platform/upstream/gstreamer.git
3 years agotranscoder: Handle the case where several errors are posted
Thibault Saunier [Thu, 19 Nov 2020 20:55:10 +0000 (17:55 -0300)]
transcoder: Handle the case where several errors are posted

There were cases where the loop was already destroyed when we were
receiving the following message.

3 years agotranscoder: Minor refactoring to output better debug logs
Thibault Saunier [Thu, 19 Nov 2020 20:54:28 +0000 (17:54 -0300)]
transcoder: Minor refactoring to output better debug logs

3 years agohlssink2: Mark as Muxer
Thibault Saunier [Thu, 19 Nov 2020 20:51:56 +0000 (17:51 -0300)]
hlssink2: Mark as Muxer

The way it is usable by encodebin2. This is what splitmux does already.

3 years agova: decoder: Picture dups only holds GstBuffer
Víctor Manuel Jáquez Leal [Mon, 30 Nov 2020 16:12:14 +0000 (17:12 +0100)]
va: decoder: Picture dups only holds GstBuffer

Also removes the warning log message at destroying buffers when picture free()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1841>

3 years agova: Remove gst_va_decoder_destroy_buffers()
Víctor Manuel Jáquez Leal [Mon, 30 Nov 2020 14:01:01 +0000 (15:01 +0100)]
va: Remove gst_va_decoder_destroy_buffers()

Since GstVaDecodePicture is destroyed completely with its free() function and
it's used as destroy notify by codecs picture, there's no need to call
gst_va_decoder_destroy_buffers() externally, since the codecs base classes
destroy the codec picture when it's required.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1841>

3 years agova: Destroy picture unreleased buffers when finalize.
He Junyan [Thu, 26 Nov 2020 06:04:31 +0000 (14:04 +0800)]
va: Destroy picture unreleased buffers when finalize.

The current way of GstVaDecodePicture's finalize will leak some
resource such as parameter buffers and slice data.
The current way deliberately leaves these resource releasing logic
to va decoder related function and trigger a warning if we free the
GstVaDecodePicture without releasing these resources.
But in practice, sometimes, you do not have the chance to release
these resource before picture is freed. For example, H264/Mpeg2
support multi slice NALs/Packets for one frame. It is possible that
we already succeed to parse and generate the first several slices
data by _decode_slice(), but then we get a wrong slice NAL/packet
and fail to parse it. We decide to discard the whole frame in the
decoder's base class, it just free the current picture and does not
trigger sub class's function again. In this kind of cases, we do
not have the chance to cleanup the resource, and the resource will
be leaked.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1841>

3 years agoqroverlay: Reuse the same OverlayComposition object when possible
Thibault Saunier [Sat, 21 Nov 2020 22:00:02 +0000 (19:00 -0300)]
qroverlay: Reuse the same OverlayComposition object when possible

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1829>

3 years agoqroverlay: Rework basing it on overlaycomposition
Thibault Saunier [Fri, 20 Nov 2020 14:28:25 +0000 (11:28 -0300)]
qroverlay: Rework basing it on overlaycomposition

The base class is now a bin which wraps the `overlaycomposition`
element and implements the `draw` signal.

This way we support all the video formats the GstVideoOverlayComposition
API supports and the blending code can be reused. It is also possible
to have the blending happen in the sinks now.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1829>

3 years agod3d11h264dec: Reconfigure decoder object on DPB size change
Seungha Yang [Wed, 25 Nov 2020 20:55:29 +0000 (05:55 +0900)]
d3d11h264dec: Reconfigure decoder object on DPB size change

Even if resolution and/or bitdepth is not updated, required
DPB size can be changed per SPS update and it could be even
larger than previously configured size of DPB. If so, we need
to reconfigure DPB d3d11 texture pool again.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1839>

3 years agoaudio: Use new AudioFormatInfo::fill_silence function
Marijn Suijten [Wed, 25 Nov 2020 16:52:42 +0000 (17:52 +0100)]
audio: Use new AudioFormatInfo::fill_silence function

The function is renamed to be properly associated with AudioFormatInfo
(its instance) instead of AudioFormat (an unrelated enum), see [1] for
the rename itself.

