platform/upstream/gstreamer.git
10 years agosrtpdec: Fix GstCaps memory leak
Miguel París Díaz [Fri, 3 Jan 2014 08:59:34 +0000 (09:59 +0100)]
srtpdec: Fix GstCaps memory leak

https://bugzilla.gnome.org/show_bug.cgi?id=721382

10 years agopo: update POTFILES
Piotr Drąg [Tue, 31 Dec 2013 21:38:03 +0000 (22:38 +0100)]
po: update POTFILES

https://bugzilla.gnome.org/show_bug.cgi?id=721293

10 years agoandroidmedia: Add new color format, and enhance debug output
Jan Schmidt [Tue, 31 Dec 2013 12:18:54 +0000 (23:18 +1100)]
androidmedia: Add new color format, and enhance debug output

Add a new color format seen on my Galaxy S3
(OMX_SEC_COLOR_FormatNV12Tiled = 0x7fc00002) to the table,
but don't actually implement it - the decoder doesn't choose it.

Remove an assert that makes the plugin fail noisily and take the app down
if it sees a color format it doesn't recognise (just skip the codec instead)

Modify the debug output when plugin scanning to print color format info to
make this sort of thing easier in the future.

10 years agompegtsmux: Don't disrupt buffer state in the clip function
Jan Schmidt [Thu, 19 Dec 2013 12:00:12 +0000 (23:00 +1100)]
mpegtsmux: Don't disrupt buffer state in the clip function

Collectpads assumes that it can pass any buffer to the clip function
for adjustment, some of which are artificially injected - so don't
adjust global timestamp tracking there. Instead, only adjust the
buffer timestamps and use them directly in the collection function.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=698748

10 years agosndfile: emit midi-base-note tag
Stefan Sauer [Mon, 30 Dec 2013 14:49:57 +0000 (15:49 +0100)]
sndfile: emit midi-base-note tag

Query instrument data. Use both 'loop_info' and 'instrument' to inform about the
basenote.

10 years agojpegparse: Remove unneeded call
Edward Hervey [Mon, 30 Dec 2013 12:28:52 +0000 (13:28 +0100)]
jpegparse: Remove unneeded call

'parse' is already provided by the parent variable

10 years agoopencv: The plugin works with 2.4.7 too
Sebastian Dröge [Sun, 29 Dec 2013 12:31:53 +0000 (13:31 +0100)]
opencv: The plugin works with 2.4.7 too

Reported by kwm on IRC.

10 years agoopusparse: Don't send caps in GstBaseParse::start()
Sebastian Dröge [Sat, 28 Dec 2013 11:39:25 +0000 (12:39 +0100)]
opusparse: Don't send caps in GstBaseParse::start()

This is too early and will confuse the event order. The other
code that sets the caps is at the right position and does it
properly already.

10 years agomfc: fix input dequeue for odroid
Víctor Manuel Jáquez Leal [Tue, 24 Dec 2013 23:15:28 +0000 (23:15 +0000)]
mfc: fix input dequeue for odroid

https://bugzilla.gnome.org/show_bug.cgi?id=721027

10 years agoopusenc: increase max payload size to 4000 bytes
Vincent Penquerc'h [Fri, 27 Dec 2013 14:29:46 +0000 (14:29 +0000)]
opusenc: increase max payload size to 4000 bytes

1275 is the maximum size of a frame, but the encoder may return
up to 3 frames, and we need a few extra bytes for TOC, etc. We
use 4000, which is a bit more, and suggested in the libopus docs.

10 years agompegts: fix g-i annotations and gtk-doc markup
Tim-Philipp Müller [Fri, 27 Dec 2013 10:25:30 +0000 (10:25 +0000)]
mpegts: fix g-i annotations and gtk-doc markup

10 years agodashdemux: remove stream loop thread
Thiago Santos [Fri, 20 Dec 2013 13:05:22 +0000 (10:05 -0300)]
dashdemux: remove stream loop thread

Download and push from the same task, makes code a lot simpler
to maintain. Also pushing from separate threads avoids deadlocking
when gst_pad_push blocks due to downstream queues being full.

