2object: Fix translation of quantization
The V4L2_MAP_QUANTIZATION macro has been fixed to something a lot saner,
fix our replica accordingly. The new macro now simply set the quantization
to full range is the pixel formats is RGB based, or if the JPEG
colorspace is used.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3952>
2024-07-30 20:26:06 +00:00
Guillaume Desmottes
beb1cef3d4 rsvgoverlay: add debug category
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7246>
2024-07-30 17:51:51 +00:00
David Rosca
6fe5d9013d vaapi: Fix sps_max_dec_pic_buffering_minus1 value in h265 decoder
Fixes decoding SLPPLP_A_VIDYO_2 sample.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7242>
2024-07-30 17:31:07 +00:00
Daniel Morin
4ff98d66bb onnx: fix formating
Code alignement was not alway consistent

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7205>
2024-07-30 16:50:15 +00:00
Víctor Manuel Jáquez Leal
43edff4e29 vaapi: tests: fix uninitialized variable compiler warning
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7279>
2024-07-30 15:06:25 +00:00
Edward Hervey
b7828d237b nlecomposition: Don't leak QoS events
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7276>
2024-07-30 14:22:54 +00:00
Edward Hervey
a38e244794 vulkan: Add missing since markers
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7277>
2024-07-30 13:46:42 +00:00
Edward Hervey
30c2f67e49 bad: Add missing plugin since
These predate current stable release. Was never spotted since they weren't built
on the CI

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7277>
2024-07-30 13:46:42 +00:00
Nirbheek Chauhan
fa25dd4de5 vtenc: Do not set the CRF/Quality parameter with ProRes
It's not supported with ProRes, setting the property will fail.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7232>
2024-07-30 12:42:17 +00:00
Nirbheek Chauhan
cbadb0b4ae vtenc: Log warnings when setting a property fails
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7232>
2024-07-30 12:42:17 +00:00
Nirbheek Chauhan
f8711239e4 vtenc: Add max-frame-delay property
This controls the number of frames allowed in the compression window.
Not all encoders and implementations support it; I've only managed to
successfully use it with ProRes.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7232>
2024-07-30 12:42:17 +00:00
Nirbheek Chauhan
30d2b8895a vtenc: Support emulating CBR mode with data rate limits
CBR is only supported on Apple Silicon, and this "emulation" works
surprisingly well. We set the window size to a single frame, and don't
set ABR at all.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7232>
2024-07-30 12:42:17 +00:00
Nirbheek Chauhan
d327105228 vtenc: Add new property for setting data rate limits
This proxies kVTCompressionPropertyKey_DataRateLimits, except it
only supports a single limit for now.
https://developer.apple.com/documentation/videotoolbox/kvtcompressionpropertykey_dataratelimits

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7232>
2024-07-30 12:42:17 +00:00
Nirbheek Chauhan
4207b4e281 vtenc: Add support for constant bitrate encoding
Only supported on macOS 13.0+ and iOS 16.0+ on Apple Silicon.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7232>
2024-07-30 12:42:17 +00:00
Nirbheek Chauhan
2e39aefe22 vtenc: Eliminate some needless complex code
We do not need a helper that takes a lock to fetch the values of these
properties. There is no race being prevented.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7232>
2024-07-30 12:42:17 +00:00
Nirbheek Chauhan
3621b82c63 vtenc: Fix setting of quality property
gst_vtenc_set_quality() will never actually set the VT compression
property, because it tries to set it on self->session which is not
initialized at this point.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7232>
2024-07-30 12:42:17 +00:00
Edward Hervey
c623f2db21 validate: reporter: Don't pollute logs
There was a stray g_printerr introduced by 2a4b9c8dc1

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7275>
2024-07-30 11:39:02 +02:00
Jan Schmidt
0faec707a0 adaptivedemux: Fail cleanly if parsebin is not installed
Detect a failure to construct a parsebin and error out
cleanly instead of trying to operate on a null pointer.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6969>
2024-07-30 00:06:50 +00:00
Jan Schmidt
213726ca41 adaptivedemux2: Post a bus error when failing to start download
If a download completely fails to start, due to malformed URI or so,
post a bus error instead of just stalling out with no indication
why.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6969>
2024-07-30 00:06:50 +00:00
Jan Schmidt
f2a18ab277 adaptivedemux2: Implement file:// URI handling
Add the ability to play HLS and DASH from local files

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6969>
2024-07-30 00:06:50 +00:00
Jan Schmidt
9dc1d68e2f codec-utils: Recognise mjpg mime type
Add mjpg to gst_codec_utils_caps_from_mime_codec_single()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6969>
2024-07-30 00:06:50 +00:00
Jan Schmidt
ef0e822559 hlsng: Check caps are not null after parsing HLS CODECS tag
If the mime codec wasn't recognised, caps will be NULL and cause
a critical

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6969>
2024-07-30 00:06:50 +00:00
Jan Schmidt
7bd51afd04 gsturi: Ensure file:// URIs retain //
Add the // back after the scheme for file URIs so 'file:///path/to/file'
doesn't become 'file:/path/to/file' in gst_uri_to_string()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6969>
2024-07-30 00:06:50 +00:00
Víctor Manuel Jáquez Leal
bff24e5ab7 vulkanupload: comment zero value usage with VK_ACCESS_NONE
Zero is used only for Vulkan version prior to 1.3, because it wasn't defined
before.