[1]: https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/940

3 years agoplayer: Fix get_current_subtitle_track annotation
Philippe Normand [Thu, 5 Nov 2020 17:14:22 +0000 (17:14 +0000)]
player: Fix get_current_subtitle_track annotation

As the info returned is a new object, the annotation should be transfer-full,
similarly to the get_current_{audio,video}_track() implementations.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1775>

3 years agova: allocator: add a memory pool object helper
Víctor Manuel Jáquez Leal [Mon, 23 Nov 2020 19:44:27 +0000 (20:44 +0100)]
va: allocator: add a memory pool object helper

Since both allocators use a memory pool, with its mutex and cond, this patch
refactors it into a single internal object, implementing a generic GstMemory
pool.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1815>

3 years agova: pool, allocator: honor GST_BUFFER_POOL_ACQUIRE_FLAG_DONTWAIT
Víctor Manuel Jáquez Leal [Tue, 17 Nov 2020 13:53:05 +0000 (14:53 +0100)]
va: pool, allocator: honor GST_BUFFER_POOL_ACQUIRE_FLAG_DONTWAIT

In order to honor GST_BUFFER_POOL_ACQUIRE_FLAG_DONTWAIT in VA pool, allocators'
wait_for_memory() has to be decoupled from their prepare_buffer() so it could be
called in pools' acquire_buffer() if the flag is not set.

wait_for_memory() functions are blocking so the received memories are assigned
to the fist requested buffer, if multithreaded calls. For this a new mutex were
added.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1815>

3 years agova: allocator: broadcast when flushing
Víctor Manuel Jáquez Leal [Tue, 17 Nov 2020 12:18:37 +0000 (13:18 +0100)]
va: allocator: broadcast when flushing

This patch handles when the bufferpool request a new buffer while
flushing.

Also fixes the usage of g_cond_wait(), which demands to be used
inside a loop to avoid spurious wakeups.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1815>

3 years agova: allocator: free allocator when a mem is held
Víctor Manuel Jáquez Leal [Tue, 17 Nov 2020 12:17:03 +0000 (13:17 +0100)]
va: allocator: free allocator when a mem is held

An application, using for example appsink, can hold buffers from any
va allocator after setting the pipeline to NULL. We need to destroy
the allocator when that memory is unrefed.

This patch juggles a bit with the allocator reference count in
memories in order to achieve this:

1. When memory is created no alloc ref is modified
2. When memory is released, alloc ref is decreased
3. When memory is reassiged to a buffer, alloc ref is increased
4. When memory is flushed, alloc ref is increased becase it is going
   to be decreased in gst_memory_unref()

Also this patch moves the deallocation of member variables to
finalize() rather than dispose()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1815>

3 years agova: allocator: dmabuf: initialize cond
Víctor Manuel Jáquez Leal [Mon, 23 Nov 2020 16:01:52 +0000 (17:01 +0100)]
va: allocator: dmabuf: initialize cond

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1815>

3 years agowebrtc: Make ssrc map into separate data structures
Olivier Crête [Fri, 20 Nov 2020 22:32:44 +0000 (17:32 -0500)]
webrtc: Make ssrc map into separate data structures

They now contain a weak reference and that could be freed later
causing strange crashes as GWeakRef are not movable.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1766>

3 years agowebrtcstats: Get the remote-inbound stats from the right RTPSource
Olivier Crête [Fri, 16 Oct 2020 01:23:08 +0000 (21:23 -0400)]
webrtcstats: Get the remote-inbound stats from the right RTPSource

This also means that we need to get the clock-rate from the codec instead
of from the RTPSource, as the remote one doesn't include a clock rate.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1766>

3 years agowebrtcbin: Implement getting stats for a specific pad
Olivier Crête [Thu, 15 Oct 2020 23:36:45 +0000 (19:36 -0400)]
webrtcbin: Implement getting stats for a specific pad

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1766>

3 years agowebrtcstats: Also return the raw rtpsource stats for more information
Olivier Crête [Sat, 10 Oct 2020 22:21:19 +0000 (18:21 -0400)]
webrtcstats: Also return the raw rtpsource stats for more information

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1766>

3 years agowebrtcstats: Avoid copy of GstStructure
Olivier Crête [Sat, 10 Oct 2020 00:59:58 +0000 (20:59 -0400)]
webrtcstats: Avoid copy of GstStructure

Instead transfer the ownership to the new structure

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1766>

3 years agowebrtcstats: Remove receiver side when sending
Olivier Crête [Sat, 10 Oct 2020 00:45:10 +0000 (20:45 -0400)]
webrtcstats: Remove receiver side when sending

Those are just invalid and just reflect what we sent. We'd need to parse the
RTCP XR packets from the other side to know more about those.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1766>

3 years agowebrtcstats: Extract statistics from the rtpjitterbuffer
Olivier Crête [Sat, 10 Oct 2020 00:27:40 +0000 (20:27 -0400)]
webrtcstats: Extract statistics from the rtpjitterbuffer

