Our decoder implementation does not use downstream d3d11 pool for
decoding because of special requirement of D3D11/DXVA. So preallocation
using the downstream buffer pool will waste GPU memory in most cases.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2211>
This reverts commit 652773de36 and
modifies it to rename the caps field name to coded-picture-structure.
It was previously removed because it confuses the decoder and we didn't
have a valid use case for including it in the encoded caps at this
stage. We now do have such a use case but still don't want to confuse
the decoder, so the field is renamed.
However, it is still not accurate without looking at the SEI picture
structure of each frame, so it was named coded-picture-structure. If its
value is "frame" it is most likely progressive, if it's "field" it is
most likely interlaced or mixed.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2177>
The current way names the level by the number of B frames it contains, the
less it contains, the higher level it is. So the non ref B frames are in the
lowest layer and the B frames in the highest level refer to I/P frames.
But the widely used way is just the opposite, the ref B frames are in the
lower level and non ref B frames are at the highest level.
The is just a terminology change, and does not have any effect for compression
result and quality.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2149>
timeapi.h is missing in our MinGW toolchain. Include mmsystem.h
header instead, which defines struct and APIs in case of our MinGW
toolchain. Note that in case of native Windows10 SDK (MSVC build),
mmsystem.h will include timeapi.h
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2153>
In case of re-syncing (i.e. moving to another partition to avoid too much of an
interleave), there was previously no checks to figure out whether a given
partition was already fully handled (i.e. when coming across it again after a
previous resync).
In order to handle this at least for single-track partitions, check whether we
have reached the essence track duration, and if so skip the partition.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2150>
The essence track position should only be overriden if we sucesfully switched to
another position. In case of EOS we do not want to override it else we would
increase the track position *again* at the end of this function
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2150>
This field is used by DXVA/NVDEC/VA, and each specification
describes (NVDEC is not well documented) that it's the number of
bits used in short_term_ref_pic_set().
DXVA doesn't explicitly mention that whether the size of
emulation preventation bytes (EPB) is inclusive or not, but
VA is clearly specifying that it's the size after removing
EPB. Excluding EPB size here makes more sense therefore.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1930>
They are part of gst_dep already and we have to make sure to always have
gst_dep. The order in dependencies matters, because it is also the order
in which Meson will set -I args. We want gstreamer's config.h to take
precedence over glib's private config.h when it's a subproject.
While at it, remove useless fallback args for gmodule/gio dependencies,
only gstreamer core needs it.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2031>
Added GstVaFeature enum type, and new parameter for VA allocator's
set_format() and get_format(). Also added a new parameter in VA pool
gst_va_pool_new_with_config() and
gst_buffer_pool_config_set_va_allocation_params().
This new parameter will define if derived images will by used for
buffer mapping.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2057>
Expose description of vendor for user information, similar to
the description property of d3d11device.
Also, set description and DRM device path on GstContext structure
so that user can read them and it will be printed on terminal
when gst-launch-1.0 is used
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2064>
By default, the classification is
"Converter/Filter/Colorspace/Scaler/Video/Hardware", but if VA
post-processor driver supports either color balance, skin tone
enhancement, sharpening or noise reduction, "Effect" is added.
Thus, if vapostproc ranking is raised, it can be chosen by
autovideosink.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2066>
g_signal_disconnect*() doesn't stop any existing callbacks from running
which means that if the notify::state callback is in progress in one
thread and the data channel object is finalize()ed in another thread,
then there could be a use-after-free trying lock the data channel
object.
We can't reasonably use a GWeakRef as we don't have a 'parent' object to
free the GWeakRef after the data channel is finalized. This is also
complicated by the fact that the application can hold a reference to the
data channel object that would live beyond the lifetime of webrtcbin
itself.
We solve this by implementing a ghetto weak-ref solution internally with
a list of outstanding data channels.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1664>
If things progress fast enough, some state changes may not be seen be
the waiting code.