10 years agouridownloader: fix deadlock near EOS
Thiago Santos [Fri, 20 Dec 2013 22:42:52 +0000 (19:42 -0300)]
uridownloader: fix deadlock near EOS

A deadlock can happen when the source sends EOS when
being put to NULL as the object lock is being held by the
thread that sets the element to NULL and is needed by
the event handler.

10 years agodash: detect EOS when segments are generated by template
Thiago Santos [Fri, 20 Dec 2013 22:39:16 +0000 (19:39 -0300)]
dash: detect EOS when segments are generated by template

Use the Period duration do know when to end the stream

10 years agodashdemux: do not try to access buffer after losing the ref
Thiago Santos [Fri, 13 Dec 2013 20:26:10 +0000 (17:26 -0300)]
dashdemux: do not try to access buffer after losing the ref

For obvious reasons

10 years agodashdemux: use a separate GstUriDownloader per stream
Thiago Santos [Fri, 13 Dec 2013 20:25:37 +0000 (17:25 -0300)]
dashdemux: use a separate GstUriDownloader per stream

Avoids one download having to wait for another to finish
before starting

10 years agodashdemux: store GstActiveStream to avoid getting it every time
Thiago Santos [Mon, 9 Dec 2013 16:55:20 +0000 (13:55 -0300)]
dashdemux: store GstActiveStream to avoid getting it every time

GstActiveStream is used everywhere to operate on the MPD client,
better store it in the GstDashDemuxStream to avoid getting it
everytime

10 years agodashdemux: simplify locking for streams
Thiago Santos [Mon, 9 Dec 2013 16:33:53 +0000 (13:33 -0300)]
dashdemux: simplify locking for streams

Use a single lock for all streams instead of having separate locks.
This makes maintenance easier and at most points we would need
a single lock before iterating on all streams data. So not much
is gained from individual locks.

10 years agodashdemux: fix multiple period mpd playback
Thiago Santos [Mon, 9 Dec 2013 14:28:25 +0000 (11:28 -0300)]
dashdemux: fix multiple period mpd playback

Make dash playlists with multiple periods work again by waiting
to switch the periods when all streams have reached the end of
the current period. The stream_loop is responsible for advancing
the period, but the download loops will already start downloading
data for the next period as soon as possible.

10 years agodashdemux: add combine flows function
Thiago Santos [Thu, 5 Dec 2013 05:53:18 +0000 (02:53 -0300)]
dashdemux: add combine flows function

Use a flow combination function to decide the result from the
stream_loop

10 years agodashdemux: handle multiple languages
Thiago Santos [Wed, 4 Dec 2013 14:30:22 +0000 (11:30 -0300)]
dashdemux: handle multiple languages

Handle multiple languages by using the not-linked return to stop
the download task for that stream. It can be reactivated when
a reconfigure event is received. Stopping the unused streams is
relevant to save network bandwidth

10 years agodashdemux: Use 1 download task per stream
Thiago Santos [Tue, 3 Dec 2013 19:16:09 +0000 (16:16 -0300)]
dashdemux: Use 1 download task per stream

Instead of having a single download task for all streams, this
commit makes each stream have its own download loop, allowing
parallel download of fragments.

10 years agodashdemux: refactor fragment fetching into smaller functions
Thiago Santos [Tue, 3 Dec 2013 03:12:08 +0000 (00:12 -0300)]
dashdemux: refactor fragment fetching into smaller functions

Makes it easier to maintain and extend. This is a first step into
adding multi language support to dashdemux

10 years agodashdemux: run gst-indent
Thiago Santos [Mon, 2 Dec 2013 20:41:01 +0000 (17:41 -0300)]
dashdemux: run gst-indent

Lots of indentation issues making it very annoying to commit
because of the indent hook. Fix this for once.

10 years agodashdemux: expose all streams
Thiago Santos [Mon, 2 Dec 2013 20:31:41 +0000 (17:31 -0300)]
dashdemux: expose all streams

always expose all streams instead of only exposing one of each type.

