platform/upstream/gstreamer.git
3 years agopbutils: Expose functions for getting a file extension for caps and flags for describ...
Sebastian Dröge [Thu, 1 Jul 2021 09:41:11 +0000 (12:41 +0300)]
pbutils: Expose functions for getting a file extension for caps and flags for describing the format of the caps

This information was available internally already but not available from
the outside.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1221>

3 years agoplaybin/uridecodebin: Emit source-setup signal early before doing the scheduling...
Sebastian Dröge [Wed, 4 Aug 2021 07:06:02 +0000 (10:06 +0300)]
playbin/uridecodebin: Emit source-setup signal early before doing the scheduling query

Some elements will require the source to be set up properly before the
scheduling query returns useful results, e.g. appsrc and giostreamsrc.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1241>

3 years agoext: alsa: Fix fallback paths for setting buffer and period times
Devarsh Thakkar [Thu, 10 Jun 2021 07:55:23 +0000 (00:55 -0700)]
ext: alsa: Fix fallback paths for setting buffer and period times

Below fallback paths were introduced in
https://github.com/GStreamer/gst-plugins-base/commit/9759810d8206b5f1aa199f98599caec3630a1813
if setting period time after buffer time failed :
1) Set period time and then buffer time if it doesn't work
2) Set only buffer time
3) Set only period time

These all were not functioning properly since they were using old
copy of snd_pcm_hw_params_t which already had some fields set
as per previous try and this was causing issues as driver was
referring to that old value while trying to set them again in
fallback paths.

So now we always use the initial copy of snd_pcm_hw_params_t
for every fallback  and same is also being done at
https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/commit/557c4295107dc7374c850b0bd5331dd35e8fdd0f

Also we change the sequence to set period time earlier than
buffer time since period bytes being the smaller unit, most of the times
if underlying alsa device has a dependency then it is of period bytes
to be a multiple of some value (as per underlying DMA constraint)
and rest of the parameters like buffer bytes need to be adjusted
as per period bytes.

The same sequence is also followed in alsa-utils at
https://github.com/alsa-project/alsa-utils/commit/9b621eeac4d55c4e881f093be5b163ca07d01b63

Fix 2) and 3) scenarios by returning success if the exclusive setting is passed
and not doing any further setting for buffer time or period time.

Add new fallback path of not setting any buffer time and period time
if all above fallback paths fail. The same is also being
followed at aforementioned pulseaudio commit.

In case of alsasink, remove the retry goto label, since it is not
required anymore as fallback paths take care of setting default
values if driver is not accepting any of the fallback paths.

Use separate label for exit to free params structs and return err
code. This also fixes leak in no_rate goto path in alsasink

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1212>

3 years agovideoencoder: pass upstream HDR information through codec state
Jakub Adam [Tue, 25 May 2021 19:16:48 +0000 (21:16 +0200)]
videoencoder: pass upstream HDR information through codec state

Don't copy HDR metadata from sink pad, because its caps may not have
been set yet if GstVideoEncoder::negotiate is called from
GstVideoEncoder::set_format, as e.g. vpx encoder does.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1175>

3 years agovideoutils: add HDR metadata fields to GstVideoCodecState
Jakub Adam [Tue, 25 May 2021 19:15:53 +0000 (21:15 +0200)]
videoutils: add HDR metadata fields to GstVideoCodecState

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1175>

3 years agovideo-overlay-composition: Allow empty overlay compositions
Sebastian Dröge [Mon, 16 Aug 2021 07:19:07 +0000 (10:19 +0300)]
video-overlay-composition: Allow empty overlay compositions

Allowing to pass NULL to the constructor removes the need to
special-case the first rectangle in calling code and generally
simplifies application code.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1256>

3 years agogstrtspconnection: Add support to ignore x-server header reply
Per Förlin [Tue, 1 Jun 2021 13:27:31 +0000 (15:27 +0200)]
gstrtspconnection: Add support to ignore x-server header reply

When connecting to an RTSP server in tunnled mode (HTTP) the server
usually replies with a x-server header. This contains the address
of the intended streaming server. However some servers return an
"invalid" address. Here follows two examples when it might happen.

1. A server use Apache combined with a separate RTSP process to handle
   Https request on port 443. In this case Apache handle TLS and
   connects to the local RTSP server, which results in a local
   address 127.0.0.1 or ::1 in the x-server reply. This address is
   returned to the actual RTSP client in the x-server header.
   The client will receive this address and try to  connect to it
   and fail.

2. The client use a ipv6 link local address with a specified scope id
   fe80::aaaa:bbbb:cccc:dddd%eth0 and connects via Http on port 80.
   The RTSP server receives the connection and returns the address
   in the x-server header. The client will receive this address and
   try to connect to it "as is" without the scope id and fail.

In the case of streaming data from RTSP servers like 1. and 2. it's
useful to have the option to simply ignore the x-server header reply
and continue using the original address.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1192>

3 years agosdp: Avoid using g_memdup() since it is deprecated
Nirbheek Chauhan [Fri, 13 Aug 2021 14:05:23 +0000 (19:35 +0530)]
sdp: Avoid using g_memdup() since it is deprecated

g_memdup() is deprecated since GLib 2.68 and we want to avoid
deprecation warnings with recent versions of GLib. Instead of using
g_memdup2(), we can simply use the new gst_buffer_new_memdup() added
in 1.19.x

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1254>

3 years agoglbuffer: support persistent buffer mappings
Matthew Waters [Fri, 4 Jun 2021 08:32:07 +0000 (18:32 +1000)]
glbuffer: support persistent buffer mappings

Requires OpenGL 4.4 or EXT_buffer_storage

Current mesa exposes GL_ARB_buffer_storage when retrieving the relevant
functions returns no-ops and causes failures.

