platform/upstream/gstreamer.git
11 years agortp: Add an rtpsbcdepay element
Arun Raghavan [Thu, 28 Mar 2013 16:46:36 +0000 (16:46 +0000)]
rtp: Add an rtpsbcdepay element

Pretty straightforward - takes SBC encapsulated in RTP, depayloads, and
pushes out SBC buffers.

https://bugzilla.gnome.org/show_bug.cgi?id=690582

11 years agortp: fix SBC payloader
Tim-Philipp Müller [Wed, 27 Mar 2013 22:18:34 +0000 (22:18 +0000)]
rtp: fix SBC payloader

Init RTP buffer on stack correctly, so mapping it works
without criticals and the payloader actually works.

11 years agodirectsoundsink: Check for a subset instead of non-empty intersection in accept-caps
Sebastian Dröge [Tue, 26 Mar 2013 13:44:36 +0000 (14:44 +0100)]
directsoundsink: Check for a subset instead of non-empty intersection in accept-caps

11 years agodirectsoundsink: Properly handle the filter caps in get_caps()
Sebastian Dröge [Tue, 26 Mar 2013 13:39:53 +0000 (14:39 +0100)]
directsoundsink: Properly handle the filter caps in get_caps()

11 years agodirectsoundsink: Don't unnecessarily get the parent class in class_init
Sebastian Dröge [Tue, 26 Mar 2013 13:35:38 +0000 (14:35 +0100)]
directsoundsink: Don't unnecessarily get the parent class in class_init

The trampoline generated by G_DEFINE_TYPE does that already.

11 years agoUse %03u for format in gst_pad_create_stream_id_printf()
David Schleef [Tue, 26 Mar 2013 01:02:10 +0000 (18:02 -0700)]
Use %03u for format in gst_pad_create_stream_id_printf()

11 years agocapssetter: Prevent unneeded caps copying and allocation
Sebastian Dröge [Mon, 25 Mar 2013 09:12:03 +0000 (10:12 +0100)]
capssetter: Prevent unneeded caps copying and allocation

11 years agocapssetter: Pass any or filter caps upstream
Dirk Van Haerenborgh [Fri, 1 Feb 2013 13:33:41 +0000 (14:33 +0100)]
capssetter: Pass any or filter caps upstream

capsetter accepts anything and just forwards different caps,
as such it should return ANY caps on the sinkpad.

https://bugzilla.gnome.org/show_bug.cgi?id=693005

11 years agovp8enc: Fix for divide by zero when using 0/1 framerate
Tom Greenwood [Wed, 6 Mar 2013 13:17:54 +0000 (13:17 +0000)]
vp8enc: Fix for divide by zero when using 0/1 framerate

https://bugzilla.gnome.org/show_bug.cgi?id=695709

11 years agowavparse: expose CUE sheet items as tracks not chapter entries in TOC
Tim-Philipp Müller [Sun, 24 Mar 2013 17:55:55 +0000 (17:55 +0000)]
wavparse: expose CUE sheet items as tracks not chapter entries in TOC

https://bugzilla.gnome.org/show_bug.cgi?id=677306

11 years agoflacenc: add more example pipelines
Tim-Philipp Müller [Sat, 23 Mar 2013 13:11:02 +0000 (13:11 +0000)]
flacenc: add more example pipelines

11 years agowavenc: add some example pipelines
Tim-Philipp Müller [Sat, 23 Mar 2013 12:59:26 +0000 (12:59 +0000)]
wavenc: add some example pipelines

11 years agowavenc: add TOC support
Anton Belka [Wed, 20 Mar 2013 18:38:40 +0000 (21:38 +0300)]
wavenc: add TOC support

https://bugzilla.gnome.org/show_bug.cgi?id=680998

11 years agoqtdemux: make empty subtitle buffer recognition more robust
Matej Knopp [Sat, 23 Mar 2013 03:56:36 +0000 (04:56 +0100)]
qtdemux: make empty subtitle buffer recognition more robust

https://bugzilla.gnome.org/show_bug.cgi?id=696244

11 years agopngenc: unmap source frame when done
David Schleef [Mon, 4 Mar 2013 23:49:06 +0000 (15:49 -0800)]
pngenc: unmap source frame when done