And expose them as standardised webrtc statistics

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1766>

3 years agowebrtcbin: Store the rtpjitterbuffer instances to extract stats from them
Olivier Crête [Fri, 9 Oct 2020 22:45:57 +0000 (18:45 -0400)]
webrtcbin: Store the rtpjitterbuffer instances to extract stats from them

Store them as web refs to avoid having to worry about freeing later and because
the new-jitterbuffer is on a different thread

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1766>

3 years agowebrtcstats: Document all RTP missing fields according to the latest spec
Olivier Crête [Fri, 9 Oct 2020 23:59:18 +0000 (19:59 -0400)]
webrtcstats: Document all RTP missing fields according to the latest spec

Just document all the missing fields and document which ones will never
be implemented because they depend on the codec or depayloader

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1766>

3 years agowebrtcstats: RTCP computed RTT is only available at sender
Olivier Crête [Fri, 9 Oct 2020 23:38:15 +0000 (19:38 -0400)]
webrtcstats: RTCP computed RTT is only available at sender

The receiver doesn't have the information to compute it.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1766>

3 years agowebrtcstats: Remove redundant lines
Olivier Crête [Thu, 8 Oct 2020 21:11:30 +0000 (17:11 -0400)]
webrtcstats: Remove redundant lines

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1766>

3 years agowebrtc: Also remove rtcp_transport from the structure
Olivier Crête [Wed, 4 Nov 2020 22:06:02 +0000 (17:06 -0500)]
webrtc: Also remove rtcp_transport from the structure

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1765>

3 years agowebrtc: Remove APIs to set transport on sender/receiver
Olivier Crête [Tue, 3 Nov 2020 00:55:46 +0000 (19:55 -0500)]
webrtc: Remove APIs to set transport on sender/receiver

They're not not used ever.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1765>

3 years agowebrtc: Remove non rtcp-mux code
Olivier Crête [Tue, 3 Nov 2020 00:49:55 +0000 (19:49 -0500)]
webrtc: Remove non rtcp-mux code

RTCP mux is now always required by the WebRTC spec

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1765>

3 years agonvcodec: Assume 25fps if framerate is invalid when calculating latency
Julian Bouzas [Fri, 20 Nov 2020 15:01:03 +0000 (15:01 +0000)]
nvcodec: Assume 25fps if framerate is invalid when calculating latency

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1826>

3 years agocodecs: h264decoder: fix memory leak
Víctor Manuel Jáquez Leal [Fri, 20 Nov 2020 21:26:14 +0000 (22:26 +0100)]
codecs: h264decoder: fix memory leak

gst_h264_dbp_get_picture_all() returns a full transfer of the GArray, which
needs be unrefed. But it is not unrefed in
gst_h264_decoder_find_first_field_picture() leaking it.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1827>

3 years agompegts: Documentation fixes
Edward Hervey [Fri, 20 Nov 2020 15:07:36 +0000 (16:07 +0100)]
mpegts: Documentation fixes

gtk-doc was complaining :)

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1825>

3 years agoqroverlay: unset executable flag on source files
Tim-Philipp Müller [Fri, 20 Nov 2020 13:24:24 +0000 (13:24 +0000)]
qroverlay: unset executable flag on source files

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1824>

3 years agoqroverlay: fix auto detection of json-glib for plugin
Tim-Philipp Müller [Fri, 20 Nov 2020 13:22:48 +0000 (13:22 +0000)]
qroverlay: fix auto detection of json-glib for plugin

Only want to check for json-glib when libqrencode was found,
but also it shouldn't be required but depend on the option.

Fixes #1465

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1824>

3 years agod3d11: Add support for packed 4:2:2 and 4:4:4 10bits formats
Seungha Yang [Thu, 19 Nov 2020 12:15:25 +0000 (21:15 +0900)]
d3d11: Add support for packed 4:2:2 and 4:4:4 10bits formats

Add support for Y210 and Y410 formats which are commonly used format
for en/decoders on Windows. Note that those formats cannot be used for
render target (output) of shader.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1821>

3 years agoopenh264dec: Accept constrained-high and progressive-high profiles
Olivier Crête [Fri, 2 Oct 2020 22:47:16 +0000 (18:47 -0400)]
openh264dec: Accept constrained-high and progressive-high profiles

They're just subsets of the high profile.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1634>