Improves throughput of uploads by roughly 30%-60% and download throughput by
roughly 10-30% across depending on the exact scenario and hardware.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1191>

3 years agoexamples: win32-videooverlay: Add support for testing gst_video_overlay_set_render_re...
Seungha Yang [Fri, 30 Jul 2021 14:57:20 +0000 (23:57 +0900)]
examples: win32-videooverlay: Add support for testing gst_video_overlay_set_render_rectangle

Add keyboard handler to test gst_video_overlay_set_render_rectangle()
API for Windows video elements

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1235>

3 years agoexamples: win32-videooverlay: Use d3d11videosink by default
Seungha Yang [Fri, 30 Jul 2021 14:04:57 +0000 (23:04 +0900)]
examples: win32-videooverlay: Use d3d11videosink by default

d3d11videosink was promoted to have primary rank and
it's recommended videosink element on Windows

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1235>

3 years agotests: appsink: Add reverse stepping test case
Seungha Yang [Thu, 8 Jul 2021 08:47:28 +0000 (17:47 +0900)]
tests: appsink: Add reverse stepping test case

To demonstrate reverse stepping issue of
https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/848

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1223>

3 years agoexamples: win32-videooverlay: Add test option for threading scenario
Seungha Yang [Thu, 3 Jun 2021 10:15:22 +0000 (19:15 +0900)]
examples: win32-videooverlay: Add test option for threading scenario

Add an option to test the case where window thread and pipeline handling
thread are different. Mainly to test the HWND leak fixed by
https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2302

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1188>

3 years agortpbasedepayload: remove object locking an extension
Matthew Waters [Tue, 27 Jul 2021 07:44:02 +0000 (17:44 +1000)]
rtpbasedepayload: remove object locking an extension

Doing that is fraught with danger of deadlocks and is not conceptually
part of the API contract.  The object lock is generally intended for
internal-object-use only.

If another lock is needed, that should be added separately.

This lock was erronously added as part of:
https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1118

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1233>

3 years agogldownload: use the GstGLSyncMeta in all cases
Matthew Waters [Tue, 27 Jul 2021 03:30:56 +0000 (13:30 +1000)]
gldownload: use the GstGLSyncMeta in all cases

fixes qmlglsrc ! gldownload ! videoconvert in some cases.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1232>

3 years agoglcontext: egl: Stop comparing native surface pointer
Nicolas Dufresne [Thu, 22 Jul 2021 21:11:26 +0000 (17:11 -0400)]
glcontext: egl: Stop comparing native surface pointer

This was noticed with wayland, sometimes the newly create native
handle can have the same pointer (even though its new). This lead
to unwanted errors or crash.

Fixes #927

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1231>

3 years agoglwindow: Add "window-handle-changed" signal
Nicolas Dufresne [Thu, 22 Jul 2021 21:02:51 +0000 (17:02 -0400)]
glwindow: Add "window-handle-changed" signal

This allow other objects to clear any wrapper object that depends
on the previous handle, and properly re-create the new wrappers without
having to resort into doing pointer comparison.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1231>

3 years agoRevert "glwindow: wayland: Skip redoing surfaces if window haven't changed"
Nicolas Dufresne [Fri, 23 Jul 2021 17:39:34 +0000 (13:39 -0400)]
Revert "glwindow: wayland: Skip redoing surfaces if window haven't changed"

This reverts commit aba6bd7822f4c0f572765bfaada76f454a594317.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1231>

3 years agoappsrc: serialize custom events with buffers flow
Guillaume Desmottes [Mon, 22 Feb 2021 12:17:18 +0000 (13:17 +0100)]
appsrc: serialize custom events with buffers flow

Application may want to inject events to the pipeline and keep them
synchronized with the buffers flow.

Fix #247

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1046>

3 years agoappsink: add API to catch events
Guillaume Desmottes [Fri, 19 Feb 2021 12:32:48 +0000 (13:32 +0100)]
appsink: add API to catch events

There is currently no way for users to receive incoming events from
appsink while keeping them properly serialized with the buffers flow.
This can be especially useful when application is injecting custom
downstream events into the pipeline and needs to know when they reached
appsink.

Solving this by adding a new signal notifying about new incoming events
and a set of action signals and method to pull those events.
The API is actually pulling the samples and events all together as they
are actually fetched from the same queue.
Having a specific API to pull only events would have the side effect of
discarding samples (and pulling samples would discard events) making
this API not convenient for users.

Partially fix #247

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1046>

3 years agoappsink: factor out dequeue_object()
Guillaume Desmottes [Fri, 19 Feb 2021 13:45:08 +0000 (14:45 +0100)]
appsink: factor out dequeue_object()

No semantic change, will be used to implement new event API.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1046>

3 years agoglwindow: wayland: Skip redoing surfaces if window haven't changed
Nicolas Dufresne [Tue, 20 Jul 2021 13:37:58 +0000 (09:37 -0400)]
glwindow: wayland: Skip redoing surfaces if window haven't changed

The problem is that EGLNativeWindowSurface and wl_egl_surface are the
same object underneath, so we must recreate both together. As an
optimization, the EGLNativeWindowSurface wrapper is only re-created
if the window_handle changed.