This is more aligned with gstreamer's way of working. Allows the user
to select the stream that it wants to use by linking its pad and leaving
the unused ones as unlinked.

10 years agoeglglessink: Properly clean up renderbuffers and EAGL context
Sebastian Dröge [Tue, 24 Dec 2013 10:51:16 +0000 (11:51 +0100)]
eglglessink: Properly clean up renderbuffers and EAGL context

10 years agoadpcmenc: Properly set output format
Sebastian Dröge [Mon, 23 Dec 2013 14:33:42 +0000 (15:33 +0100)]
adpcmenc: Properly set output format

Otherwise this will just error out if we only set caps on the srcpad.

10 years agoAutomatic update of common submodule
Tim-Philipp Müller [Sun, 22 Dec 2013 22:33:51 +0000 (22:33 +0000)]
Automatic update of common submodule

From dbedaa0 to d48bed3

10 years agopo: set gettext domain in Makevars so we don't have to patch the generated Makefile...
Tim-Philipp Müller [Sun, 22 Dec 2013 21:56:03 +0000 (21:56 +0000)]
po: set gettext domain in Makevars so we don't have to patch the generated Makefile.in.in

https://bugzilla.gnome.org/show_bug.cgi?id=705455

10 years agoapplemedia: vtdec: set the correct stride in the video meta
Alessandro Decina [Sun, 22 Dec 2013 16:46:40 +0000 (17:46 +0100)]
applemedia: vtdec: set the correct stride in the video meta

10 years agoapplemedia: atdec: fix aac decoding on Mavericks
Alessandro Decina [Sun, 22 Dec 2013 16:44:01 +0000 (17:44 +0100)]
applemedia: atdec: fix aac decoding on Mavericks

10 years agomodplug: small seeking code reshuffle
Stefan Sauer [Fri, 20 Dec 2013 17:12:53 +0000 (18:12 +0100)]
modplug: small seeking code reshuffle

Put the code that checks stop values together. Don't clamp a cur value of -1 to 0 (should not happen right now).

10 years agosfdec: make _stop() non fatal when we have an error already
Stefan Sauer [Fri, 20 Dec 2013 17:10:14 +0000 (18:10 +0100)]
sfdec: make _stop() non fatal when we have an error already

Having no open file in _stop() is okay. Just reset state.

10 years agosfdec: add event function and seeking support
Stefan Sauer [Fri, 20 Dec 2013 17:08:33 +0000 (18:08 +0100)]
sfdec: add event function and seeking support

10 years agosfdec: add taglist support
Stefan Sauer [Wed, 18 Dec 2013 07:12:07 +0000 (08:12 +0100)]
sfdec: add taglist support

Map the metadata strings and a bunch of info-fields to GStreamer tags.

10 years agosfdec: add query implementation for position and duration
Stefan Sauer [Tue, 17 Dec 2013 21:25:03 +0000 (22:25 +0100)]
sfdec: add query implementation for position and duration

10 years agosndfile: rewrite sndfile for 1.0
Stefan Sauer [Tue, 17 Dec 2013 16:56:32 +0000 (17:56 +0100)]
sndfile: rewrite sndfile for 1.0

Add a sfdec for a start. Instead of a source plugin, this is a demuxer/decoder combination. This makes it work with auto-plugging.

10 years agomssdemux: updating docs explaining how it works
Thiago Santos [Fri, 20 Dec 2013 03:39:34 +0000 (00:39 -0300)]
mssdemux: updating docs explaining how it works

mssdemux changed a lot and the docs were not correct anymore.

Also adds the 'Adaptive' category to its details

10 years agomssdemux: do not push event holding object lock
Thiago Santos [Thu, 19 Dec 2013 17:30:13 +0000 (14:30 -0300)]
mssdemux: do not push event holding object lock

Can lead to deadlocks as the push might block downstream in
serialized event cases.

