platform/upstream/gstreamer.git
11 years agofrei0r: add the build prefix path in the search dirs
Andoni Morales Alastruey [Fri, 28 Jun 2013 10:25:59 +0000 (12:25 +0200)]
frei0r: add the build prefix path in the search dirs

11 years agodashdemux: Cleanup memory allocation/deallocation
Greg Rutz [Tue, 25 Jun 2013 16:26:24 +0000 (10:26 -0600)]
dashdemux: Cleanup memory allocation/deallocation

Ensure that g_free/xmlFree is used correctly based on how the
memory was allocated.

When deallocating GLists, there were many places that were using
g_list_foreach and g_list_free.  Converted these occurrences to
call g_list_free_full.

Add NULL checks to all xmlFree calls since the documentation does
not guarantee that passing NULL is safe

In places where we are strdup'ing memory allocated by libxml2,
changed those calls to use xmlMemStrdup().

There were several places where we were missing g_slice_free when
deallocating a top-level node structure.

https://bugzilla.gnome.org/show_bug.cgi?id=702837

11 years agompegdesc: Extend list of known descriptors
Edward Hervey [Mon, 24 Jun 2013 11:02:40 +0000 (13:02 +0200)]
mpegdesc: Extend list of known descriptors

Add the various ATSC ones

11 years agompegtsbase: Only warn for added streams when needed
Edward Hervey [Mon, 24 Jun 2013 06:24:58 +0000 (08:24 +0200)]
mpegtsbase: Only warn for added streams when needed

We use add_stream(stream_type:-1) to ensure a programs' PCR Stream is
also taken into account. For most programs this will re-use an
existing ES stream.

So only warn that we are re-adding a stream if it was already present
AND it is not to ensure the PCR stream is taken into account.

11 years agompegtspacketizer: Speed up section parsing
Edward Hervey [Sun, 23 Jun 2013 13:12:58 +0000 (15:12 +0200)]
mpegtspacketizer: Speed up section parsing

Only create subtables when needed. It was previously creating one every
single time ... to check if one was present.
And speed up code to detect whether a subtable was already present or not.

Overall makes section pushing 2 times faster.

11 years agompegdesc: Extend list of known descriptors
Edward Hervey [Mon, 24 Jun 2013 05:40:20 +0000 (07:40 +0200)]
mpegdesc: Extend list of known descriptors

From ITU-T H222.0 (06/2012). Freely available from itu.int

11 years agompegtspacketizer: Speedup packet parsing
Edward Hervey [Sun, 23 Jun 2013 13:26:14 +0000 (15:26 +0200)]
mpegtspacketizer: Speedup packet parsing

Avoid using bitshifting where not needed.

Speeds up packet parsing by 3% (this is *the* hotspot of mpeg-ts elements).

11 years agocodecparsers: Fix gtk-doc
Edward Hervey [Sat, 22 Jun 2013 06:57:12 +0000 (08:57 +0200)]
codecparsers: Fix gtk-doc

Detected by initial pre-run of gi-scanner.

11 years agodaaladec: Add decoder element
Sebastian Dröge [Sun, 23 Jun 2013 09:29:23 +0000 (11:29 +0200)]
daaladec: Add decoder element

11 years agodaalaenc: Add encoder element
Sebastian Dröge [Sun, 23 Jun 2013 08:04:29 +0000 (10:04 +0200)]
daalaenc: Add encoder element

11 years agodaala: Initial version of a Daala plugin
Sebastian Dröge [Sat, 22 Jun 2013 09:27:37 +0000 (11:27 +0200)]
daala: Initial version of a Daala plugin

11 years agosrtp: Fix dist
Edward Hervey [Sun, 23 Jun 2013 07:05:29 +0000 (09:05 +0200)]
srtp: Fix dist

The enums should not be dist-ed and instead be re-generated when
compiling.

11 years agompegtspacketizer: Better detect already seen section
Edward Hervey [Wed, 19 Jun 2013 18:39:54 +0000 (20:39 +0200)]
mpegtspacketizer: Better detect already seen section

In some cases (NIT on highly-populated DVB-C operator for example), there
will be more than one section emitted for the same subtable and version
number.

In order not to lose those updates for the same version number, we checked
against the CRC of the previous section we parsed.

The problem is that, while it made sure we didn't lose any information, it
also meant that if the same section came back (same version, same CRC) later
on we would re-process it, re-parse it and re-emit it.

This version improves on that by keeping a list of previously observed CRC
for identical PID/subtable/version-number and will only process sections if
they really were never seen in the past (as opposed to just before).

On a 30s clip, this brings down the number of NIT section parsing from 4541
down to 663.

https://bugzilla.gnome.org/show_bug.cgi?id=614479

11 years agowinscreencap: Initial port to 1.0
Sebastian Dröge [Thu, 20 Jun 2013 20:12:11 +0000 (22:12 +0200)]
winscreencap: Initial port to 1.0

11 years agoandroidmedia: Fix copying of raw video frames on Samsung Galaxy S3 with Exynos 4 SOC
Xavi Artigas [Thu, 20 Jun 2013 07:52:31 +0000 (09:52 +0200)]
androidmedia: Fix copying of raw video frames on Samsung Galaxy S3 with Exynos 4 SOC

11 years agoshmsink: unref buffer if no clients are connected
Aleix Conchillo Flaque [Wed, 19 Jun 2013 19:52:52 +0000 (12:52 -0700)]
shmsink: unref buffer if no clients are connected

If no client has received the command, unref the buffer. This will
make sure that the shared memory area does not get filled with buffers
no one knows about.

https://bugzilla.gnome.org/show_bug.cgi?id=702684

11 years agompegtspacketizer: Speed up NIT parsing
Edward Hervey [Wed, 19 Jun 2013 17:29:27 +0000 (19:29 +0200)]
mpegtspacketizer: Speed up NIT parsing

Avoid duplicating the delivery structure, and instead give it to the
receiving container.