On Mesa, this would cause crash, which will be fixed by:
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11979

And will lead to proper errors in the future or on other GL stack. This
issue was encounter using a permanent GstGLDisplay after cycling one of
multiple independent pipelines through NULL state.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1230>

3 years agoglwindow: wayland: Remove redundant create_surfaces call
Nicolas Dufresne [Tue, 20 Jul 2021 13:36:22 +0000 (09:36 -0400)]
glwindow: wayland: Remove redundant create_surfaces call

The surfaces will be created in _roundtrip_async, so no need to call
this early. This should cause no functional difference.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1230>

3 years agodecodebin3: improve decoder selection
Michael Olbrich [Fri, 11 Jun 2021 07:02:29 +0000 (09:02 +0200)]
decodebin3: improve decoder selection

Currently the decoder selection is very naive: The type with the highest
rank that matches the current caps is used. This works well for software
decoders. The exact supported caps are always known and the static caps are
defined accordingly.
With hardware decoders, e.g. vaapi, the situation is different. The decoder
may reject the caps later during a caps query. At that point, a new decoder
is created. However, the same type is chosen an after several tries,
decodebin fails.

To avoid this, do the caps query while adding the decoder and try again
with other decoder types if the query fails:

1. create the decoder from the next matching type
2. add and link the decoder
3. change the decoder state to READY
4. do the caps query
   if it fails then remove the decoder again and go back to 1.
5. expose the source pad
6. sync the decoder state with the parent.

This way, the decoder is already part of the pipeline when the state change
to READY happens. So context handling should work as before.

Exposing the source pad after the query was successful is important:
Otherwise the thread from the decoder source pad may block in a blocked pad
downstream in the playsink waiting for other pads to be ready.
The thread now blocks trying to set the state back to NULL while holding
the SELECTION_LOCK. Other streams may block on the SELECTION_LOCK and the
playsink never unblocks the pad. The result is a deadlock.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1201>

3 years agogl: x11: Issue XSync to close our top level window
Nicolas Dufresne [Wed, 14 Jul 2021 20:09:41 +0000 (16:09 -0400)]
gl: x11: Issue XSync to close our top level window

This is similar action as when the window handle is modified, we now issue
XSync whenever we destroy our internal window. This ensure that the window is
properly closed before the connecgtion is dropped.

Fixes #815

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1226>

3 years agogl: wayland: Fix hinding the window on close()
Nicolas Dufresne [Wed, 14 Jul 2021 15:43:10 +0000 (11:43 -0400)]
gl: wayland: Fix hinding the window on close()

When the window is called, we properly destroy all surfaces, which effectively
will unmap that surface and should make it disapear on screen, but we also
destroy the wl_source, a GSource that is resposibble of dispatching and executing
messages to/from the Wayland server.

As a side effect, the server never gets the message and the surfaces are
"leaked" on the server. We fix this using wl_display_flush() before destroying
the wl_source.

Fixes #815

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1226>

3 years agotests: example: Add missing glx_dep when building sdlshare
Nicolas Dufresne [Thu, 15 Jul 2021 15:09:35 +0000 (11:09 -0400)]
tests: example: Add missing glx_dep when building sdlshare

Might be realted to some recent Mesa cleanup, but GLX is not longer visible
through libOpenGL, so add the missing deps now.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1227>

3 years agoexamples/qt/textureshare: add explicit dep on glx_dep
Matthew Waters [Mon, 17 May 2021 02:04:50 +0000 (12:04 +1000)]
examples/qt/textureshare: add explicit dep on glx_dep

Fixes linking:

/usr/bin/ld: subprojects/gst-plugins-base/tests/examples/gl/qt/qglwtextureshare/qglwtextureshare.p/qglrenderer.cpp.o: undefined reference to symbol 'glXGetCurrentContext'
/usr/bin/ld: /usr/lib64/libGLX.so.0: error adding symbols: DSO missing from command line

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1143>

3 years agogl: Try GLVND 'opengl' and 'glx' first
Haelwenn (lanodan) Monnier [Fri, 14 May 2021 12:10:55 +0000 (14:10 +0200)]
gl: Try GLVND 'opengl' and 'glx' first

This fixes targetting desktop OpenGL without libGL.so

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1143>

3 years agoaudioaggregator: Only post QoS messages if the property is enabled
Sebastian Dröge [Mon, 12 Jul 2021 06:37:24 +0000 (09:37 +0300)]
audioaggregator: Only post QoS messages if the property is enabled

Previously one of the branches did not check for the property value. To
avoid this in the future, check inside the QoS calculation function
instead.

As a side effect this now always prints the debug messages into the logs
when samples are dropped, which is useful information even without the
QoS messages.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1224>

3 years agoaudioaggregator: Resync on the next buffer when dropping a buffer on discont resyncing
Sebastian Dröge [Fri, 9 Jul 2021 06:49:15 +0000 (09:49 +0300)]
audioaggregator: Resync on the next buffer when dropping a buffer on discont resyncing

If a buffer is dropped during resyncing on a discont because either its
end offset is already before the current output offset of the
aggregator or because it fully overlaps with the part of the current
output buffer that was already filled, then don't just assume that the
next buffer is going to start at exactly the expected offset. It might
still require some more dropping of samples.

This caused the input to be mixed with an offset to its actual position
in the output stream, causing additional latency and wrong
synchronization between the different input streams.

Instead consider each buffer after a discont as a discont until the
aggregator actually resynced and starts mixing samples from the input
again.

Also update the start output offset of a new input buffer if samples
have to be dropped at the beginning. Otherwise it might be mixed too
early into the output and overwrite part of the output buffer that
already took samples from this input into account.

Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/issues/912
which is a regression introduced by https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1180/

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1224>

3 years agoaudiomixer: Add test for QoS message posting
Olivier Crête [Wed, 26 May 2021 22:20:02 +0000 (18:20 -0400)]
audiomixer: Add test for QoS message posting

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1209>

3 years agoaudio aggregator: Post QoS message when dropping audio
Olivier Crête [Wed, 26 May 2021 14:38:18 +0000 (10:38 -0400)]
audio aggregator: Post QoS message when dropping audio

Post a QoS message every time some audio samples are dropped.
Also print log messages to make it easier to debug

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1209>

3 years agoaudio aggregator: Count samples that are dropped or processed
Olivier Crête [Tue, 25 May 2021 22:05:05 +0000 (18:05 -0400)]
audio aggregator: Count samples that are dropped or processed

Keep a count of samples that are dropped or processed as statistics

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1209>

3 years agoaudio aggregator: Add QoS property to pad
Olivier Crête [Fri, 21 May 2021 20:16:50 +0000 (16:16 -0400)]
audio aggregator: Add QoS property to pad

Add a property to emit a QoS message whenever any data is dropped.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1209>

3 years agoaudio aggregator: Rename property enum to match class name
Olivier Crête [Fri, 21 May 2021 20:10:06 +0000 (16:10 -0400)]
audio aggregator: Rename property enum to match class name

Add "CONVERT" into the property enum as we're going to add an
enum specifically for the base pad.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1209>

3 years agovideodecoder: add API to receive subframes
Stéphane Cerveau [Tue, 24 Sep 2019 15:14:10 +0000 (17:14 +0200)]
videodecoder: add API to receive subframes

A video decoder can now receive subframes and start decoding
instead of waiting for the full frame to be complete.
Subframe support will reduce latency as described in the
video encoder base class.

A unit test illustrating this API is available in
tests/check/libs/videodecoder.c.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/454>

3 years agogl/context/wgl: Add missing NULL init
Seungha Yang [Sat, 3 Jul 2021 10:36:06 +0000 (19:36 +0900)]
gl/context/wgl: Add missing NULL init

The value of uninitialized local variable is varying depending
on compiler and not guaranteed to be NULL initialized.
That results in pointing random address instead of expected function pointer.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1222>

3 years agovideotestsrc: Add SMPTE75 RP-219 color bars conformant
Francisco Javier Velázquez-García [Fri, 12 Mar 2021 12:55:38 +0000 (13:55 +0100)]
videotestsrc: Add SMPTE75 RP-219 color bars conformant

Implement 8-bit values of SMPTE RP 2019-1:2014.  The bar widths and
heights are the result of fractions as integers.  The remainders of
widths are distributed in a way that they match the values in Table
C.1 (a) in the specification.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1063>

3 years agovideotestsrc: Add a start parameter to _blend_line
Jan Alexander Steffens (heftig) [Fri, 12 Mar 2021 19:58:40 +0000 (20:58 +0100)]
videotestsrc: Add a start parameter to _blend_line

Makes it easier to paint part of a line.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1063>

3 years agovideotestsrc: Keep tmpline unchanged in_convert_tmpline
Jan Alexander Steffens (heftig) [Fri, 12 Mar 2021 19:57:13 +0000 (20:57 +0100)]
videotestsrc: Keep tmpline unchanged in_convert_tmpline

This will allow us to repeatedly
call it to render subsequent lines.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1063>

3 years agoappsink: fix incorrect return nullability
Michael de Gans [Tue, 15 Jun 2021 20:22:55 +0000 (13:22 -0700)]
appsink: fix incorrect return nullability

This commit fixes the annoations for return nullability on several
GstAppSink functions. This was causing bindings to be generated
incorrectly.

Fixes #914

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1203>

3 years agocompositor: Add scaling policy to support PAR-aware scaling
Seungha Yang [Wed, 19 May 2021 07:22:46 +0000 (16:22 +0900)]
compositor: Add scaling policy to support PAR-aware scaling

Adding "sizing-policy" property for user to be able to specify
scaling policy (aspect-ratio for example).
At the moment, supported mode is only keep-aspect-ratio, but we might
be able to add more policies such as cropping, etc.

Fixes: https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/issues/696
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1156>

3 years agovideo: Deprecate gst_video_sink_center_rect()
Seungha Yang [Wed, 19 May 2021 11:11:15 +0000 (20:11 +0900)]
video: Deprecate gst_video_sink_center_rect()

... and add gst_video_center_rect() method as a replacement.
The method is useful for outside of videosink subclasses as well
but the old naming might be able to mislead people.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1156>

3 years agortpbasepayload: don't write empty extension header
Jakub Adam [Mon, 24 May 2021 17:11:51 +0000 (19:11 +0200)]
rtpbasepayload: don't write empty extension header

When some header extensions are present but none decides to write any
data to the currently processed RTP buffer, remove the extension data
section.

Resulting RTP buffer wasn't formatted correctly.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1173>

3 years agortpbuffer: Add gst_rtp_buffer_remove_extension_data()
Jakub Adam [Mon, 24 May 2021 17:02:42 +0000 (19:02 +0200)]
rtpbuffer: Add gst_rtp_buffer_remove_extension_data()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1173>

3 years agortpbasepayload: map RTP buffer READWRITE when setting headers
Jakub Adam [Mon, 24 May 2021 17:01:24 +0000 (19:01 +0200)]
rtpbasepayload: map RTP buffer READWRITE when setting headers

GstRTPHeaderExtension::write can map the RTP buffer for reading. If that
happens on a buffer that is already mapped WRITE-only by the payloader,
the payloader's mapping gets invalidated (GstRTPBuffer::map will point
to a different instance of GstMemory).