11 years agoqtmux: Fix test regression with one buffer streams
David Schleef [Fri, 22 Mar 2013 22:14:15 +0000 (15:14 -0700)]
qtmux: Fix test regression with one buffer streams

11 years agoqtdemux: split large raw audio samples
David Schleef [Wed, 6 Mar 2013 01:00:17 +0000 (17:00 -0800)]
qtdemux: split large raw audio samples

In order to deal with a file that has samples that are 24 seconds
long.  Seeking still doesn't work with such files.

11 years agoqtmux: Remove documentation for dts-method
David Schleef [Fri, 22 Mar 2013 18:54:08 +0000 (11:54 -0700)]
qtmux: Remove documentation for dts-method

11 years agoqtmux: deprecate dts-method property
David Schleef [Fri, 22 Mar 2013 20:24:33 +0000 (13:24 -0700)]
qtmux: deprecate dts-method property

11 years agoqtmux: Fix problems causing bad durations in file
David Schleef [Thu, 14 Mar 2013 00:08:03 +0000 (17:08 -0700)]
qtmux: Fix problems causing bad durations in file

- Fix up out-of-order incoming DTS values.
- Fix duration of initial sample.

11 years agoqtmux: fix all timestamps once first_ts is determined
David Schleef [Wed, 13 Mar 2013 02:08:26 +0000 (19:08 -0700)]
qtmux: fix all timestamps once first_ts is determined

11 years agoqtmux: Use PTS/DTS from incoming buffers
David Schleef [Fri, 15 Feb 2013 00:34:34 +0000 (16:34 -0800)]
qtmux: Use PTS/DTS from incoming buffers

Remove old DTS guessing code.

11 years agoqtmux: expose mulaw caps
Nicola Murino [Mon, 18 Mar 2013 11:30:50 +0000 (12:30 +0100)]
qtmux: expose mulaw caps

https://bugzilla.gnome.org/show_bug.cgi?id=696052

11 years agoRequire Orc >= 0.4.17
Tim-Philipp Müller [Fri, 22 Mar 2013 10:50:34 +0000 (10:50 +0000)]
Require Orc >= 0.4.17

Orc 0.4.17 fixes a bunch crashes on i386 and RPi when orc
functions can't be compiled and the fallback function is
supposed to be used. Also fixes some issues on PowerPC.

https://bugzilla.gnome.org/show_bug.cgi?id=684399
https://bugzilla.gnome.org/show_bug.cgi?id=693862

11 years agoqtdemux: fix sample leak when processing private qt tags
Rodolfo Schulz de Lima [Fri, 22 Mar 2013 08:47:17 +0000 (08:47 +0000)]
qtdemux: fix sample leak when processing private qt tags

https://bugzilla.gnome.org/show_bug.cgi?id=696355

11 years agoqtmux: set stream language code from tag
Matej Knopp [Fri, 22 Mar 2013 01:24:01 +0000 (02:24 +0100)]
qtmux: set stream language code from tag

https://bugzilla.gnome.org/show_bug.cgi?id=696358

11 years agoqtdemux: send GAP events for subtitle streams
Matej Knopp [Thu, 21 Mar 2013 01:55:06 +0000 (02:55 +0100)]
qtdemux: send GAP events for subtitle streams

https://bugzilla.gnome.org/show_bug.cgi?id=696244

11 years agoqtdemux: ignore empty subtitle buffers
Matej Knopp [Thu, 21 Mar 2013 01:53:24 +0000 (02:53 +0100)]
qtdemux: ignore empty subtitle buffers

https://bugzilla.gnome.org/show_bug.cgi?id=696244

11 years agoqtdemux: recognize SBTL subtype for subtitles
Matej Knopp [Thu, 21 Mar 2013 01:52:07 +0000 (02:52 +0100)]
qtdemux: recognize SBTL subtype for subtitles

https://bugzilla.gnome.org/show_bug.cgi?id=696244

11 years agoflacparse: add support for the toc-select event
Anton Belka [Sun, 17 Mar 2013 13:27:03 +0000 (16:27 +0300)]
flacparse: add support for the toc-select event

