Commit graph

59505 commits

Author SHA1 Message Date
Ludvig Rappe
4a1d8eac31 pbutils: Add function for parsing H.264 extradata
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1265>
2021-08-30 08:49:33 +00:00
Edward Hervey
1fe15bb61c qtdemux: Force stream-start push when re-using EOS'd streams
When re-using streams, we *do* need to push a `stream-start` event downstream if
we previously were EOS'd. Failure to do that would never remove the EOS status
on all downstream elements and cause weird issues.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1067>
2021-08-27 14:40:02 +02:00
Brad Smith
7db1040346 deinterlace: Use proper ASM output format for *BSD OS
FreeBSD/NetBSD/OpenBSD amd64 use the ELF binary format.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1066>
2021-08-27 06:41:41 +00:00
Matthew Waters
c906ccb79f element: NULL the lists of contexts in dispose()
If dispose() is called more than once, we may double unref the list of
GstContext's.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/875>
2021-08-27 05:40:55 +00:00
Matthew Waters
50661c1aa9 qmlgl: don't critical on input events before input format has been set
Accessing the unset GstVideoInfo would result in criticals

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1065>
2021-08-27 13:34:01 +10:00
Sebastian Dröge
5472252688 docs: Add Since marker to "twcc-feedback-interval" property
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 11:53:58 +03:00
Havard Graff
32cdea7c73 docs: update with "twcc-feedback-interval"
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Tulio Beloqui
266c2d0619 rtptwcc: changes to use rtp buffer arrival time and current time.
For TWCC we are more interested to track the arrival time (receive side)
and the current time (sender side) of the buffers rather than the
running time.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Knut Inge Hvidsten
0440cb12de rtptwcc: add payloadtype to RTPTWCCPacket
The consumer of the stats can then separate between different media-types,
and do individual stats for each of them.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Havard Graff
8194ab13f7 rtptwcc: make enabling TWCC sticky
Meaning that if a caps comes along that does NOT have TWCC in it,
this does not turn of TWCC for the rest, as this is in fact
completely allowed. (To have some payload-types not containing TWCC
seqnums).

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Havard Graff
b66c6714fa rtptwcc: move TWCC-logic over to the TWCC-manager
Prevent cluttering up the rtpsession, and keeping things localized.

Also write TWCC-seqnums for *all* streams in the session if configured by
caps.

A while back WebRTC was not doing TWCC for audio, basically breaking the
whole idea of a "transport-wide seqnuencenumber" applying for all bundled
streams. However, they have since fixed this, and now it no longers
makes sense to be able to single out certain payloadtypes for
use with TWCC, rather just including them all.

This also makes using RTX, RED, FEC etc much simpler, as it will apply
to them all as they enter the rtpsession.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Havard Graff
ee361bf958 rtptwcc: fix warning
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Tulio Beloqui
4ef0ce282e rtptwcc: fixes and optimizations around run-length chunks
Co-authored-by: Havard Graff <havard.graff@gmail.com>
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Havard Graff
219749c40c rtptwcc: fix seqnum-wrap
Using the proper API to do this is obviously an improvement, and
adding a test for the case of a packet-loss when the seqnum wrap
is also a good idea.

Co-authored-by: Tulio Beloqui <tulio.beloqui@pexip.com>
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Tulio Beloqui
3b14a24630 rtptwcc: fixed feedback packet count overflow that allowed late
packets to be processed

Co-authored-by: Havard Graff <havard.graff@gmail.com>
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Tulio Beloqui
3484f21b95 rtptwcc: fixed parsing of old sequence number
Co-authored-by: Havard Graff <havard.graff@gmail.com>
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Tulio Beloqui
abf4b57a1c rtptwcc: fixed guint8 overflow of feedback packet count
Co-authored-by: Havard Graff <havard.graff@gmail.com>
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Havard Graff
be5fab15e0 rtptwcc: add feedback-interval
To allow RTCP TWCC reports to be scheduled on a timer instead of per
marker-bit.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Havard Graff
ddcde96efe rtptwcc: remove _set_send_packet_ts
Not in use.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Havard Graff
c8400120f1 rtptwcc: make twcc-tests more deterministic
They were a bit racy.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/927>
2021-08-25 08:36:06 +00:00
Olivier Blin
3823f94311 eglimage: fix redefinition of EGLuint64KHR
It is already defined in gst/gl/egl/gstegl.h

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1262>
2021-08-25 09:20:09 +02:00
Jan Alexander Steffens (heftig)
148ac71a1f pad: Keep IDLE probe hook alive during immediate callback
When the probe returns GST_PAD_PROBE_REMOVE and gets called concurrently
from the streaming thread while we're in the callback here, the hook has
already been destroyed by the time we've reacquired the object lock.
Consequently, cleanup_hook gets passed an invalid pointer.