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1173>

3 years agortphdrext: Make all fields private
Olivier Crête [Thu, 24 Jun 2021 18:56:11 +0000 (14:56 -0400)]
rtphdrext: Make all fields private

The presence of a method and a field with the same name confuses the C#
binding generator. As there are accessor functions for all the fields,
let's just make them private.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1218>

3 years agogst: don't use volatile to mean atomic
Olivier Crête [Sat, 26 Jun 2021 16:50:58 +0000 (12:50 -0400)]
gst: don't use volatile to mean atomic

volatile is not sufficient to provide atomic guarantees and real atomics
should be used instead.  GCC 11 has started warning about using volatile
with atomic operations.

https://gitlab.gnome.org/GNOME/glib/-/merge_requests/1719

Discovered in gst-plugins-good#868

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1219>

3 years agovideo-converter: Set up matrix tables only once.
Jan Schmidt [Fri, 25 Jun 2021 13:42:34 +0000 (23:42 +1000)]
video-converter: Set up matrix tables only once.

When configuring a multi-thread converter, only allocate the
shared colour conversion matrices once for the first thread,
to avoid allocating multiple times and leaking memory.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1216>

3 years agovideo-converter: Set up gamma tables only once
Jan Alexander Steffens (heftig) [Thu, 6 May 2021 17:01:41 +0000 (19:01 +0200)]
video-converter: Set up gamma tables only once

When the video converter is using multiple threads, the gamma tables
were created multiple times, leaking the tables set up for the previous
thread.

Only calculate the tables once.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1140>

3 years agoaudio-converter: Free config when gst_audio_converter_new fails
Jan Alexander Steffens (heftig) [Thu, 6 May 2021 16:22:45 +0000 (18:22 +0200)]
audio-converter: Free config when gst_audio_converter_new fails

The config got leaked when parameter validation fails.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1140>

3 years agoglprototypes: Add GST_GL_API_OPENGL to available version of sync
Seungha Yang [Fri, 25 Jun 2021 06:24:21 +0000 (15:24 +0900)]
glprototypes: Add GST_GL_API_OPENGL to available version of sync

Make sync APIs usable if supported, even when GST_GL_API_OPENGL3 is
not selected

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1215>

3 years agogstrtspconnection: Add IPv6 support for tunneled mode
Per Förlin [Fri, 23 Apr 2021 16:03:20 +0000 (18:03 +0200)]
gstrtspconnection: Add IPv6 support for tunneled mode

An IPv6 address must be specified within [] brackets.
Add brackets for IPv6 address used for tunneled mode,
for non-tunneled this is already supported.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1145>

3 years agovideodecoder: Call drain() rather then finish() on segment-done
Nicolas Dufresne [Wed, 16 Jun 2021 18:49:14 +0000 (14:49 -0400)]
videodecoder: Call drain() rather then finish() on segment-done

The finish() virtual function documentation state that "Sub-classes can refuse
to decode new data after." Though, it is very common to issue a non-flushing
seek after that event in gapless playback uses case. This fixes potential
stalls with code using segment seeks, by using drain() virtual funciton
instead.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1206>

3 years agooggdemux: fix a race in push mode when performing the duration seek
Matthew Waters [Tue, 8 Jun 2021 04:55:36 +0000 (14:55 +1000)]
oggdemux: fix a race in push mode when performing the duration seek

There may be two or more threads involved here however the important
interaction is the use of ogg->seeK_event_drop_till value that was only
set in the push-mode seek-event thread and could race with upstream
sending e.g. and EOS (or data).

Scenario is this:
1. oggdemux performs a seek to near the end of the file to try and find
   the duration. ogg->push_state is set to PUSH_DURATION.
2. Seek is picked up by the dedicated seek event thread and sets
   ogg->seek_event_drop_till to the seek event's seqnum.
3. Most operations are blocked or dropped waiting on the duration to
   be determined and processing continues until a duration is found.
4. Two branching options for how this ultimately plays out
4a. The source is too fast and we receive an EOS event which is dropped
    because ogg->push_state == PUSH_DURATION.  In this case everything
    works.
4b. We hit our 'almost at the end' check in
    gst_ogg_pad_handle_push_mode_state() and attempt to seek back to the
    beginning (or to a user-provided seek).  This seek is marshalled to
    the seek event thread without setting ogg->seek_event_drop_till but
    with change ogg->push_state = PUSH_PLAYING.  If an EOS event or
    e.g. buffers arrive from upstream before the seek event thread has
    picked up the seek event, then the EOS/data is processed as if it
    came as a result of the seek event.  This is the case that fails.