Fix by:
1. keeping a list of all the state changes
2. waiting checks each entry and if the relevant state is found, all
states up to and including then are removed.
This ensures that any waits will see all the state sets.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1664>
Input (sink pads) is the already-ssrc-muxed stream with the relevant rtp
sdes header extensions already applied:
- mid
- stream-id
- repaired-stream-id
Output (src pads) have the pads separated into individual ssrc's as
that's what rtpbin gives us.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1664>
Each rtpbin exposed recv_src pad is now exposed as webrtcbin src_%u pad
now with no meaining applied to the value of %u. Previously this used
to mean the mline in the SDP. If this is is still required, then the
transceiver can be retrieved from the pad and the "mlineindex" property
from the transciever. The "mid" is also retrievable from the
transceiver.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1664>
When creating a transceiver when creating an answer, the media kind of the
transceiver was never set correctly initially. This would lead to a
GST_WARNING being produced about changin a transceiver's media kind.
Fix by retrieving the GstSDPMedia kind from the offer instead as the answer
GstSDPMedia has not been set as the answer caps have not been chosen yet.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1664>
In order to other plugins use gstva objects, such as allocators and buffer
pools, this merge request move them from the va plugin to the gstva library.
This objects are not exposed in <gst/va/gstva.h> since they are not expected
to be used by users, only by plugin implementators.
Because of the surface copy design, which is used to implement allocator's
mem_copy() virtual function, depends on the vafilter, which is kept inside
the plugin, memory copy through VAPosproc is disabled and removed temporarly.
Also added some missing parameter validation.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2048>
Untabifying header file.
The logging category was moved from the plugin generic category to
the display category. It can argue that video formats hacks are
display dependant.
Added validations for input parameters.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2048>
... and add more encoding options.
QSV API supports dynamic bitrate change without IDR insertion.
That's more efficient way of runtime encoding option update
than starting from new sequence with IDR per bitrate option change.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2039>
../gst-libs/gst/mpegts/gst-dvb-section.c:206:9: error: variable 'i' set but not used [-Werror,-Wunused-but-set-variable]
guint i = 0, allocated_events = 12;
^
../gst-libs/gst/mpegts/gst-dvb-section.c:365:9: error: variable 'i' set but not used [-Werror,-Wunused-but-set-variable]
guint i = 0, allocated_streams = 12;
^
../gst-libs/gst/mpegts/gst-dvb-section.c:543:9: error: variable 'i' set but not used [-Werror,-Wunused-but-set-variable]
guint i = 0, allocated_streams = 12;
^
../gst-libs/gst/mpegts/gst-dvb-section.c:885:9: error: variable 'i' set but not used [-Werror,-Wunused-but-set-variable]
guint i = 0, allocated_services = 8;
^
../gst-libs/gst/mpegts/gst-dvb-section.c:1316:9: error: variable 'i' set but not used [-Werror,-Wunused-but-set-variable]
guint i = 0, allocated_services = 8;
^
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2046>
Check that `self` and `self->callback` are defined. `self` can be set to
`NULL` in `remove_listener`, and `self->callback` can be set to `NULL`
inside `gst_amc_surface_texture_jni_set_on_frame_available_callback`.
This can cause a segfault since the Java object can outlive the C
object, and call the callback after `remove_listener` is called.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2024>
Produce an error if we try to use the feature of holding capture buffer
but it is not supported by the driver. Ingoring this can lead to stalls
as the driver will run-out of capture buffer to decode into. This
affects slice decoders but also split-field interlaced decoding.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2009>
This flag is set when the stream is interlaced and the specific
slice is made of single parity fields rather the paired at the
macroblock layer. This is rarely needed in late decoding process
but the Rockchip RKVDEC HW interface requires it, hence needs to
be passed through V4L2 Stateless interface.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2009>
Some problematic H265 stream may miss the reference frame in the DPB,
and get some message like: "No short term reference picture for xxx".