Select tracks from the CUE sheet by sending a toc-select
event based on the uid in the TOC.

https://bugzilla.gnome.org/show_bug.cgi?id=540891

11 years agomp4mux: in faststart mode, don't output up to 4 kB of garbage at the end.
Michael Smith [Wed, 20 Mar 2013 01:09:31 +0000 (18:09 -0700)]
mp4mux: in faststart mode, don't output up to 4 kB of garbage at the end.

11 years agosbcparse: pack multiple frames into one output buffer
Tim-Philipp Müller [Wed, 20 Mar 2013 00:35:17 +0000 (00:35 +0000)]
sbcparse: pack multiple frames into one output buffer

Don't output a single buffer for every tiny SBC frame

11 years agov4l2: fix compilation against newer kernel headers as on FC19
Bastien Nocera [Mon, 18 Mar 2013 14:59:35 +0000 (14:59 +0000)]
v4l2: fix compilation against newer kernel headers as on FC19

11 years agodeinterlace: fix infinite loop on EOS with non-default methods or fields
Kishore Arepalli [Thu, 14 Mar 2013 13:12:05 +0000 (14:12 +0100)]
deinterlace: fix infinite loop on EOS with non-default methods or fields

Fixes problem of infinite loop in gst_deinterlace_reset_history.
Last field in the history was never deinterlaced because idx becomes negative.

Happens e.g. with method=scalerbob fields=bottom or
method=greedyl fields=top

https://bugzilla.gnome.org/show_bug.cgi?id=695644
https://bugzilla.gnome.org/show_bug.cgi?id=693173

11 years agodvdemux: don't return FALSE when dropping sink events
Kishore Arepalli [Tue, 12 Mar 2013 09:48:31 +0000 (09:48 +0000)]
dvdemux: don't return FALSE when dropping sink events

Fixes problem in conjunction with avidemux.

https://bugzilla.gnome.org/show_bug.cgi?id=695643

11 years agoavimux: change raw video caps order so that GRAY8 is last
Tim-Philipp Müller [Tue, 12 Mar 2013 00:16:18 +0000 (00:16 +0000)]
avimux: change raw video caps order so that GRAY8 is last

People like colours.

https://bugzilla.gnome.org/show_bug.cgi?id=695543

11 years agortph264pay: Don't use upstream caps with peer_query_caps ()
Ognyan Tonchev [Mon, 11 Mar 2013 13:50:41 +0000 (14:50 +0100)]
rtph264pay: Don't use upstream caps with peer_query_caps ()

Calling gst_pad_peer_query_caps () on the src pad with the caps
upstream can produce as a filter from gst_rtp_h264_pay_getcaps ()
is wrong and makes caps negotiation fail if upstream caps are not
NULL.

https://bugzilla.gnome.org/show_bug.cgi?id=695629

11 years agoavimux: support raw BGR
Dirk Van Haerenborgh [Sun, 10 Mar 2013 08:10:18 +0000 (09:10 +0100)]
avimux: support raw BGR

https://bugzilla.gnome.org/show_bug.cgi?id=695543

11 years agoavidemux: support raw video with negative height
Dirk Van Haerenborgh [Sun, 10 Mar 2013 08:25:34 +0000 (09:25 +0100)]
avidemux: support raw video with negative height

https://bugzilla.gnome.org/show_bug.cgi?id=695541

11 years agoautodetect checktest: Do not fail without videosink
Jonas Holmberg [Tue, 5 Mar 2013 13:40:56 +0000 (14:40 +0100)]
autodetect checktest: Do not fail without videosink

If there is no videosink available autovideosink will contain a
fakesink instead which needs special treatment in the unit test.