The fix is two-fold:
1. Preemptively set ogg->seek_event_drop_till when setting the seek
   event so that data and other events can be dropped correctly.
2. In addition to dropping and EOS events while ogg->push_state ==
   PUSH_DURATION, also drop any EOS events that are received before the
   seek event has been processed by also tracking the seqnum of the seek.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1196>

3 years agoaudiobasesink: Fix of double lock release
Sergei Kovalev [Mon, 21 Jun 2021 14:06:14 +0000 (14:06 +0000)]
audiobasesink: Fix of double lock release

Add missing "return;" which prevents double lock release.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1208>

3 years agoUpdate COPYING
Corentin Damman [Mon, 21 Jun 2021 08:34:07 +0000 (08:34 +0000)]
Update COPYING

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1207>

3 years agogl/context: fill a GError on platform-specific fill_info() error
Matthew Waters [Wed, 16 Jun 2021 05:55:49 +0000 (15:55 +1000)]
gl/context: fill a GError on platform-specific fill_info() error

Fixes bindings assuming that GError is always set on error:
https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/809#note_957493
https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/809#note_957494
https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/809#note_957498

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1204>

3 years agoCheck mandatory ClockTime arguments
François Laignel [Wed, 16 Jun 2021 10:13:21 +0000 (12:13 +0200)]
Check mandatory ClockTime arguments

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1205>

3 years agogl/wayland: Use consistent wl_display when creating work queue for proxy wrapper
Scott Moreau [Fri, 11 Jun 2021 15:27:46 +0000 (09:27 -0600)]
gl/wayland: Use consistent wl_display when creating work queue for proxy wrapper

Without this, glimagesink since wayland 727c7903 fails with
gst-launch-1.0: ../src/wayland-client.c:2181: wl_proxy_set_queue:
Assertion 'proxy->display == queue->display' failed.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1200>

3 years agogl/viv-fb: fix gl plugin hang when run with viv-fb backend
Haihua Hu [Mon, 7 Jun 2021 09:54:46 +0000 (17:54 +0800)]
gl/viv-fb: fix gl plugin hang when run with viv-fb backend

below commit change the window resize thread and cause viv-fb backend
hang, need move resize code after window->open is called. Otherwise,
the resize message will send to a thread that not start running and
window resize call will waiting forever.

Commit:        b887db1efe816c0c28b60a6842fa9005a26c1502
glwindow: fix racy resize updates

Take locks around resize handling and marshall all resizes to the
windowing thread by default.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1195>

3 years agoeglimage: Add AV12 DMABuf import support
Nicolas Dufresne [Thu, 10 Jun 2021 16:41:31 +0000 (12:41 -0400)]
eglimage: Add AV12 DMABuf import support

The per plane format mapping for AV12 was missing, which would force
raw upload.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1199>

3 years agoeglimage: Fix wrong stride when importing DMABuf
Nicolas Dufresne [Thu, 10 Jun 2021 16:39:34 +0000 (12:39 -0400)]
eglimage: Fix wrong stride when importing DMABuf

When the code was ported to use component index instead of plane
index, the call to GST_VIDEO_INFO_PLANE_STRIDE() was accidently ported
to use component index, but this macro takes a plane index.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1199>

3 years agoappsrc: When dropping buffers before handling the initial segment use the latest...
Sebastian Dröge [Tue, 8 Jun 2021 14:27:55 +0000 (17:27 +0300)]
appsrc: When dropping buffers before handling the initial segment use the latest input segment

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1198>

3 years agoappsrc: Correctly check if this is the first buffer that was queued
Sebastian Dröge [Tue, 8 Jun 2021 10:05:46 +0000 (13:05 +0300)]
appsrc: Correctly check if this is the first buffer that was queued

By checking the queue length this would also count events and caps. We
already keep track of the number of buffers separately so just use that.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1197>

3 years agoappsrc: signal enough-data even when leaking
Mathieu Duponchelle [Sat, 5 Jun 2021 00:35:30 +0000 (02:35 +0200)]
appsrc: signal enough-data even when leaking

this is convenient for application that wish to monitor whether
the appsrc is leaking.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1193>

3 years agortp/header: Add missing `array length` annotation to read/write methods
Marijn Suijten [Thu, 3 Jun 2021 19:14:42 +0000 (21:14 +0200)]
rtp/header: Add missing `array length` annotation to read/write methods

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1190>

3 years agovideo: Sort video formats correctly
Sebastian Dröge [Thu, 3 Jun 2021 10:40:33 +0000 (13:40 +0300)]
video: Sort video formats correctly

AV12 should be right after A420 because it is the same format with just
one plane less, instead of being next to I420/NV12 which don't have an
alpha channel.

RGBP should be before GBR because it's the same format except for the
more canonical component order.

See https://gitlab.freedesktop.org/gstreamer/gstreamer-rs/-/merge_requests/790
which actually checks on the CI if the algorithm defined in
video-format.h is implemented correctly.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1189>

3 years agoglvideomixer: hold extra pad ref while calling parent
Matthew Waters [Thu, 3 Jun 2021 05:57:39 +0000 (15:57 +1000)]
glvideomixer: hold extra pad ref while calling parent

Our subsequent cleanup needs a ref on the pad and calling the parent may
release the last reference and could cause a use-after-free.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1187>

3 years agogl/stereo: fix a coupld of caps leaks
Matthew Waters [Thu, 3 Jun 2021 05:57:17 +0000 (15:57 +1000)]
gl/stereo: fix a coupld of caps leaks

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1187>

3 years agogl/display: remove choosing egl fallback from GST_GL_PLATFORM
Matthew Waters [Sat, 22 May 2021 05:42:17 +0000 (15:42 +1000)]
gl/display: remove choosing egl fallback from GST_GL_PLATFORM

If GST_GL_WINDOW is unset but GST_GL_PLATFORM=egl, then we were choosing
to create an GstGLDisplayEGL directly instead of going through the any
more specific windowing system implementation (X11, Wayland).