So there may be empty entries in ref_pic_list0/1 when passing to
decode_slice() function of sub class. We need to check the NULL pointer.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2018>
Specify modules to look for OpenEXR when CMake is used, as we may have
CMake config files instead of pkg-config files that result from building
OpenEXR, which may be built with CMake which is typically the case on Visual
Studio builds.
In this case, Meson does seem to find the 'OpenEXR' package with CMake
after trying pkg-config, but does not consider it enough without the
'modules:' argument.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2014>
This reverts commit 3cad3455377d5a22faa138d9df840257059776c8.
That commit was breaking the association between an audio and
a video track in the standard case.
In practice, to support carrying separate MediaStream, we are
going a way to map what MediaStreamTrack belong to what MediaStream,
but that will require some thinking about the API.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2023>
From https://datatracker.ietf.org/doc/html/draft-ietf-mmusic-msid-16:
> Multiple media descriptions with the same value for msid-id and
> msid-appdata are not permitted.
Our previous implementation of simply using the CNAME as the msid
identifier and the name of the transceiver as the msid appdata was
misguided and incorrect, and created issues when bundling multiple
video streams together: the ontrack event was emitted with the same
streams for the two bundled medias, at least in Firefox.
Instead, use the transceiver name as the identifier, and expose
a msid-appdata property on transceivers to allow for further
customization by the application. When the property is not set,
msid-appdata can be left empty as it is specified as optional.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2003>
WebKit is not going to render anything until a URI is set, leading to a
WPE posting a `WPE View did not render a buffer` error message. To avoid
requiring the user to know it if they only want to use
`wpesrc::load-bytes` we can just use `about:blank` as default and
everything will work as users would expect.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1492>
Adding new encoder elements nvd3d11{h264,h265}enc for Direct3D11
input support and re-written nvcuda{h264,h265}enc elements.
Newly writeen elements have some differences compared with old
nv{h264,h265}enc including non-backward compatible changes.
* RGBA is not a supported input format any more:
New elements will support only YUV formats to avoid implicit conversion
done by hardware. Ideally it should be done by upstream element
in order to have more control on it. Moreover, RGBA support can cause
redundant RGBA -> YUV conversion if multiple encoders are
used for the same RGBA input
* Subsampled planar format support is dropped:
I420 and YV12 format are not supported formats for Direct3D11.
Although it's supported in CUDA mode, it's not a hardware friendly
memory layout and it will waste GPU memory since UV planes
will have large padding due to the memory layout requirement of NVENC.
* GL support is dropped: Similar to the RGBA case,
GL support in encoder would be suboptimal if GL input is
used by multiple encoders, because each encoder will copy GL memory
into CUDA memory.
Upstream cudaupload element can be used for GL <-> CUDA
interop instead.
* No more pre-allocation of encoder input surfaces. New implementation
will use input CUDA memory without copy (zero-copy) or
will copy into a NVENC's input buffer struct in case of
system memory input.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1997>
Dispatches a list of active touch events to the wpe view on each
received TOUCH_FRAME event. Touch inputs currently only move the cursor,
since wpe doesn't seem to support clicking/scrolling or zooming with
touch input.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1633>
Since the strings are empty for GST_MSDK_CAPS_MAKE_WITH_DMABUF_FEATURE
and GST_MSDK_CAPS_MAKE_WITH_VA_FEATURE, when excuting
gst-inspect-1.0.exe msdkh265enc, there will be convert static caps error
because of the extra semicolon between two empty strings. Now macro
definitions are added to avoid this issue.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2004>
Pass the current frame to the duplicate_picture callback. This makes it easier
to set the frame's output_buffer if we already have one available. Also
documented that unlike VP9, it is not optional to implement this as the
picture will populate the DPB if it is a key-frame. To ensure this, remove the
default implementation.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1992>
The system_frame_number is notably used by V4L2 decoder as a unique
indentifier for the frame that was decoded. This value is used to tell driver
which frame to reference, as V4L2 does not have an efficient mechanism to
otherwise pass back the frames.