3 years agod3d11h264dec: Accept constrained-high and progressive-high profiles
Olivier Crête [Fri, 2 Oct 2020 22:47:06 +0000 (18:47 -0400)]
d3d11h264dec: Accept constrained-high and progressive-high profiles

They're just subsets of the high profile.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1634>

3 years agomsdkh264dec: Accept constrained-high and progressive-high profiles
Olivier Crête [Fri, 2 Oct 2020 22:46:56 +0000 (18:46 -0400)]
msdkh264dec: Accept constrained-high and progressive-high profiles

They're just subsets of the high profile.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1634>

3 years agonvdec: Accept progressive-high and contrained-high profiles
Olivier Crête [Tue, 22 Sep 2020 19:42:37 +0000 (15:42 -0400)]
nvdec: Accept progressive-high and contrained-high profiles

They're subsets of the high profiles with no interlacing and
no B-frames for constrained

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1634>

3 years agocodecparsers: av1: add the set_operating_point() API.
He Junyan [Mon, 28 Sep 2020 05:33:00 +0000 (13:33 +0800)]
codecparsers: av1: add the set_operating_point() API.

The av1 can support multi layers when scalability is enabled. We
need an API to set the operating point and filter the OBUs just
belonging to some layers(the layers are specified by the operating
point).

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: Add an API to reset the annex_b state only.
He Junyan [Fri, 9 Oct 2020 08:13:28 +0000 (16:13 +0800)]
codecparsers: av1: Add an API to reset the annex_b state only.

In practice, we encounter streams that have one or more temporal units
error. When that kind of error temporal units is in annex b format, the
whole temporal unit should be discarded.
But the temporal units before it are correct and can be used. More
important, because of the error temporal unit, the parser is in a wrong
state and all later temporal unit are also parsed uncorrectly.
We need to add this API to reset the annex_b state only when we meet
some temporal unit error.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: clean the seen_frame_header in parse_tile_group().
He Junyan [Fri, 9 Oct 2020 08:01:35 +0000 (16:01 +0800)]
codecparsers: av1: clean the seen_frame_header in parse_tile_group().

The current seen_frame_header is not cleaned correctly. According
to the spec, it should be cleaned when tiles are parsed completely.
Also delete a verbose seen_frame_header init in reset_state().

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: fix a typo in parse_metadata_scalability
He Junyan [Tue, 29 Sep 2020 05:15:37 +0000 (13:15 +0800)]
codecparsers: av1: fix a typo in parse_metadata_scalability

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: Do not assert in identify_one_obu when check annex b size.
He Junyan [Mon, 28 Sep 2020 10:22:08 +0000 (18:22 +0800)]
codecparsers: av1: Do not assert in identify_one_obu when check annex b size.

Some buggy stream just writes the wrong temporal unit and frame size in
the stream. We should return failure rather than assert to abort.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: Add unknow AV1 profile define for saint check.
He Junyan [Tue, 22 Sep 2020 11:16:30 +0000 (19:16 +0800)]
codecparsers: av1: Add unknow AV1 profile define for saint check.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: Improve the parse_tile_info.
He Junyan [Fri, 24 Jul 2020 06:54:37 +0000 (14:54 +0800)]
codecparsers: av1: Improve the parse_tile_info.

1. store more tile info when parse tile group.
   The column, row, tile offset and tile data size are all useful for
   decoder process, especially for HW kind decoder such as VAAPI dec.
   Also fix the tile group skip size for each tile data.
2. No min_inner_tile_width requirement in newest spec.
3. Calculate the sbs of each tile for both uniform tile and non-uniformi
   tile.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: Fix a tile info read typo in frame header.
He Junyan [Tue, 28 Jul 2020 09:25:44 +0000 (17:25 +0800)]
codecparsers: av1: Fix a tile info read typo in frame header.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: Fix a typo when get value of segmentation params.
He Junyan [Tue, 25 Aug 2020 11:44:48 +0000 (19:44 +0800)]
codecparsers: av1: Fix a typo when get value of segmentation params.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: add valid check for global motion params.
He Junyan [Tue, 25 Aug 2020 08:33:26 +0000 (16:33 +0800)]
codecparsers: av1: add valid check for global motion params.