11 years agodtmf: move dtmf plugin from -bad to -good
Tim-Philipp Müller [Sat, 9 Mar 2013 01:18:30 +0000 (01:18 +0000)]
dtmf: move dtmf plugin from -bad to -good

https://bugzilla.gnome.org/show_bug.cgi?id=687416

11 years agoMerge branch 'dtmf-moved-from-bad'
Tim-Philipp Müller [Sat, 9 Mar 2013 00:30:38 +0000 (00:30 +0000)]
Merge branch 'dtmf-moved-from-bad'

https://bugzilla.gnome.org/show_bug.cgi?id=687416

11 years agoosxaudio: add support for iOS using the RemoteIO AudioUnit
Andoni Morales Alastruey [Tue, 5 Mar 2013 20:22:18 +0000 (21:22 +0100)]
osxaudio: add support for iOS using the RemoteIO AudioUnit

11 years agoosxaudio: add a façade for the CoreAudio API
Andoni Morales Alastruey [Tue, 5 Mar 2013 20:17:52 +0000 (21:17 +0100)]
osxaudio: add a façade for the CoreAudio API

11 years agoAutomatic update of common submodule
Tim-Philipp Müller [Thu, 7 Mar 2013 00:00:41 +0000 (00:00 +0000)]
Automatic update of common submodule

From 2de221c to 04c7a1e

11 years agomatroska: Include config.h, it's needed for _stdint.h
Sebastian Dröge [Sun, 3 Mar 2013 10:59:31 +0000 (11:59 +0100)]
matroska: Include config.h, it's needed for _stdint.h

11 years agoflacparse: Fix (wrong) use of uninitialized variable compiler warning
Sebastian Dröge [Sun, 3 Mar 2013 10:53:04 +0000 (11:53 +0100)]
flacparse: Fix (wrong) use of uninitialized variable compiler warning

11 years agoqtdemux: add variant field to H.263 caps
Tim-Philipp Müller [Sat, 2 Mar 2013 13:59:52 +0000 (13:59 +0000)]
qtdemux: add variant field to H.263 caps

avdec_h263 won't get plugged otherwise.

11 years agoqtdemux: skip disabled tracks
Arnaud Vrac [Fri, 22 Feb 2013 18:06:52 +0000 (19:06 +0100)]
qtdemux: skip disabled tracks

ISO/IEC 14496-12 specifies disabled tracks should be completely
ignored, so just do it.

Avoids deadlock during prerolling for some files.

Also prevents 'chapter' subtitle tracks from showing up.

https://bugzilla.gnome.org/show_bug.cgi?id=693993
https://bugzilla.gnome.org/show_bug.cgi?id=628790

11 years agotests: re-add suppression for GValueArray warnings to unit test as well
Tim-Philipp Müller [Mon, 25 Feb 2013 09:58:13 +0000 (09:58 +0000)]
tests: re-add suppression for GValueArray warnings to unit test as well

11 years agotests: use relative include for out-of-tree builds in dtmf test
Jonas Holmberg [Thu, 28 Feb 2013 12:25:06 +0000 (13:25 +0100)]
tests: use relative include for out-of-tree builds in dtmf test

11 years agospectrum: remove the since doc-comment from 0.10
Stefan Sauer [Thu, 28 Feb 2013 07:46:59 +0000 (08:46 +0100)]
spectrum: remove the since doc-comment from 0.10

11 years agolevel: add a "post-messages" property and deprecate "message"
Stefan Sauer [Thu, 28 Feb 2013 07:44:18 +0000 (08:44 +0100)]
level: add a "post-messages" property and deprecate "message"

In spectrum this was changed from 0.10 to 1.0, lets do this here too.