For this reason, and because it is more ligical, copy the original
system_frame_number into the duplicate picture instead of using the current
frame.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1992>
Showing existing keyframe have special meaning in AV1. All the references
frame will be refreshed with the original keyframe information. The refresh
process (7.20) is implemented by saving data from the frame_header into the
state. To fix this special case, load all the relevant information into the
frame_header.
As there is nothing happening in between this and the loading of the key-frame
into the state, this patch also remove the separate API function, using it
internally instead.
Fixes#1090
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1971>
We bind transceivers' fec_percentage property to the FEC encoder
percentage property, and with the binding bidirectional a deadlock
was introduced by the latest changes from !1762:
We take hold of the transceiver's object lock, then add the binding
and set the property to its initial value on the encoder, which causes
set_property to deadlock in the transceiver when the binding kicks in.
Changing the binding type to DEFAULT (source to target) is enough
to address the deadlock and still serves the original intent.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1967>
Holding previously decoded but not outputted pictures even after
new_sequence is not a safe approach in various aspect.
However, we cannot drain out DPB on new_sequence() unconditionally,
because there is a case where decoder should drop decoded pictures
if NoOutputOfPriorPicsFlag is set.
To detect NoOutputOfPriorPicsFlag before the new_sequence() call,
this patch splits decoding process into two path, one for nal unit parsing
in order to detect NoOutputOfPriorPicsFlag and then each nal unit
will be decoded.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1937>
1. Always set the according GstVaH264EncFrame pointer when GstVideoCodecFrame
pointer is assigned, which can make the logic safe.
2. Fix the forgotten change in _sort_by_frame_num. Its input pointer now is
GstVideoCodecFrame type.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1935>
Make it posible to configure the element to obtain the timestamps from
reference timestamp meta data instead of using the ntp-offset property,
or estimating its own offset. Currently the only time format supported
is "timestamp/x-unix", i.e. UTC time expressed in the unix time epoch.
In addition the custom event GstNtpOffset has been renamed to
GstOnvifTimestamp, to reflect that it is not necessarily used to convey
the ntp-offset. As a consequence we had to modify a couple of files in
the rtsp-server as well.
Fixes#984
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1683>
On GstVideoDecoder::{drain,flush}, we send null packet with
CUVID_PKT_ENDOFSTREAM flag to drain out decoder. Which will
reset CUVID parser as well.
To continue decoding after the drain, the next input buffer
should include sequence headers otherwise CUVID parser will
not report any decodeable frame.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1911>
There could be a case where the new program has the same program number as the
previous one ... but is actually located on a PID previously used for elementary
stream. In that case the program is guaranteed to not be an update of the
previous program but a completely new one.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1893>
We need to be able to look for programs by their PID also. Using a hash table
was a bit sub-par (and overkill) for storing a range of programs.
This is needed because there could potentially be two programs with the same
program id but different PMT PID (while one is being deactivated the new one
would "exist").
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1893>
Instead of using GstMiniObject to hold H264 frame, now it uses a plain
structure. Besides, instead of holding a reference to
GstVideoCodecFrame, the H264 frame structure is set as a
GstVideoCodecFrame user data.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1856>
And use the output segment position for the outgoing timestamp while it
is. This is needed to delay the calculation of `output_ts_offset` until
we actually have a usable timestamp, as tsmux will output a few initial
packets while `last_ts` is still unset.
Without this, the calculation would use the initial `0` value, which did
not have the intended effect of making VBR mode behave like CBR mode,
but always calculated an offset equal to the selected start time.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1884>
When there is vpp scaling downstream, we need to make sure SFC is not
triggered because vpp may fall into passthrough mode which causes
the decoder negotiation to create src caps with vpp scaled width/height.