Just for readability.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7247>
2024-07-29 17:10:15 +02:00
Víctor Manuel Jáquez Leal
fad06c9b6f vkimagebufferpool: refactor how image usage is set
Now that driver version is expected to be equal or superior to 1.3.275 the bug
in NVIDIA and RADV regarding usage is solved, we can revert commit b7ded81f7b.

Also this patch sets the internal usage variable after all the validation are
run, thus the state don't keep an invalid usage.

Finally, the now unused supported_usage variable is dropped.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7247>
2024-07-29 17:10:15 +02:00
Víctor Manuel Jáquez Leal
bb9eb6f477 vkimagebufferpool: add encoding usage as video usage
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7247>
2024-07-29 17:10:15 +02:00
Víctor Manuel Jáquez Leal
541e2e93ca vkimagebufferpool: reset the number of profiles at set_config()
Virtual method set_config() can be called several times, and if the number of
profiles counter isn't reset the pool will reach an error state.

The purpose of number of profiles is to check the number of valid vulkan video
profiles (two in the case of transcoding use-case, for example) so it's local to
set_config() virtual method.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7247>
2024-07-29 17:10:14 +02:00
Víctor Manuel Jáquez Leal
d8e384085a vulkan: encoder and decoder runtime check for driver version 1.3.275
Which is the one checked in meson. See commit 21ee264d65

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7247>
2024-07-29 17:10:14 +02:00
Philippe Normand
bd64374174 parsebin: accept-caps handling for elements with unusual pad names
In case the last element of the parse chain doesn´t have a sink pad named
"sink", send the accept-caps query to the first sink pad of the element.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7255>
2024-07-29 12:41:33 +00:00
Víctor Manuel Jáquez Leal
8faefccfd5 va: refactor dmabuf handle close
Moved the close loop into a function guarded for non-win32 platforms.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7254>
2024-07-29 10:34:56 +02:00
Seungha Yang
080a838fca qsv: Fix critical warnings
Fixing warnings
GStreamer-CRITICAL **: 01:21:25.862: gst_value_set_int_range_step:
assertion 'start < end' failed

Although when QSV runtime reports a codec is supported, resolution query
fails sometimes, espeically VP9 encoder case on Windows.
Don't try to register an element if resolution query returned an error

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7250>
2024-07-27 19:33:52 +00:00
Nirbheek Chauhan
0bb336f34c svtav1enc: Fix segfault when flushing
gst_video_encoder_get_oldest_frame() is nullable, and will signal that
all frames are handled by returning NULL.

Fixes #3650

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7249>
2024-07-27 19:21:37 +01:00
Seungha Yang
f04f6f43c3 d3d12frame: Fix frame copy method
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7243>
2024-07-27 04:18:47 +09:00
Seungha Yang
28a7adf4dd d3d12memory: Check heap flag before trying to create NT handle
CreateSharedHandle() will fail eventually if the resource was created
with non-shared heap. Instead of trying to create handle blindly,
validate resource first.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7243>
2024-07-27 04:17:03 +09:00
Seungha Yang
1c8c5ed457 dwrite: Prefer d3d12 resource allocated with shared heap
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7243>
2024-07-27 03:39:22 +09:00
Seungha Yang
4a6daad517 d3d12: Suppress fence data object leak report
We don't release GstD3D12Device intentionally, thus
a GstD3D12FenceDataPool owned by a device will not be released
but that's expected leak.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7243>
2024-07-27 03:23:22 +09:00
Seungha Yang
9a8d7db5de d3d12: Fix debug category name
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7243>
2024-07-27 03:23:22 +09:00
Seungha Yang
88efaa35a8 d3d12download: Do not overwrite fence of non-writable memory
A fence configured in GstD3D12Memory should be used only for
write access to be completed. And because d3d12 -> d3d11 copy path
is read access to d3d12 resource, we should not set fence to
memory. Otherwise another read access to the d3d12 resource
will wait for d3d11 device context's copy operation although
simultaneous read access is allowed.

Use background thread to keep d3d12 resource and wait for d3d11 device's
copy operation instead.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7243>
2024-07-27 03:23:22 +09:00
Jan Schmidt
1ebbcc31ef va: Fix dmabuf handle leaks
Close dmabuf handles manually when they're not going to
be passed into GStreamer FD memory, to avoid fd handle
leaks.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7240>
2024-07-26 10:32:48 +00:00
Jordan Yelloz
f11735c97b tsmux: Adjust byte counter when adjusting bitrate
When configured in constant bitrate mode, the muxer computes timing information
using the configured bitrate and the byte counter (now = bytes sent / byterate).

When an application changes the bitrate in CBR mode during playback, the
relationship between bytes sent and bitrate is no longer valid so new timing
values will be off by the ratio of the old bitrate to the new bitrate.
Furthermore, it will upset the way that padding is generated.

pad_stream() works by trying to fit the byte counter to now * byterate.
The result is that when decreasing bitrate, the muxer stalls, waiting until the
byte counter is in agreement with now * byterate. Also, when increasing
bitrate, the padding will spike in volume until the byte counter fits with
now * byterate.

If the byte counter is scaled by the ratio of new bitrate / old bitrate when
adjusting bitrate, then padding is generated in a way that applications would
more likely expect.

One detail this change doesn't yet address is whether the next PCR will match up
optimally with the previous PCR right after the byte counter is scaled. In that
case, some correction may be necessary. Also, perhaps the user should be
prevented from changing from bitrate=0 to bitrate=nonzero during playback since
it's not straightforward how to scale the byte counter in that case.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7158>
2024-07-25 15:02:11 +00:00