11 years agotests: Add tests for dtmfsrc
Olivier Crête [Wed, 27 Feb 2013 23:56:50 +0000 (18:56 -0500)]
tests: Add tests for dtmfsrc

11 years agotests: Fix ref leak in dtmf test
Olivier Crête [Wed, 27 Feb 2013 21:15:27 +0000 (16:15 -0500)]
tests: Fix ref leak in dtmf test

11 years agortpmp4gdepay: streamtype is not put by all RTSP server, not make it optional
Olivier Crête [Tue, 26 Feb 2013 19:18:20 +0000 (14:18 -0500)]
rtpmp4gdepay: streamtype is not put by all RTSP server, not make it optional

Specific case here is Wowza 3.5.0

11 years agolevel: put back deprecation warnings
Thomas Vander Stichele [Sun, 24 Feb 2013 23:35:58 +0000 (00:35 +0100)]
level: put back deprecation warnings

11 years agolevel: send last message on EOS
Thomas Vander Stichele [Sun, 24 Feb 2013 16:00:14 +0000 (17:00 +0100)]
level: send last message on EOS

11 years agoavidemux: push mode: handle some more 0-size buffer cases
Mark Nauwelaerts [Sat, 23 Feb 2013 13:34:35 +0000 (14:34 +0100)]
avidemux: push mode: handle some more 0-size buffer cases

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=684944

11 years agomatroskamux: fix up example pipeline in docs
Tim-Philipp Müller [Sat, 23 Feb 2013 18:50:52 +0000 (18:50 +0000)]
matroskamux: fix up example pipeline in docs

11 years agopulsesink: Update segdone periodically
Arun Raghavan [Tue, 20 Nov 2012 06:44:07 +0000 (12:14 +0530)]
pulsesink: Update segdone periodically

This makes sure that we update segdone based on the read index received
during latency updates. As the comment notes, we make some compromises
to deal with the fact that segdone is a segment multiple, while the read
index offers finer granularity. The updates are also not very often
(100ms since that is how often automatic timing updates are provided).

All this is required for the baseaudiosink sample alignment code to work
at all.

https://bugzilla.gnome.org/show_bug.cgi?id=694257

11 years agortpsession: Fix wrong code organisation in case of collision
Paul HENRYS [Wed, 13 Feb 2013 09:46:54 +0000 (10:46 +0100)]
rtpsession: Fix wrong code organisation in case of collision

change_ssrc field of RTPSession should be set before calling
rtp_session_schedule_bye_locked () as this function will call reconsider function
that will wake up rtcp_thread which will call rtp_session_on_timeout () that will
check change_ssrc to change the ssrc.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=694184

11 years agoalpha: improve descriptions of chroma keying-related properties and enums
Jean-François Fortin Tam [Thu, 21 Feb 2013 16:15:23 +0000 (11:15 -0500)]
alpha: improve descriptions of chroma keying-related properties and enums

https://bugzilla.gnome.org/show_bug.cgi?id=694374

11 years agoalpha: Do not override the method with custom r/g/b values
Youness Alaoui [Thu, 21 Feb 2013 20:01:15 +0000 (15:01 -0500)]
alpha: Do not override the method with custom r/g/b values

Depending on the order g_object_set() calls aare made, the
target r/g/b settings will override the method if set to
green/blue. Change that so we do not use the target-r/g/b values
unless the method is set to custom.

https://bugzilla.gnome.org/show_bug.cgi?id=694374

11 years agoauparse: do not leak src_caps
Ognyan Tonchev [Wed, 20 Feb 2013 14:46:43 +0000 (15:46 +0100)]
auparse: do not leak src_caps

https://bugzilla.gnome.org/show_bug.cgi?id=694275

11 years agortpsession: only delay RTCP when we are a sender
Wim Taymans [Wed, 20 Feb 2013 19:03:27 +0000 (21:03 +0200)]
rtpsession: only delay RTCP when we are a sender

Only delay the RTCP thread when we are a sender, which we can know because we
have a send_rtp_src pad. Otherwise we might delay the RTCP thread if we
are only a receiver and then there is no code path that wakes up the
RTCP thread and we end up without RTCP packets.