10 years agomssdemux: lock around 'cancelled' flag
Thiago Santos [Thu, 19 Dec 2013 17:29:42 +0000 (14:29 -0300)]
mssdemux: lock around 'cancelled' flag

Prevents race conditions when pipeline is seeking near eos

10 years agoaudiomixer: Also resync timestamps on the RESYNC flag
Sebastian Dröge [Thu, 19 Dec 2013 20:59:09 +0000 (21:59 +0100)]
audiomixer: Also resync timestamps on the RESYNC flag

10 years agoeglglessink: Don't set the EAGLContext of the main thread and don't reuse it either
Sebastian Dröge [Thu, 19 Dec 2013 20:36:50 +0000 (21:36 +0100)]
eglglessink: Don't set the EAGLContext of the main thread and don't reuse it either

... instead create a new context for every sink instance.

https://bugzilla.gnome.org/show_bug.cgi?id=720421

10 years agoapplemedia: corevideobuffer: fix for planar formats
Alessandro Decina [Thu, 19 Dec 2013 07:57:41 +0000 (08:57 +0100)]
applemedia: corevideobuffer: fix for planar formats

10 years agoapplemedia: vtdec: remove silly stride requirement
Alessandro Decina [Thu, 19 Dec 2013 06:54:07 +0000 (07:54 +0100)]
applemedia: vtdec: remove silly stride requirement

This is legacy from the 0.10 times, and it isn't clear how it was useful back
then also.'

10 years agodirectfb: Add hack to fix compile with broken DirectFB define
Jan Schmidt [Thu, 19 Dec 2013 00:11:28 +0000 (11:11 +1100)]
directfb: Add hack to fix compile with broken DirectFB define

Compilation breaks because some versions of DirectFB redefine
__no_instrument_function__, which is used in the G_GNUC_NO_INSTRUMENT
macro.

10 years agomssdemux: fix segfault when trying to access codec_data
Thiago Santos [Tue, 17 Dec 2013 20:55:58 +0000 (17:55 -0300)]
mssdemux: fix segfault when trying to access codec_data

Check if codec_data was obtained before trying to use it to
prevent a segfault

10 years agomssdemux: also generate caps from audiotag attribute
Thiago Santos [Tue, 17 Dec 2013 20:54:31 +0000 (17:54 -0300)]
mssdemux: also generate caps from audiotag attribute

Some audio will have its audio format defined by the AudioTag
instead of the Fourcc. Fallback to using AudioTag if Fourcc
isn't available

10 years agomssdemux: make sure streams stop properly on cancelled
Thiago Santos [Tue, 17 Dec 2013 19:31:52 +0000 (16:31 -0300)]
mssdemux: make sure streams stop properly on cancelled

specially when the gst_task_join is waiting for the task
to go stop and the task itself sets its back to paused, causing
a deadlock on exit.

10 years agomssdemux: track segments per stream
Thiago Santos [Tue, 17 Dec 2013 16:16:58 +0000 (13:16 -0300)]
mssdemux: track segments per stream

As streams now flow independently, the GstSegment needs to be put
on each stream so they can track the position of each one correctly
instead of being mixed in a single segment

10 years agomssdemux: push language code tag
Guillaume Desmottes [Thu, 14 Nov 2013 14:23:42 +0000 (15:23 +0100)]
mssdemux: push language code tag

https://bugzilla.gnome.org/show_bug.cgi?id=712358

10 years agomssmanifest: store the language of the stream
Guillaume Desmottes [Thu, 14 Nov 2013 14:14:18 +0000 (15:14 +0100)]
mssmanifest: store the language of the stream

https://bugzilla.gnome.org/show_bug.cgi?id=712358

10 years agomssdemux: remove the stream loop task
Thiago Santos [Mon, 16 Dec 2013 19:14:24 +0000 (16:14 -0300)]
mssdemux: remove the stream loop task

Download and push from the same task, makes code a lot simpler
to maintain. Also pushing from separate threads avoids deadlocking
when gst_pad_push blocks due to downstream queues being full

10 years agomssdemux: Improve logging
Thiago Santos [Fri, 13 Dec 2013 20:31:11 +0000 (17:31 -0300)]
mssdemux: Improve logging

Show the stream's pad on log messages to make easier to debug
issues in the multiple threads

10 years agomssdemux: improve flow return handling
Thiago Santos [Tue, 10 Dec 2013 21:08:40 +0000 (18:08 -0300)]
mssdemux: improve flow return handling

Handle different flow returns both from the streaming and the
downloading loops

10 years agomssdemux: remove stream locks
Thiago Santos [Tue, 10 Dec 2013 18:41:00 +0000 (15:41 -0300)]
mssdemux: remove stream locks

Simplify the locking by using a single lock instead of having one
lock per stream. This still works and is simpler to maintain.