The global motion params and its matrix values need to be verified
before we use them. If it is invalid, we should notify the decoder
that it should not be used.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: uint8 range is not enough for av1_bitstreamfn_ns
He Junyan [Tue, 25 Aug 2020 07:25:56 +0000 (15:25 +0800)]
codecparsers: av1: uint8 range is not enough for av1_bitstreamfn_ns

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: delete duplicated GST_AV1_GM_ABS_ALPHA_BITS define.
He Junyan [Tue, 25 Aug 2020 07:25:06 +0000 (15:25 +0800)]
codecparsers: av1: delete duplicated GST_AV1_GM_ABS_ALPHA_BITS define.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agotest: av1parser: update the test result because of bug fixing.
He Junyan [Thu, 27 Aug 2020 13:33:14 +0000 (21:33 +0800)]
test: av1parser: update the test result because of bug fixing.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: Improve the loop filter setting.
He Junyan [Mon, 24 Aug 2020 07:29:56 +0000 (15:29 +0800)]
codecparsers: av1: Improve the loop filter setting.

1. loop_filter_ref_deltas should be int because it needs to compare
   with 0.
2. Move the loop filter init logic to setup_past_independence() and
   load_previous(), which make it more precise with the spec.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: Fix a error report for metadata obu.
He Junyan [Fri, 14 Aug 2020 06:40:49 +0000 (14:40 +0800)]
codecparsers: av1: Fix a error report for metadata obu.

The metadata OBUs, for example, ITUT_T35 has an undefined payload such
as itu_t_t35_payload_bytes field in AV1 spec, which may cause the failure
of parsing the trailings bits. We can give a warning and ignore this kind
of errors.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: Fix a level index bug in sequence.
He Junyan [Tue, 28 Jul 2020 07:06:04 +0000 (15:06 +0800)]
codecparsers: av1: Fix a level index bug in sequence.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agocodecparsers: av1: all ref idx should be gint8.
He Junyan [Fri, 24 Jul 2020 04:49:10 +0000 (12:49 +0800)]
codecparsers: av1: all ref idx should be gint8.

All the ref index need to compare with 0 in reference index decision
algorithm. We also need to init them to -1.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1464>

3 years agova: h264dec: Add support for interlaced stream
Seungha Yang [Sat, 14 Nov 2020 09:48:05 +0000 (18:48 +0900)]
va: h264dec: Add support for interlaced stream

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1812>

3 years agova: h264dec: implement gst_va_h264_dec_new_field_picture()
Víctor Manuel Jáquez Leal [Mon, 16 Nov 2020 15:29:04 +0000 (16:29 +0100)]
va: h264dec: implement gst_va_h264_dec_new_field_picture()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1812>

3 years agova: h264dec: Fix picture_height_in_mbs_minus1
Seungha Yang [Sat, 14 Nov 2020 11:46:30 +0000 (20:46 +0900)]
va: h264dec: Fix picture_height_in_mbs_minus1

Fix for interlaced stream (when sps->frame_mbs_only_flag is equal to one)

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1812>

3 years agova: h264dec: admit baseline if stream obeys A.2
Víctor Manuel Jáquez Leal [Mon, 16 Nov 2020 15:29:46 +0000 (16:29 +0100)]
va: h264dec: admit baseline if stream obeys A.2

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1812>

3 years agocodecs: h264decoder: Add support for field ref picture list modification
Seungha Yang [Sat, 14 Nov 2020 15:20:54 +0000 (00:20 +0900)]
codecs: h264decoder: Add support for field ref picture list modification

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1812>

3 years agocodecs: h264decoder: Add more option arguments for reference picture getter
Seungha Yang [Tue, 17 Nov 2020 09:39:56 +0000 (18:39 +0900)]
codecs: h264decoder: Add more option arguments for reference picture getter

In case that "pic_order_cnt_type" is equal to zero, ref picture
list for B slice should not include non-existing picture
as per spec 8.2.4.2.3. And, the second field is not needed
for the process of frame picture reference list construction
since it needs to be frame unit, not field picture in that case.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1812>

3 years agocodecs: h264decoder: Split gap picture as well if needed
Seungha Yang [Tue, 17 Nov 2020 09:59:35 +0000 (18:59 +0900)]
codecs: h264decoder: Split gap picture as well if needed

field pair pictures might be required for reference list
depending on context.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1812>

3 years agoh264dec: Fix POC calculation for type 0
Nicolas Dufresne [Thu, 5 Nov 2020 23:09:06 +0000 (18:09 -0500)]
h264dec: Fix POC calculation for type 0

This is mostly for future use as it only fixes the caclulation for interlaced
cases, the case of frame seemed correct already.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1812>

3 years agocodecs: h264decoder: Don't try to construct RefPicList0 and RefPicList1 if not required
Seungha Yang [Mon, 16 Nov 2020 18:11:46 +0000 (03:11 +0900)]
codecs: h264decoder: Don't try to construct RefPicList0 and RefPicList1 if not required