Keep another reference to the hook alive to avoid this situation.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/873>
2021-08-24 15:13:19 +02:00
Tim-Philipp Müller
43f2fd8081 qtdemux: add depth for ProRes 4:4:4:4 variants if available
Might be 24bpp in case an alpha channel is coded but
the image is always opaque.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1061>
2021-08-24 12:35:47 +00:00
Ruslan Khamidullin
f510d48ecf qtmux: for Apple ProRes, allow overriding pixel bit depth for 4:4:4:4 variants
e.g. when exporting an opaque image, yet with alpha channel.

Apple ProRes certification requires that, when a ProRes-writing
application *knows* that the entire frame is opaque, the application
writes only RGB without alpha even when the clip is RGBA. For that,
this tiny change allows the app to override pixel depth when writing ProRes.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1061>
2021-08-24 12:35:47 +00:00
Seungha Yang
c654f86859 video-converter: Add support for A420 to RGB fast path
Add fast path for A420 -> RGB format conversion

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1245>
2021-08-24 11:09:28 +00:00
Havard Graff
068c2a71ba vpxdec: Fix direct rendering, avoid holding write access
When a buffer is pushed downstream, we should try not to hold the
buffer mapped with write access. Doing so would often lead to
an unneccesary memcpy later.

For instance, gst_buffer_make_writable() in
gst_video_decoder_finish_frame() will cause a memcpy because of
_memory_get_exclusive_reference().

We know that we can perform a two-step remap when using system
memory, as this will not cause the location of the memory to
change.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/812>
2021-08-23 14:31:37 +00:00
Matthew Waters
6605174358 isomp4/mux: add a function for seeking to a specific output byte position
We do it enough times that this makes sense.  Also add a debug log line
for the seek position requested.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1060>
2021-08-23 04:17:36 +00:00
Matthew Waters
9dc5cb34ae isomp4/mux: don't overwrite with a bigger moov when fragmenting
When outputting fragmented mp4, with a seekable downstream, we rewrite
the moov to maybe add a duration to the mvex.  If we start by not
writing the initial moov->mvex->mhed duration and then overwrite with a
moov containing mhed atom, the moov's will have different sizes and
could overwrite subsequent data and result in an unplayable file.

e.g. The initial moov would be of size 842 and the final moov would have
a size of 862.

Fix by always pushing out the mhed duration in the moov when
fragmenting.

Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/issues/898

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1060>
2021-08-23 04:17:36 +00:00
Matthew Waters
0d27e6f86e isomp4: actually make streamable fallback work
We weren't setting the fragment_mode field anymore now that the
implementation doesn't change based on the value of the streamable
property.  This lead to invalid files.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1060>
2021-08-23 04:17:36 +00:00
Matthew Waters
d806486503 isomp4: fix trun data offset handling
The trun offset was missing a calculation for one of the box type
headers.

Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/issues/866

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1060>
2021-08-23 04:17:36 +00:00
Matthew Waters
467829358c isomp4/mux: fixes for fragmented mp4 output
Various buffer offset calculations were not quite correct in all cases.

Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/issues/866

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1060>
2021-08-23 04:17:36 +00:00
Seungha Yang
f5a79ce05f compositor: Fix crash while drawing background and/or blending for subsampled YUV
Fix crash caused by out-of-bounds memory accesses when drawing
background and/or blending. This fix is conceptually identical to the
approach as the commit of 8ff5079e5e

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1229>
2021-08-21 01:19:06 +09:00
Edward Hervey
637b0d8dc2 concat: Properly propagate seqnum of segment events
Was broken by https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/819

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/871>
2021-08-20 16:35:53 +02:00
Mathieu Duponchelle
ebb6b9778a encoding-profile: ignore more encoding private fields
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1249>
2021-08-20 14:20:25 +00:00
Thibault Saunier
acf98372a3 smartencoder: Respect user stream-format when specified
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1249>
2021-08-20 14:20:25 +00:00
Mathieu Duponchelle
4aa72cea4f smartencoder: clean up and extend accepted formats
* Add support for H265

* Don't overwrite original codec_data / streamheader in the output
  caps, but instead allow them to change and send them to the
  combiner at the right moment: encoder caps, reencoded GOP,
  original caps, original GOP(s), and potentially encoder caps
  and rencoded last GOP.

* For H264 / H265, force usage of a format with inband SPS / PPS
  (avc3 / hev1), this is cleaner than misadvertising avc1, hvc1 and
  some muxers like mp4mux will actually advertise both differently.