10 years agomssdemux: avoid downloading not-linked streams
Thiago Santos [Tue, 12 Nov 2013 12:58:31 +0000 (09:58 -0300)]
mssdemux: avoid downloading not-linked streams

When a stream gets a not-linked return, it will be marked as so and
won't download any more new fragments until a reconfigure event
is received. This will make mssdemux expose all pads, but only download
fragments for the streams that are actually being used.

Relying on the pads being linked/unlinked isn't enough in this scenario
as there might be an input-selector downstream that is actually discarding
buffers for a given linked pad.

When streams are switching, the old active stream can be blocked because
input-selector will block not-linked streams. In case the mssdemux's
stream loop is blocked pushing a buffer to a full queue downstream it will
never unblock as the queue will not drain (input-selector is blocking).

In this scenario, stream switching will deadlock as input-selector is
waiting for the newly active stream data and the stream_loop that would
push this data is blocked waiting for input-selector.

To solve this issue, whenever an stream is reactivated on a reconfigure
it will enter into the 'catch up mode', in this mode it can push buffers
from its download thread until it reaches the currrent GstSegment's position.
This works because this timestamp will always be behind or equal to the maximum
timestamp pushed for all streams, after pushing data for this timestamp,
the stream will go back to default and be pushed sequentially from the main
streaming thread. By this time, the input-selector should have already
released the thread.

https://bugzilla.gnome.org/show_bug.cgi?id=711849

10 years agompegvideoparse: Fix documentation of GstMpegVideoSequenceScalableMode
Luis de Bethencourt [Wed, 18 Dec 2013 21:26:17 +0000 (16:26 -0500)]
mpegvideoparse: Fix documentation of GstMpegVideoSequenceScalableMode

10 years agopkgconfig: Don't include libraries in the generic gst-plugins-bad pc file
Sebastian Dröge [Wed, 18 Dec 2013 12:48:54 +0000 (13:48 +0100)]
pkgconfig: Don't include libraries in the generic gst-plugins-bad pc file

10 years agompegts: Install generated headers
Mariusz Buras [Wed, 18 Dec 2013 12:45:22 +0000 (13:45 +0100)]
mpegts: Install generated headers

https://bugzilla.gnome.org/show_bug.cgi?id=720673

10 years agoapplemedia: vtdec: Enable hardware accelerated decoding on 10.9
Davide Bertola [Tue, 17 Dec 2013 09:01:42 +0000 (10:01 +0100)]
applemedia: vtdec: Enable hardware accelerated decoding on 10.9

10 years agocurl: Fix indention
Sebastian Dröge [Tue, 17 Dec 2013 11:17:51 +0000 (12:17 +0100)]
curl: Fix indention

10 years agocurlftpsink: Optionally create a temporary file during FTP transfer/upload
Haridass Selvaraj [Tue, 17 Dec 2013 11:06:13 +0000 (12:06 +0100)]
curlftpsink: Optionally create a temporary file during FTP transfer/upload

https://bugzilla.gnome.org/show_bug.cgi?id=711620

10 years agodecklinksink: Give preference to the order of the filter caps when querying caps
Sebastian Dröge [Tue, 17 Dec 2013 09:23:15 +0000 (10:23 +0100)]
decklinksink: Give preference to the order of the filter caps when querying caps

10 years agodecklinksink: Only intersect with the filter caps if there actually is a filter
Sebastian Dröge [Tue, 17 Dec 2013 09:21:26 +0000 (10:21 +0100)]
decklinksink: Only intersect with the filter caps if there actually is a filter