11 years agov4l2: Add support of dmabuf
Benjamin Gaignard [Tue, 19 Feb 2013 10:47:20 +0000 (11:47 +0100)]
v4l2: Add support of dmabuf

v4l has add a new IOCTL to export a buffer by using dmabuf.
This patch allow to use this new IOTCL if it has been defined in videodev2.h
I introduce a new IO mode (GST_V4L2_IO_DMABUF) to enable this way of working.

https://bugzilla.gnome.org/show_bug.cgi?id=693826

11 years agoqtdemux: fix up dodgy code that tries to fix up a broken moov atom
Tim-Philipp Müller [Mon, 18 Feb 2013 20:04:05 +0000 (20:04 +0000)]
qtdemux: fix up dodgy code that tries to fix up a broken moov atom

After gst_buffer_new_and_alloc() gst_buffer_copy_into() will likely
append to the already-existing memory instead of filling it.

11 years agoqtdemux: fix potential crash on short MOOV atom
Tim-Philipp Müller [Mon, 18 Feb 2013 16:32:13 +0000 (16:32 +0000)]
qtdemux: fix potential crash on short MOOV atom

Don't unmap short MOOV atom buffer twice, which happened
in the case where we don't fix up the MOOV atom.

Fixes crashes when thumbnailing partial mp4 file where
the MOOV atom is still incomplete.

https://bugzilla.gnome.org/show_bug.cgi?id=694010

11 years agosouphttpsrc: set SOUP_VERSION_{MIN_REQUIRED,MAX_ALLOWED} to suppress deprecations...
Tim-Philipp Müller [Sat, 16 Feb 2013 16:49:22 +0000 (16:49 +0000)]
souphttpsrc: set SOUP_VERSION_{MIN_REQUIRED,MAX_ALLOWED} to suppress deprecations with newer versions

https://bugzilla.gnome.org/show_bug.cgi?id=693911

11 years agosoup: use default proxy resolver instead of deprecated GNOME proxy resolver
Tim-Philipp Müller [Sat, 16 Feb 2013 15:47:02 +0000 (15:47 +0000)]
soup: use default proxy resolver instead of deprecated GNOME proxy resolver

Apparently there's no reason to use it any longer. Drop libsoup-gnome
dependency while at it, now that we don't need anything from it any
more (it only consists entirely of deprecated API now anyways).

https://bugzilla.gnome.org/show_bug.cgi?id=693911

11 years agotests: fix some h264 caps
Tim-Philipp Müller [Fri, 15 Feb 2013 15:43:43 +0000 (15:43 +0000)]
tests: fix some h264 caps

Doesn't fix anything in particular, but is
still needed here for correctness.

11 years agoaudiopanorama: remove channel-mask from caps
Stefan Sauer [Fri, 15 Feb 2013 07:19:24 +0000 (08:19 +0100)]
audiopanorama: remove channel-mask from caps

The channel-mask is only needed for channels>2 which we don't do.

11 years agov4l2: don't check stride for encoded formats
Benjamin Gaignard [Fri, 15 Feb 2013 15:21:21 +0000 (16:21 +0100)]
v4l2: don't check stride for encoded formats

Don't try to check the stride for encoded formats. Some drivers output
something != 0 and then we don't want to fail on that.