  Unfortunately, while mp4 supports updating the codec_data and using
  avc1 with no in-band SPS / PPS updates, it turns out some decoders
  (eg chrome / firefox) don't handle this particularly well and stop
  decoding after the reencoded GOP. We could expose a switch to
  force usage of avc1 / hvc1 nevertheless, but for now stick to
  requiring that the parser output SPS / PPS in-band with
  config-interval=-1 (that has not changed)

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1249>
2021-08-20 14:20:25 +00:00
Seungha Yang
1ae8b61ec0 compositor: Add "max-threads" property
Adding new property for user to be able to set expected the maximum
number of blend task threads. This can be useful in case that user
wants to restrict the number of parallel task runners for system
resource management or debugging/development purpose.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1242>
2021-08-20 18:43:26 +09:00
Théo MAILLART
adc565ff4a tests: elementfactory: add element creation tests
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/736>
2021-08-20 01:41:30 +00:00
Théo MAILLART
aadf84837b elementfactory: enable construct only property passing
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/736>
2021-08-20 01:41:30 +00:00
Mathieu Duponchelle
9bd8d608d5 matroska-mux: support H264 avc3 / H265 hev1
The matroska codec specs is unfortunately vague on the subject,
stating for H264:

AVC/H.264 stored as described in [@!ISO.14496-15]

and for H265:

HEVC/H.265 stored as described in [@!ISO.14496-15]

This spec however specifies multiple stream formats, our
implementation has opted for interpreting this as avc1 / hvc1,
both of which disallow in-band SPS.

Most decoders however will support in-band SPS / PPS, and
this commit gives the option to explicitly mux in avc3 / hev1,
which allows changing stream parameters on the fly, that is
useful for smart encoding for example.

When either of these stream formats are picked as the input,
changes in codec_data / tier / level / profile do not cause
renegotiation failure, a warning is logged however as it isn't
clear how compliant such a stream is.

The stream-format field is correctly ordered in the template
caps to avoid selecting potentially non-compliant options on
automatic negotiation.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1047>
2021-08-20 00:16:43 +00:00
Mathieu Duponchelle
cb75eda13b isomp4/qtmux: allow renegotiating when tier / level / profile change
Those are carried either in codec_data or in-band SPS (for avc3),
and it is OK for those to change, though decoders obviously need
to support it.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1047>
2021-08-20 00:16:43 +00:00
Mathieu Duponchelle
896c49cf49 isomp4/qtmux: accept video/x-h264, stream-format=avc3
The main difference between avc1 and avc3 is that avc3 is allowed
to contain in-band SPS / PPS. In practice decoders will always use
in-band parameter sets anyway, but it is cleaner to explicitly
advertise it.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1047>
2021-08-20 00:16:43 +00:00
Mathieu Duponchelle
fa835d686f isomp4/qtmux: make sure to switch to next chunk on new caps
For example, with single video sink pad, and new codec_data is
received, current_chunk_offset must be reset to -1 for the
aggregate loop to open a new chunk.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1047>
2021-08-20 00:16:43 +00:00
Mathieu Duponchelle
e069824c7d isomp4/atoms: fix multiple stsd entries
stsd entries are serialized in reverse order (starting from
g_list_last()), and must be prepended to the entry list for their
index to be correct when referenced from stsc entries.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1047>
2021-08-20 00:16:43 +00:00
Olivier Crête
a4a1782872 tracer: Add new tracer to list loaded elements and other features
This new tracer will list loaded elements and plugins. This should
make it easier to generate minimal builds of GStreamer.

This also traces other features such as typefind functions, device
providers and dynamic types.

The format of the output of gst-stats should match the parameters
expected by the meson based gst-build system.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/782>
2021-08-18 17:01:27 -04:00
Olivier Crête
8332b44a2a tracers: Add tracepoint when a plugin feature it loaded
This makes it possible to trace which ones are loaded in a specific
program to make nice statistics.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/782>
2021-08-18 17:01:27 -04:00
Arun Raghavan
2c6be7373f matroska-mux: Add a timestamp-offset property
Adds a user-controllable timestamp offset to clusters and blocks. This
should be useful if we want to have timestamps that have significance
outside of the current file (for example, we might set the offset to the
wallclock when the file is being created, or some other common base, if
we want to correlate streams across multiple files).

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1051>
2021-08-18 10:51:15 -04:00
Sebastian Dröge
bf71ef17e3 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>
2021-08-18 12:06:16 +00:00
Sebastian Dröge
52bca104e4 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>
2021-08-18 09:07:07 +00:00
Edward Hervey
2b01467934 pad: Ensure last flow return is set on sink pads in push mode
The last flow return field was never updated on sink pads in push mode. This
fixes it and makes it consistent.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/868>
2021-08-18 10:25:08 +02:00