10 years agovc1parse: Some compiler warning fixes to satisfy XCode compiler
Todd Agulnick [Mon, 16 Dec 2013 05:05:31 +0000 (21:05 -0800)]
vc1parse: Some compiler warning fixes to satisfy XCode compiler

https://bugzilla.gnome.org/show_bug.cgi?id=720513

10 years agoid3tag: Read image-description from the info structure, not the caps
Sebastian Dröge [Mon, 16 Dec 2013 15:07:44 +0000 (16:07 +0100)]
id3tag: Read image-description from the info structure, not the caps

10 years agoconfigure.ac: Fix typo in bz2 check
Aurélien Zanelli [Mon, 16 Dec 2013 10:14:29 +0000 (11:14 +0100)]
configure.ac: Fix typo in bz2 check

https://bugzilla.gnome.org/show_bug.cgi?id=720532

10 years agoid3tag: Write image-type into the APIC tag
Sebastian Dröge [Mon, 16 Dec 2013 10:51:00 +0000 (11:51 +0100)]
id3tag: Write image-type into the APIC tag

10 years agovc1parse: Post VIDEO_CODEC tag
Sebastian Dröge [Mon, 16 Dec 2013 09:25:22 +0000 (10:25 +0100)]
vc1parse: Post VIDEO_CODEC tag

10 years agopngparse: Post VIDEO_CODEC tag
Sebastian Dröge [Mon, 16 Dec 2013 09:24:28 +0000 (10:24 +0100)]
pngparse: Post VIDEO_CODEC tag

10 years agompegvideoparse: Use pbutils API to create the VIDEO_CODEC tag string
Sebastian Dröge [Mon, 16 Dec 2013 09:22:24 +0000 (10:22 +0100)]
mpegvideoparse: Use pbutils API to create the VIDEO_CODEC tag string

10 years agompeg4videoparse: Post VIDEO_CODEC tag
Sebastian Dröge [Mon, 16 Dec 2013 09:21:00 +0000 (10:21 +0100)]
mpeg4videoparse: Post VIDEO_CODEC tag

10 years agoh265parse: Post VIDEO_CODEC tag
Sebastian Dröge [Mon, 16 Dec 2013 09:19:36 +0000 (10:19 +0100)]
h265parse: Post VIDEO_CODEC tag

10 years agoh264parse: Post VIDEO_CODEC tag
Sebastian Dröge [Mon, 16 Dec 2013 09:18:23 +0000 (10:18 +0100)]
h264parse: Post VIDEO_CODEC tag

10 years agoh263parse: Post VIDEO_CODEC tag
Sebastian Dröge [Mon, 16 Dec 2013 09:16:14 +0000 (10:16 +0100)]
h263parse: Post VIDEO_CODEC tag

10 years agodiracparse: Post VIDEO_CODEC tag
Sebastian Dröge [Mon, 16 Dec 2013 09:13:36 +0000 (10:13 +0100)]
diracparse: Post VIDEO_CODEC tag

10 years agodecklinksrc: Process EOS even when there are no incoming video frames.
Todd Agulnick [Sat, 14 Dec 2013 23:12:53 +0000 (18:12 -0500)]
decklinksrc: Process EOS even when there are no incoming video frames.

https://bugzilla.gnome.org/show_bug.cgi?id=680700

10 years agoapplemedia: vtdec: handle level 3 in get_dpb_max_mb_s_from_level. oops.
Alessandro Decina [Sat, 14 Dec 2013 14:55:42 +0000 (15:55 +0100)]
applemedia: vtdec: handle level 3 in get_dpb_max_mb_s_from_level. oops.