11 years agoudpsrc: use g_socket_set_option() to set buffer size with newer GLib versions
Tim-Philipp Müller [Fri, 15 Feb 2013 14:11:36 +0000 (14:11 +0000)]
udpsrc: use g_socket_set_option() to set buffer size with newer GLib versions

So we have to worry less about portability.

https://bugzilla.gnome.org/show_bug.cgi?id=692400

11 years agojpegdec: remove sof-marker from template caps for now
Tim-Philipp Müller [Thu, 14 Feb 2013 14:13:27 +0000 (14:13 +0000)]
jpegdec: remove sof-marker from template caps for now

Now that the subset check actually works, this breaks
things with demuxers that don't put a "sof-marker"
in their jpeg caps, and we don't have a good parser
to plug either yet.

11 years agojpegenc: Put the SOF marker into the caps
Sebastian Dröge [Wed, 13 Feb 2013 11:32:10 +0000 (12:32 +0100)]
jpegenc: Put the SOF marker into the caps

11 years agortp-payloading: Fix unit test caps and AMR depayloader sink template caps
Sebastian Dröge [Wed, 13 Feb 2013 11:02:46 +0000 (12:02 +0100)]
rtp-payloading: Fix unit test caps and AMR depayloader sink template caps

Fields were missing from the actual caps, or too many fields
existed in the template caps.

11 years agoaacparse: Fix caps used in the unit test
Sebastian Dröge [Wed, 13 Feb 2013 10:53:01 +0000 (11:53 +0100)]
aacparse: Fix caps used in the unit test

The AAC caps passed were incomplete.

11 years agowavpack: Fix unit tests, width is now called depth in the caps in 1.0
Sebastian Dröge [Wed, 13 Feb 2013 10:49:40 +0000 (11:49 +0100)]
wavpack: Fix unit tests, width is now called depth in the caps in 1.0

11 years agotests: make souphttpsrc unit test work even if http_proxy is set
Tim-Philipp Müller [Tue, 12 Feb 2013 23:31:22 +0000 (23:31 +0000)]
tests: make souphttpsrc unit test work even if http_proxy is set

We're testing with an http server on localhost, but don't support
an exception list for the http_proxy, so just unset the environment
variable to make sure we can run this test properly even if the
environment has http_proxy set.

Also, don't skip all tests if there is an issue with the SSL server,
just run the non-SSL tests then.

https://jenkins.qa.ubuntu.com/view/Raring/view/JHBuild%20Gnome/job/jhbuild-amd64-gst-plugins-good/

11 years agoqtdemux: extract codec_data for ProRes
Michael Smith [Tue, 12 Feb 2013 20:53:52 +0000 (12:53 -0800)]
qtdemux: extract codec_data for ProRes

11 years agoavimux: Fixing buffer leak in gst_avi_mux_do_buffer
Tim 'mithro' Ansell [Thu, 7 Feb 2013 14:02:10 +0000 (01:02 +1100)]
avimux: Fixing buffer leak in gst_avi_mux_do_buffer

gst_avi_mux_do_buffer was leaking data from gst_collect_pads_pop.

11 years agoavidemux: correct duration for audio VBR buffers in pull mode
Mark Nauwelaerts [Sun, 10 Feb 2013 14:10:32 +0000 (15:10 +0100)]
avidemux: correct duration for audio VBR buffers in pull mode

11 years agoavidemux: proper position reporting and push mode timestamping
Mark Nauwelaerts [Fri, 8 Feb 2013 20:28:02 +0000 (21:28 +0100)]
avidemux: proper position reporting and push mode timestamping

... and align current_total semantics in push and pull mode,
which tracks bytes for CBR and blocks for VBR.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=691481

11 years agortpsession: delay RTCP until first RTP packet
Wim Taymans [Fri, 8 Feb 2013 16:05:27 +0000 (17:05 +0100)]
rtpsession: delay RTCP until first RTP packet

Delay sending the first RTCP packet until we have sent the first RTP packet.
Otherwise we will send out a Receiver Report instead of a sender report.