This patch includes bitstream's original size in first query with
downstream in gst_msdkdec_src_caps, which is the same for what we do for
color format in this query. This is to ensure SFC scaling starts to
work only when downstream directly asks for a different size instead of
through vpp.
Note that here SFC scaling follows the same behavior as msdkvpp:
if user only changes width or height, e.g. dec ! video/x-raw,width=xx !,
the height will be modified to the value which fits the original DAR.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1838>
* Hide GstCudaMemory member variables
* Make GstCudaAllocator object GstCudaContext independent
* Set offset/stride of memory correctly via video meta
* Drop GST_BUFFER_POOL_OPTION_VIDEO_ALIGNMENT support.
This implementation actually does not support custom alignment
because we allocate device memory via cuMemAllocPitch
of which alignment is almost uncontrollable
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1834>
* cudaupload/download
- Specify only formats actually we can deal with
nvcodec elements, not all video formats
- Supports CUDA output for download and input for upload in order
to make passthrough possible, like other upload/download elements.
* cudabasetransform
- Reset conversion element if upstream CUDA memory
holds different CUDA context and the element can accept it.
This is the same behavior as corresponding d3d11 filter elements.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1834>
This was to support very old V4L2 kernel. As we moved to DMABuf and can now
detach buffers on renegotiation, the buffer it tries to fix no longer exist.
The risk to blocking indefinitly the application does still exist though.
Fixes#1070
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1861>
When we negotiate with downstream, We should use the intersected
caps of input and output to decide the alignment and stream format.
The current code just uses the input caps which may lack the stream
format.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1837>
The demux now outputs the AV1 stream in "tu" alignment, so we do not need
to detect the input alignment. But the annex b stream format is not recognized
by the demux, we still need to detect that stream format for the first input.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1837>
Decoders that required frame aligmment and didn't have an associated
alpha decoder were skipped. This is because the parser was constructing
caps based on the software alpha decoder, which specify super-frame
alignment.
Iterate over the caps to filter the one that have a matching codec-alpha, with
the semantic the no codec-alpha field means codec-alpha=false. Then if
everything was removed, callback to the original, so that the first non-alpha
decoder will be picked.
Fixes#820
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1855>
Currently for copying the coded buffer onto a GStreamer buffer, the
coded buffer is mapped two times: one for getting the size, and later
for do the actual copy. We can avoid this by doing directly in the
element rather than in the general encoder object.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1845>
We need to always add the RTX/RED/ULPFEC elements as rtpbin will only
call us once to request aux/fec senders/receivers.
We also need to regenerate the media section of the SDP instead of
blindly copying from the previous offer.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1762>
Remove the symbolic link `gst-uninstalled` which points to `gst-env`.
The `uninstalled` is the old name and the project should stick to a
single name for the procedure.
Remove the term from all the files, exceptions are variables from
dependencies like `uninstalled_variables` from pkgconfig and
`meson-uninstalled`.
Adjust mentions of the script in the documentation and README.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1743>
Do not maintain similar build instructions within each gst-plugins-*
subproject and the subproject/gstreamer subproject. Use the build
instructions from the mono-repository and link to them via hyperlink.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1743>
This a new VA-API implementation of a H264 encoder.
It can control the GOP and parameter settings, while the MV searching,
VCL and the rate control algorithm are implemented by VA drivers and HW.
It supports most of the common usage options in H264, but still lacks
of look ahead, field, B frame weighted prediction, etc.
Co-authored-by: Victor Jaquez <vjaquez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1051>
The g_queue_clear_full() and g_array_copy() functions in the glib
may not be available for the current glib version check, so we add
helper functions to wrap it.
This should be deleted after the glib version bumps.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1051>
The fd was in different meanings on windows:
POSIX read and write use the fd as a file descriptor.
The gst_poll use the fd as a WSASocket.
This patch use WSASocket as default on windows. This is a temporary measure, because IPC has many different implement. There may be a better way in the future.
See #1044
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1791>