The 'create an GstGLDisplayEGL when GST_GL_PLATFORM=egl' was a fallback
as we did not have entries for all EGL-using window systems previously.
Now that we do, the fallback can be removed.  An EGLDisplay can still
be created by setting GST_GL_WINDOW=egl or as one option.

Fixup of https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1154

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1169>

3 years agoUse g_memdup2() where available and add fallback for older GLib versions
Tim-Philipp Müller [Sun, 23 May 2021 22:51:27 +0000 (23:51 +0100)]
Use g_memdup2() where available and add fallback for older GLib versions

g_memdup() is deprecated since GLib 2.68 and we want to avoid
deprecation warnings with recent versions of GLib.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1171>

3 years agoBack to development
Tim-Philipp Müller [Tue, 1 Jun 2021 14:28:24 +0000 (15:28 +0100)]
Back to development

3 years agoRelease 1.19.1
Tim-Philipp Müller [Mon, 31 May 2021 23:09:54 +0000 (00:09 +0100)]
Release 1.19.1

3 years agoglcontext: Ask for an alpha channel and fallback
Nicolas Dufresne [Wed, 26 May 2021 20:57:28 +0000 (16:57 -0400)]
glcontext: Ask for an alpha channel and fallback

While this was already possible through the GLContext machinary, this simply
request an alpha channel by default and fallback if this is not possible. This
obsolete some RPi Dispmanx hack, since this is near equivalent will allow see
through when playgin WebM Alpha or other transparent files.

Application are still free to pass their own EGLContext attribute, this is
specially for the case the application let GStreamer chose (e.g. gst-launch).

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1176>

3 years agorawbaseparse: check destination format correctly
Daniel Knobe [Mon, 31 May 2021 07:14:53 +0000 (09:14 +0200)]
rawbaseparse: check destination format correctly

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1181>

3 years agoaudiomixer: Add test for discont going backwards
Olivier Crête [Thu, 27 May 2021 02:50:22 +0000 (22:50 -0400)]
audiomixer: Add test for discont going backwards

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1180>

3 years agoaudioaggregator: Don't overwrite already written samples
Olivier Crête [Thu, 27 May 2021 20:05:04 +0000 (16:05 -0400)]
audioaggregator: Don't overwrite already written samples

On re-sync, don't forget what has already been written. Instead, just
drop any samples that overlap with parts that were already filled.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1180>

3 years agoaudiobasesrc: Fix divide by zero assertion
Seungha Yang [Thu, 27 May 2021 07:20:09 +0000 (16:20 +0900)]
audiobasesrc: Fix divide by zero assertion

GstAudioRingBufferSpec can be cleared from other thread, then
rate value will be zero

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1179>

3 years agoaudio,video-format: Make generate_raw_formats idempotent for assertions
Marijn Suijten [Wed, 26 May 2021 21:22:45 +0000 (23:22 +0200)]
audio,video-format: Make generate_raw_formats idempotent for assertions

When compiling without assertions `g_assert` and its contents disappear
resulting in no list being deserialized at all and the
`gst_{audio,video}_formats_raw` functions to return an empty collection.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1177>

3 years agoUse new gst_buffer_new_memdup()
Tim-Philipp Müller [Sun, 23 May 2021 22:48:39 +0000 (23:48 +0100)]
Use new gst_buffer_new_memdup()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1170>

3 years agocompositor: Consider the converter-config when deciding whether one pad obscures...
Sebastian Dröge [Mon, 24 May 2021 11:25:55 +0000 (14:25 +0300)]
compositor: Consider the converter-config when deciding whether one pad obscures another

If the converter configuration is set to not fill any borders, or if the
border fill color is not full opaque, then the pad has to be handled
as potentially transparent and can't be considered to obscure another
one.

This prevents pads from being wrongly skipped and doing alpha-blending
with uninitialized memory.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1172>

3 years agodoc: Update cache for AV12 pixel format
Nicolas Dufresne [Fri, 21 May 2021 19:04:16 +0000 (15:04 -0400)]
doc: Update cache for AV12 pixel format

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1152>

3 years agovideo: Update video-orc-dist
Nicolas Dufresne [Fri, 21 May 2021 17:28:32 +0000 (13:28 -0400)]
video: Update video-orc-dist

This adds the new symbols needed for AV12 support.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1152>

3 years agogl: add support for AV12
Daniel Almeida [Tue, 18 May 2021 19:20:36 +0000 (16:20 -0300)]
gl: add support for AV12

AV12 is an internally conceived format that is actually the
combination of NV12 and an alpha plane.

This format is to add to gstreamer's webM transparency support for
vp8 and vp9. To this end, two I420 streams are independently decoded
simultaneously for the actual content and the alpha plane respectively
and these are then combined into A420.

This patch adds GL conversion support so that it is possible to convert
from AV12 to RGBA for the purposes of rendering it on a display.

The reverse conversion is also supplied.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1152>

3 years agovideo: add support for AV12
Daniel Almeida [Fri, 21 May 2021 15:24:37 +0000 (12:24 -0300)]
video: add support for AV12

AV12 is an internally conceived format that is actually
the combination of NV12 and an alpha plane.

This format is to add to gstreamer's webM
transparency support for vp8 and vp9. To this end, two
I420 streams are independently decoded simultaneously for
the actual content and the alpha plane respectively
and these are then combined into A420.