10 years agoapplemedia: atdec: fix mp3 caps
Alessandro Decina [Sat, 14 Dec 2013 14:55:22 +0000 (15:55 +0100)]
applemedia: atdec: fix mp3 caps

10 years agodecklinksrc: Set GST_ELEMENT_FLAG_SOURCE on the source
Sebastian Dröge [Sat, 14 Dec 2013 17:42:11 +0000 (18:42 +0100)]
decklinksrc: Set GST_ELEMENT_FLAG_SOURCE on the source

As it does not inherit from basesrc, this flag is not automatically set
and e.g. gst_bin_iterate_sources() and other code does not consider this
element a source.

https://bugzilla.gnome.org/show_bug.cgi?id=680700

10 years agoconfigure.ac: add missing AM_CONDITIONAL for OPENEXR and SBC
Alessandro Decina [Sat, 14 Dec 2013 12:32:26 +0000 (13:32 +0100)]
configure.ac: add missing AM_CONDITIONAL for OPENEXR and SBC

Makes my automake (1.14) happier.

10 years agoapplemedia: always include <Foundation/Foundation.h>
Alessandro Decina [Sat, 14 Dec 2013 12:30:01 +0000 (13:30 +0100)]
applemedia: always include <Foundation/Foundation.h>

Fixes compilation with newest xcode.

10 years agoh264parser: not all startcodes should have 3-byte 0 prefix
Matej Knopp [Thu, 12 Dec 2013 16:49:24 +0000 (17:49 +0100)]
h264parser: not all startcodes should have 3-byte 0 prefix

The parser assumes that every time there is a 0 before the startcode,
it is part of the startcode. But that's not true.

From the specification

Byte stream NAL unit syntax
zero_byte is a single byte equal to 0x00.
  When any of the following conditions are fulfilled, the zero_byte syntax
  element shall be present.
  – the nal_unit_type within the nal_unit( ) is equal to 7 (sequence parameter
    set) or 8 (picture parameter set)
  – the byte stream NAL unit syntax structure contains the first NAL unit of an
    access unit in decoding order, as specified by subclause 7.4.1.2.3.

The problem with doing this for all startcodes is that a trailing zero can mess
up timestamps. The trailing zero gets prepended to the startcode, which will
carry the PTS and DTS of previous buffer.

https://bugzilla.gnome.org/show_bug.cgi?id=664443

10 years agoasfparse: add timestamps to packets
Thiago Santos [Tue, 10 Dec 2013 14:27:52 +0000 (11:27 -0300)]
asfparse: add timestamps to packets

re-enable the code that adds timestamps to packets after baseparse's
port

10 years agoasfparse: port to baseparse
Thiago Santos [Tue, 10 Dec 2013 13:34:02 +0000 (10:34 -0300)]
asfparse: port to baseparse

asfparse was not really functional after the port to 1.0

Now porting it to baseparse to get it working again

10 years agodvbsuboverlay: fix reading of display window in display definition segment
Benjamin Gaignard [Fri, 13 Dec 2013 10:49:52 +0000 (11:49 +0100)]
dvbsuboverlay: fix reading of display window in display definition segment

From ETSI EN 300 743 V1.3.1 (2006-11) 7.2.1 Display definition segment specifictations
the parameters of display window are in this order: Xmin, Xmax, Ymin, Ymax.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@linaro.org>
Signed-off-by: Pierre-Yves Mordret <pierre-yves.mordret@st.com>
https://bugzilla.gnome.org/show_bug.cgi?id=720382

10 years agosrtp: add 256-bit key support
Aleix Conchillo Flaqué [Fri, 13 Dec 2013 22:16:49 +0000 (14:16 -0800)]
srtp: add 256-bit key support

* ext/srtp/gstsrtp.[ch]: added GST_SRTP_CIPHER_AES_256_ICM to
  GstSrtpCipherType and new function cipher_key_size.

* ext/srtp/gstsrtpenc.c: maximum key size is now 46 characters (14 for
  the salt plus the key). If different ciphers are chosen for RTP and
  RTCP the maximum needed key size is expected.

* ext/srtp/gstsrtpdec.c: minor documentation updates.

https://bugzilla.gnome.org/show_bug.cgi?id=720434

10 years agogsmdec: process all available input frames in one go
Tim-Philipp Müller [Thu, 12 Dec 2013 13:21:15 +0000 (13:21 +0000)]
gsmdec: process all available input frames in one go

Instead of parsing, decoding and sending out
lots os little 20ms audio buffers one by one.