We were trying to construct reference picture list even for
I slice before this commit. Reference list is required only for
P, SP and B slices. Also, if all existing reference pictures
are gap pictures, we don't need to construct lists.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1812>

3 years agova: pool: Check the force_videometa for all memory types.
He Junyan [Mon, 2 Nov 2020 17:59:46 +0000 (01:59 +0800)]
va: pool: Check the force_videometa for all memory types.

force_videometa should mean that the buffer must use video meta to
map correctly. When the stride or the offset of the alloc_info is
different from the src caps, the downstream must use video meta.
So this flag should not link with the RAW caps only. All kinds of
caps(memory:VAMemory, memory:DMABuf) should have this flag.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1711>

3 years agova: basedec: Improve the decide_allocation().
He Junyan [Mon, 16 Nov 2020 16:18:22 +0000 (00:18 +0800)]
va: basedec: Improve the decide_allocation().

In decide_allocation(), we now just use the other_pool for frames
copy when the src caps is raw. This can make the logic a little
clear. There is no need for us to check the alignment and video
meta again.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1711>

3 years agova: basedec: fallback to system memory if downstream caps is any.
He Junyan [Mon, 16 Nov 2020 15:53:39 +0000 (23:53 +0800)]
va: basedec: fallback to system memory if downstream caps is any.

When the downstream element reports an ANY caps, and it also fails to
support VideoMeta, we should fallback to the system memory.
Note: the basetransform kind elements never return valid allocation
query before set_caps(). So, if a basetransform return an ANY sink
caps, we always fallback to system memory for it.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1711>

3 years agovah264dec: Fix for long term reference picture signalling
Seungha Yang [Sun, 15 Nov 2020 19:38:28 +0000 (04:38 +0900)]
vah264dec: Fix for long term reference picture signalling

Allocate a GArray which is used to fill
VAPictureParameterBufferH264.ReferenceFrames (called per frame),
instead of alloc/free per frame.
Also this commit is to fix the condition where long-term reference
picture is needed for VAPictureParameterBufferH264.ReferenceFrames
entry.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1813>

3 years agocodecs: h264decoder: Fix MMCO type 1 for interlaced stream
Seungha Yang [Sat, 14 Nov 2020 18:41:27 +0000 (03:41 +0900)]
codecs: h264decoder: Fix MMCO type 1 for interlaced stream

If field_pic_flag of current picture is equal to zero,
both field of reference field pair should be marked as
"unused for reference"

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1810>

3 years agocodecs: h264decoder: Fix MMCO type 3 for interlaced stream
Seungha Yang [Sat, 14 Nov 2020 17:59:24 +0000 (02:59 +0900)]
codecs: h264decoder: Fix MMCO type 3 for interlaced stream

Depending on short-ref picture corresponding to picNumX value,
there's a condition that only one field should be updated to
be non-reference picture.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1810>

3 years agocodecs: h264decoder: Split frame picture into field pictures if needed
Seungha Yang [Sat, 14 Nov 2020 15:55:09 +0000 (00:55 +0900)]
codecs: h264decoder: Split frame picture into field pictures if needed

In case of interlaced stream, frame pictures need to be splitted
into field for reference marking process.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1810>

3 years agocodecs: h264decoder: Add util macro for frame/field picture identification
Seungha Yang [Sun, 15 Nov 2020 15:27:28 +0000 (00:27 +0900)]
codecs: h264decoder: Add util macro for frame/field picture identification

Add a macro to check whether given GstH264Picture is for frame or field
decoding.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1810>

3 years agod3d11window: Prefer full color range for display target colorspace
Seungha Yang [Mon, 16 Nov 2020 11:44:06 +0000 (20:44 +0900)]
d3d11window: Prefer full color range for display target colorspace

We don't need to preserve input color range for transformed target
color space. Also some GPUs doesn't seem to be happy with 16-235
color range for RGB color space.
Also, since our default display target color space is
DXGI_COLOR_SPACE_RGB_FULL_G22_NONE_P709, choosing full color range
would make more sense.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1814>

3 years agowpe: Don't crash when running on X11.
Jan Schmidt [Fri, 14 Aug 2020 16:02:44 +0000 (02:02 +1000)]
wpe: Don't crash when running on X11.

Don't assume the available EGL display is a wayland display -
instead, check the the GStreamer GL context is EGL, and then
use gst_gl_display_egl_from_gl_display to create a
GstGLDisplayEGL from that, which also adds refcounting
around the underlying EGLDisplay.

Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/issues/1385

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1752>

3 years agova: allocator: dmabuf: log unknown surface format
Víctor Manuel Jáquez Leal [Fri, 13 Nov 2020 19:25:36 +0000 (20:25 +0100)]
va: allocator: dmabuf: log unknown surface format

It is possible that surface format is not assigned, keeping its default
GStreamer value: unknown, but gst_video_format_to_string() doesn't print
unknown format, so this patch does it manually.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1811>

3 years agova: allocator: dmabuf: destroy VASurface if no pooled buffer
Víctor Manuel Jáquez Leal [Fri, 13 Nov 2020 19:20:47 +0000 (20:20 +0100)]
va: allocator: dmabuf: destroy VASurface if no pooled buffer

When gst_va_dmabuf_allocator_setup_buffer_full() receives info (not NULL) it is
supposed that this buffer is not part of the allocator pool, so it has to be
de-allocated as soon it is freed.

This patch sets the destroy notify of the assigned GstVaBufferSurface if info is
not NULL.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1811>

3 years agovah264dec: Allow missing reference picture
Seungha Yang [Fri, 13 Nov 2020 18:20:19 +0000 (03:20 +0900)]
vah264dec: Allow missing reference picture

baseclass might provide reference picture list with null picture.
Ensure picture before filling picture information.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1809>

3 years agocodecs: h264decoder: Don't give up to decode due to missing reference picture
Seungha Yang [Fri, 13 Nov 2020 18:16:07 +0000 (03:16 +0900)]
codecs: h264decoder: Don't give up to decode due to missing reference picture

Missing reference picture is very common thing for broken/malformed stream.
Decoder should be able to keep decoding if it's not a very critical error.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1809>

3 years agompegtsdemux: Fix off by one error
Edward Hervey [Fri, 13 Nov 2020 16:50:03 +0000 (17:50 +0100)]
mpegtsdemux: Fix off by one error

Turns out timestamps of zero are valid :) Fixes issues with streams where the
PTS/DTS would be equal to the first PCR.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1807>

3 years agod3d11h264dec: Add support for interlaced stream
Seungha Yang [Thu, 5 Nov 2020 17:45:21 +0000 (02:45 +0900)]
d3d11h264dec: Add support for interlaced stream

Add support for interlaced stream.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1534>

3 years agocodecs: h264decoder: Add support for interlaced stream
Seungha Yang [Mon, 9 Nov 2020 16:28:03 +0000 (01:28 +0900)]
codecs: h264decoder: Add support for interlaced stream

Initial support for interlaced stream. Subclass should implement
new_field_picture() vfunc. Otherwise, baseclass will assume that
subclass doesn't support interlaced stream.

Restrictions:
* Reference picture modification process for interlaced stream
  is not implemented yet
* PAFF (Picture Adaptive Frame Field) is not properly implemented.
* Field display ordering (e.g., top-field-first) decision should
  be enhanced via picture timing SEI parsing
* Gap in field picture should be handled

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1534>

3 years agocodecs: h264decoder: Rename DPB methods
Seungha Yang [Wed, 4 Nov 2020 19:16:54 +0000 (04:16 +0900)]
codecs: h264decoder: Rename DPB methods

Clarify wheter it's for picture(field) or frame in order to
support interlaced stream, because DPB size is frame unit, not picture
in case of interlaced stream.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1534>

3 years agocodecs: h264decoder: Remove interlaced stream related constraints
Seungha Yang [Wed, 4 Nov 2020 18:47:35 +0000 (03:47 +0900)]
codecs: h264decoder: Remove interlaced stream related constraints

... and add new_field_picture() vfunc so that ensure interlaced
decoding support by subclass.
The method will be used later with interlaced stream support.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1534>

3 years agocodecs: h264decoder: Move to inline GstH264DecoderClass documentation
Seungha Yang [Thu, 12 Nov 2020 14:49:01 +0000 (23:49 +0900)]
codecs: h264decoder: Move to inline GstH264DecoderClass documentation

Don't duplicate documentation for class vfunc. Hotdoc doesn't seem
to be happy with duplicated documentation.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1534>

3 years agocodecs: h264decoder: Store reference picture type using enum value
Seungha Yang [Thu, 5 Nov 2020 16:45:36 +0000 (01:45 +0900)]
codecs: h264decoder: Store reference picture type using enum value

Managing reference picture type by using two variables
(ref and long_term) seems to be redundant and that can be
represented by using a single enum value.