See https://bugzilla.gnome.org/show_bug.cgi?id=691400

11 years agortpsession: remove dead code
Wim Taymans [Thu, 7 Feb 2013 14:06:40 +0000 (15:06 +0100)]
rtpsession: remove dead code

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=668355

11 years agortpptdemux: forward sticky events and then set caps
Paul HENRYS [Tue, 29 Jan 2013 09:48:17 +0000 (10:48 +0100)]
rtpptdemux: forward sticky events and then set caps

When a new src pad is added, first forward the sticky events and then
set the caps on the src pad

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692786

11 years agortpjitterbuffer: improve debug output
Markovtsev Vadim [Thu, 7 Feb 2013 13:32:26 +0000 (14:32 +0100)]
rtpjitterbuffer: improve debug output

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=688935

11 years agortpbin: rework cleanup of streams
Wim Taymans [Mon, 26 Sep 2011 21:42:51 +0000 (14:42 -0700)]
rtpbin: rework cleanup of streams

Move the work of cleaning up the client streams in the free_stream
function. This allows us to properly clean up the client streams when we
remove an RTP stream as well.

Based on patch by Sujay <sdatar@cisco.com>

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=660156

11 years agovideomixer2: avoid caps leak
Tim 'mithro' Ansell [Thu, 7 Feb 2013 10:40:35 +0000 (11:40 +0100)]
videomixer2: avoid caps leak

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=693307

11 years agojitterbuffer: do skew estimation only for new timestamps
Wim Taymans [Wed, 6 Feb 2013 16:15:11 +0000 (17:15 +0100)]
jitterbuffer: do skew estimation only for new timestamps

Only run the skew estimation code when we have a new RTP timestamp. If we have
the same RTP timestamp, we simply use the previous estimation. This works
because the new observation with the same RTP timestamp has to have a bigger
receiver time and is thus not going to influence the estimation except for
causing more jitter.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=640023

11 years agortspsrc: only EOS when our source sends BYE
Wim Taymans [Wed, 6 Feb 2013 12:52:26 +0000 (13:52 +0100)]
rtspsrc: only EOS when our source sends BYE

Only EOS when we receive a BYE event from the SSRC of our stream.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=675453

11 years agortspsrc: save the stream SSRC
Wim Taymans [Wed, 6 Feb 2013 12:47:51 +0000 (13:47 +0100)]
rtspsrc: save the stream SSRC

Conflicts:
gst/rtsp/gstrtspsrc.c

11 years agortspsrc: flush connection when stopping
Wim Taymans [Wed, 6 Feb 2013 12:18:18 +0000 (13:18 +0100)]
rtspsrc: flush connection when stopping

When we stop, we can flush all pending commands so that we can stop and
join the task.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=684924

11 years agospectrum: remove outdates readme
Stefan Sauer [Tue, 5 Feb 2013 21:02:13 +0000 (22:02 +0100)]
spectrum: remove outdates readme

Lets remove the readme from pre-0.1.0 that is completely irrelevant now.

11 years agoaudiopanorama: add more debug logging
Stefan Sauer [Tue, 5 Feb 2013 06:32:29 +0000 (07:32 +0100)]
audiopanorama: add more debug logging

11 years agolevel-example. avoid taking the arrays again for each channel for clarity
Stefan Sauer [Tue, 5 Feb 2013 07:26:14 +0000 (08:26 +0100)]
level-example. avoid taking the arrays again for each channel for clarity

Also introduce some blank lines for better readability and update the comments.

11 years agoaudioparsers: fix typo in noinst_headers
Rico Tzschichholz [Mon, 4 Feb 2013 18:38:41 +0000 (18:38 +0000)]
audioparsers: fix typo in noinst_headers

11 years agoaudiopanorama: further port to 1.0
Stefan Sauer [Mon, 4 Feb 2013 10:08:23 +0000 (11:08 +0100)]
audiopanorama: further port to 1.0

Transformcaps is not called with caps containing single structures anymore. Also add missing filter handling. Still does not negotiate though.