Since most hardware decoders output NV12, this patch adds
NV12+A to make the same workflow possible.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1152>

3 years agovideo: video-orc: remove trailing spaces
Daniel Almeida [Fri, 21 May 2021 15:24:00 +0000 (12:24 -0300)]
video: video-orc: remove trailing spaces

Remove unnecessary trailing spaces at the end of some orc functions.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1152>

3 years agotests: Update expectation files with sorted structure fields
Thibault Saunier [Thu, 20 May 2021 15:16:30 +0000 (11:16 -0400)]
tests: Update expectation files with sorted structure fields

3 years agomeson: Fix the name of the `sdp_deps` variable
Thibault Saunier [Thu, 20 May 2021 15:24:15 +0000 (11:24 -0400)]
meson: Fix the name of the `sdp_deps` variable

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1167>

3 years agogl/context/wgl: implement a better get_proc_address()
Matthew Waters [Thu, 20 May 2021 12:52:56 +0000 (22:52 +1000)]
gl/context/wgl: implement a better get_proc_address()

Look in opengl32.dll first, then wglGetProcAddress(), and only then
possibly from any linked in libraries.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1165>

3 years agogl/context: add opengl32.dll as a library to dlopen() on windows
Matthew Waters [Thu, 20 May 2021 12:50:23 +0000 (22:50 +1000)]
gl/context: add opengl32.dll as a library to dlopen() on windows

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1165>

3 years agooverlaycomposition: fix cut-off example code in docs
Tim-Philipp Müller [Thu, 20 May 2021 22:27:21 +0000 (23:27 +0100)]
overlaycomposition: fix cut-off example code in docs

Include everything to the end.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1164>

3 years agogl: add support for A420 conversion
Daniel Almeida [Thu, 6 May 2021 20:26:50 +0000 (17:26 -0300)]
gl: add support for A420 conversion

A420 is a four planar format similar to I420 but with an extra buffer
for alpha values.

A common use of the gl stack is for GPU format conversions using
shaders, in which case one can use glupload, glcolorconvert and
gldownload elements to upload their buffer to the GPU context, perform
the conversion on the GPU itself and then retrieve the data to CPU
context.

A420 was not supported. This patch adds said support mainly by adding
the corresponding shader to perform the conversion and updating the
supported caps.

Both A420->RGBA and RGBA->A420 conversions are supported.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1153>

3 years agoglcolorbalance: Error out on unsupported texture target types
Enrique Ocaña González [Wed, 19 May 2021 17:44:29 +0000 (19:44 +0200)]
glcolorbalance: Error out on unsupported texture target types

The issue can be reproduced on a computer with a Radeon graphics card
when trying to force GStreamer Editing Services to use GL for video
mixing in GESSmartMixer, instead of the GstCompositor that smart mixer
would normally use. This change causes the resulting video stream to
have "video/x-raw(memory:GLMemory) ... texture-target: 2D" caps (instead
of "video/x-raw ..." caps). At the PlaySink stage of the pipeline, a
GstGLImageSinkBin is plugged, with a GstGLColorBalance on it. For some
reason that is still to be debugged (and out of the scope of this
patch), gst_gl_filter_set_caps() is never called on that color balance
element, leaving filter->in_texture_target set to its default
GST_GL_TEXTURE_TARGET_NONE value. The incomplete _create_shader() logic
does the rest and silently generates a shader code that doesn't build.

This is the command I use to reproduce the issue (I'm not sure if I
would be able to isolate the issue in a simple pipeline, though):

 GST_PLUGIN_FEATURE_RANK=vaapih265enc:NONE,vaapih264enc:NONE,vaapisink:NONE,vaapidecodebin:NONE,vaapipostproc:NONE,vaapih265dec:NONE,vaapivc1dec:NONE,vaapih264dec:NONE,vaapimpeg2dec:NONE,vaapijpegdec:NONE,glvideomixer:260
 ges-launch-1.0 +clip /tmp/video.mp4

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1159>

3 years agocompositor: Fix NV12 blend operation
Nicolas Dufresne [Wed, 19 May 2021 20:35:01 +0000 (16:35 -0400)]
compositor: Fix NV12 blend operation

The full src_height/width was being used instead of the remaining
width/height for the current band. As a side effect, that value would
get erroneously reset and would cause overrun.

Fixes #887

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1160>

3 years agovideo: Pass component index not plane index
Nicolas Dufresne [Tue, 18 May 2021 19:36:36 +0000 (15:36 -0400)]
video: Pass component index not plane index

While so far it worked, we are about to introduce a format that break this
assuming. We have a format which consist of NV12 with alpha, and this format
does not have a direct mapping of the component against their plane indexes.

Fix this by using gst_video_format_info_component() introduced in 1.18 for
this purpose.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1151>

3 years agogl/context: move egl creation lower in priority on _new()
Matthew Waters [Tue, 18 May 2021 10:00:01 +0000 (20:00 +1000)]
gl/context: move egl creation lower in priority on _new()

e.g. if running a dual wgl/egl built library, then egl will always
succeed in creating the GstGLContext because almost anything could
support egl, as long as eglGetDisplay() works.

wgl, however has a check for the correct display type so should move
earlier in the tried list.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1154>

3 years agogl/display: provide a gst_gl_display_new_with_type()
Matthew Waters [Tue, 18 May 2021 09:56:13 +0000 (19:56 +1000)]
gl/display: provide a gst_gl_display_new_with_type()

Allows more fine-grained control over the exact display type that is
created.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1154>

3 years agovideoaggregator: Don't try to return void
Seungha Yang [Wed, 19 May 2021 07:23:43 +0000 (16:23 +0900)]
videoaggregator: Don't try to return void

warning C4098: 'gst_video_aggregator_parallel_convert_pad_finalize':
'void' function returning a value

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1155>