This is to sync this implementation with gstreamer-vaapi so that
make comparison between this and gstreamer-vaapi easier and also
in order to minimize the change required for subclass to be able
to support interlaced.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1534>

3 years agocodecs: h264decoder: Minor documentation fix
Seungha Yang [Tue, 10 Nov 2020 16:56:52 +0000 (01:56 +0900)]
codecs: h264decoder: Minor documentation fix

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1534>

3 years agocodecs: h264decoder: Remove DPB size related spammy debug message
Seungha Yang [Fri, 13 Nov 2020 14:18:20 +0000 (23:18 +0900)]
codecs: h264decoder: Remove DPB size related spammy debug message

It's not informative at all if SPS wasn't updated. Also we are printing
DPB size related debug message in another place already.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1806>

3 years agoh264parse: try harder to update timecode
Mathieu Duponchelle [Thu, 12 Nov 2020 21:27:08 +0000 (22:27 +0100)]
h264parse: try harder to update timecode

NumClockTS is the maximum number of timecodes the pic_timing SEI
can carry, but it is perfectly OK for it to carry fewer, and have
one of the clock_timestamp_flags set to 0.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1804>

3 years agoh264parse: fix installing of update-timecode property
Mathieu Duponchelle [Thu, 12 Nov 2020 21:32:00 +0000 (22:32 +0100)]
h264parse: fix installing of update-timecode property

Simply fixes a typo that did not have any adverse effect,
and avoid hardcoding initializer

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1805>

3 years agocodecs: h264decoder: Don't fill gap picture if it's not allowed
Seungha Yang [Thu, 12 Nov 2020 10:43:22 +0000 (19:43 +0900)]
codecs: h264decoder: Don't fill gap picture if it's not allowed

We should fill gap picture only if sps->gaps_in_frame_num_value_allowed_flag
is set.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1801>

3 years agoopenjpegenc: store stripe offset when encoding image
Aaron Boxer [Thu, 16 Apr 2020 14:06:29 +0000 (10:06 -0400)]
openjpegenc: store stripe offset when encoding image

The decoder can simply read this offset after decoding
to know where to blit the stripe to the full frame

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1800>

3 years agoopenjpegenc: take subsampling into account when calculating stripe height
Aaron Boxer [Tue, 24 Mar 2020 13:15:30 +0000 (09:15 -0400)]
openjpegenc: take subsampling into account when calculating stripe height

We calculate minimum of (stripe height * sub sampling) across all components
to ensure that all component dimensions are consistent with sub-sampling.
The last stripe for each component is simply the remaining height.

limit wavelet resolutions for "thin" stripes

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1800>

3 years agoopenjpegenc: fix memory leak from mstream
Stéphane Cerveau [Thu, 12 Mar 2020 12:41:40 +0000 (13:41 +0100)]
openjpegenc: fix memory leak from mstream

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1800>

3 years agoopenjpegenc: fail negotation in handle_frame if alignment mismatch
Aaron Boxer [Mon, 13 Jan 2020 19:00:38 +0000 (14:00 -0500)]
openjpegenc: fail negotation in handle_frame if alignment mismatch

If encoder is in stripe mode, then downstream must also support stripe

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1800>

3 years agonvcodec: Fix various typo
Seungha Yang [Thu, 12 Nov 2020 12:46:59 +0000 (21:46 +0900)]
nvcodec: Fix various typo

Not sure where the DECOCER came from

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1803>

3 years agova: comments to explain code
Víctor Manuel Jáquez Leal [Thu, 12 Nov 2020 12:33:26 +0000 (13:33 +0100)]
va: comments to explain code

There are a couple part where code seems, at least to me, a bit oscure or
confusing. So let's better add an explanation.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1802>

3 years agoadaptivedemux: Don't calculate bitrate for header/index fragments
Edward Hervey [Tue, 10 Nov 2020 13:48:28 +0000 (14:48 +0100)]
adaptivedemux: Don't calculate bitrate for header/index fragments

They are generally substantially smaller than regular fragments, and therefore
we end up pushing totally wrong bitrates downstream.

Fixes erratic buffering issues with DASH introduced by
66f5e874352016e29f555e3ce693b23474e476db

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1021>

3 years agoadaptivedemux: Store QoS values on the element
Edward Hervey [Mon, 9 Nov 2020 10:41:10 +0000 (11:41 +0100)]
adaptivedemux: Store QoS values on the element

Storing it per-stream requires taking the manifest lock which can apparenly be
hold for aeons. And since the QoS event comes from the video rendering thread
we *really* do not want to do that.

Storing it as-is in the element is fine, the important part is knowing the
earliest time downstream.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1021>