10 years agoapplemedia: vtdec: fix compiler warning (uninitialized variable)
Alessandro Decina [Wed, 11 Dec 2013 17:52:53 +0000 (18:52 +0100)]
applemedia: vtdec: fix compiler warning (uninitialized variable)

10 years agoapplemedia: vtdec: rename reorder_queue_frame_delay to reorder_queue_length
Alessandro Decina [Wed, 11 Dec 2013 07:11:50 +0000 (08:11 +0100)]
applemedia: vtdec: rename reorder_queue_frame_delay to reorder_queue_length

10 years agoapplemedia: vtdec: report latency
Alessandro Decina [Wed, 11 Dec 2013 07:10:19 +0000 (08:10 +0100)]
applemedia: vtdec: report latency

10 years agoapplemedia: vtdec: set reorder queue length to the max DPB length
Alessandro Decina [Wed, 11 Dec 2013 06:58:23 +0000 (07:58 +0100)]
applemedia: vtdec: set reorder queue length to the max DPB length

Set reorder_queue_frame_delay from the DPB size (in frames). Still not optimal,
as the DPB size is larger than the max bframe forward prediction length, but I
don't know how to compute the latter without parsing every group of pictures.

10 years agosoundtouch: Call default event handler for basetransform
Jan Schmidt [Tue, 10 Dec 2013 12:06:53 +0000 (23:06 +1100)]
soundtouch: Call default event handler for basetransform

Make pipeline end properly at EOS and not leak by passing events
to the default event handler.

10 years agoresindvd: Remove racy glib warning.
Jan Schmidt [Thu, 5 Dec 2013 13:45:50 +0000 (00:45 +1100)]
resindvd: Remove racy glib warning.

Sometimes, the pad probes can be called twice - for OOB events
travelling downstream. Clear out the pad block id, so it only
gets removed once.

10 years agogsmdec: fix decoding of audio/ms-gsm variant
Tim-Philipp Müller [Fri, 6 Dec 2013 00:22:20 +0000 (00:22 +0000)]
gsmdec: fix decoding of audio/ms-gsm variant

Alternates between 33 and 32 byte frames, but must start
with a 33 byte frame. This has been broken for ages since
the element was ported to the audio decoder base class.

https://bugzilla.gnome.org/show_bug.cgi?id=709416

10 years agoapplemedia: vtdec: set the correct video format in the video meta
Alessandro Decina [Tue, 10 Dec 2013 10:10:54 +0000 (11:10 +0100)]
applemedia: vtdec: set the correct video format in the video meta

Fixes all kinds of weird bugs when videoconvert is used with the decoder.

10 years agocodecparsers: h264: fix picture level scaling lists derivation (rule B).
Cong Zhong [Wed, 17 Apr 2013 02:44:48 +0000 (10:44 +0800)]
codecparsers: h264: fix picture level scaling lists derivation (rule B).

Fix picture level scaling lists derivation from fall-back rule set B,
as specified in 7.4.2.2. More precisely, the sequence level scaling
lists need to be used but intra and inter lists arguments were swapped.

This fixes FRExt/freh5.264 from conformance testing.

https://bugzilla.gnome.org/show_bug.cgi?id=720099

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
10 years agoapplemedia: corevideobuffer: set the correct stride for packed formats
Alessandro Decina [Sun, 8 Dec 2013 15:46:58 +0000 (16:46 +0100)]
applemedia: corevideobuffer: set the correct stride for packed formats

10 years agoapplemedia: vtdec: fix mpeg2 video support
Alessandro Decina [Sun, 8 Dec 2013 14:31:09 +0000 (15:31 +0100)]
applemedia: vtdec: fix mpeg2 video support

10 years agoapplemedia: vtdec: make jpeg decoding actually work
Alessandro Decina [Sun, 8 Dec 2013 14:25:27 +0000 (15:25 +0100)]
applemedia: vtdec: make jpeg decoding actually work