11 years agompegtbase: simplify code
Edward Hervey [Fri, 7 Jun 2013 10:53:08 +0000 (12:53 +0200)]
mpegtbase: simplify code

We don't need generic methods for push/flush since we know all the
subclass that we control

11 years agompegtsdemux: Fix for compiling without GST_DEBUG
Edward Hervey [Wed, 5 Jun 2013 13:51:58 +0000 (15:51 +0200)]
mpegtsdemux: Fix for compiling without GST_DEBUG

Protect debug-only variables/usage, and in one case remove a useless
variable.

11 years agotsparse: Ensure initial events are sent in the right order
Edward Hervey [Mon, 17 Jun 2013 06:15:54 +0000 (08:15 +0200)]
tsparse: Ensure initial events are sent in the right order

First send stream-start, then caps, then segment.

The segment we push is from upstream in push-mode. If we work in pull-mode
then we initialize the base segment to BYTES.

https://bugzilla.gnome.org/show_bug.cgi?id=702422

11 years agoshmsink: propagate events to basesink class
Aleix Conchillo Flaque [Tue, 18 Jun 2013 19:29:41 +0000 (12:29 -0700)]
shmsink: propagate events to basesink class

https://bugzilla.gnome.org/show_bug.cgi?id=702597

11 years agowayland: Add support for RGBx and RGBA formats
Benjamin Gaignard [Tue, 18 Jun 2013 11:47:54 +0000 (13:47 +0200)]
wayland: Add support for RGBx and RGBA formats

Wayland interface could offer two buffers pixels formats: WL_SHM_FORMAT_XRGB8888 and WL_SHM_FORMAT_ARGB8888.
Update waylandsink to support them and check if the format is really available.

https://bugzilla.gnome.org/show_bug.cgi?id=702112

11 years agoopencv: fix CFLAGS order and add plugins base libs includes to CXXFLAGS
Tim-Philipp Müller [Tue, 18 Jun 2013 11:46:07 +0000 (12:46 +0100)]
opencv: fix CFLAGS order and add plugins base libs includes to CXXFLAGS

https://bugzilla.gnome.org/show_bug.cgi?id=702297

11 years agoopencv: fixes build by moving the c++ header to gstsegmentation.cpp
Alban Browaeys [Fri, 14 Jun 2013 22:12:18 +0000 (00:12 +0200)]
opencv: fixes build by moving the c++ header to gstsegmentation.cpp

Fixes:
In file included from gstsegmentation.h:51:0,
                 from gstopencv.c:42:
/usr/include/opencv2/video/background_segm.hpp:47:16: fatal error: list:
No such file or directory
 #include <list>
                ^
compilation terminated.

https://bugzilla.gnome.org/show_bug.cgi?id=702297

11 years agoapplemedia: fix release of null pointer
Andoni Morales Alastruey [Tue, 18 Jun 2013 11:33:14 +0000 (13:33 +0200)]
applemedia: fix release of null pointer

11 years agoapplemedia: coremediabuffer: fix leaking of apple media structs
Thiago Santos [Tue, 18 Jun 2013 11:17:52 +0000 (13:17 +0200)]
applemedia: coremediabuffer: fix leaking of apple media structs

remember to release the apple's api buffers to avoid huge leaking

11 years agoapplemedia: move defines to CPPFLAGS for obj-c files
Andoni Morales Alastruey [Tue, 18 Jun 2013 11:02:42 +0000 (13:02 +0200)]
applemedia: move defines to CPPFLAGS for obj-c files

11 years agodshowaudiosrc: fix audiocapture producing silence
Andoni Morales Alastruey [Tue, 11 Jun 2013 09:54:41 +0000 (11:54 +0200)]
dshowaudiosrc: fix audiocapture producing silence

Configure the capture latency using the IAMBufferNegotiation
interface and try to respect the configured latency-time and buffer-time

11 years agompegvideoparse: Make sure the buffer is writable before we add the MPEG video meta
Sebastian Dröge [Fri, 14 Jun 2013 12:52:07 +0000 (14:52 +0200)]
mpegvideoparse: Make sure the buffer is writable before we add the MPEG video meta

https://bugzilla.gnome.org/show_bug.cgi?id=702174

11 years agobuild: Fix some automake warnings
Sebastian Dröge [Thu, 13 Jun 2013 11:34:37 +0000 (13:34 +0200)]
build: Fix some automake warnings

11 years agotsdemux: fix M2TS stream resync
Arnaud Vrac [Mon, 27 May 2013 20:22:32 +0000 (22:22 +0200)]
tsdemux: fix M2TS stream resync

Sync byte scan is incorrect for M2TS streams because the timestamp 4
bytes were not included in the flush size. This can result in an
infinite loop.

Rework the scan code to be clearer and work in all cases.

11 years agotsdemux: move some debug prints at the log level
Arnaud Vrac [Fri, 24 May 2013 08:59:55 +0000 (10:59 +0200)]
tsdemux: move some debug prints at the log level

11 years agotsdemux: fixes seeking in pull mode
Josep Torra [Fri, 24 May 2013 08:59:59 +0000 (10:59 +0200)]
tsdemux: fixes seeking in pull mode

Preserve the current segment and observations in pull mode seeks with
flushing.

https://bugzilla.gnome.org/show_bug.cgi?id=698050

11 years agocurlsmtpsink: removed an incorrect assert
Patricia Muscalu [Tue, 11 Jun 2013 17:45:31 +0000 (19:45 +0200)]
curlsmtpsink: removed an incorrect assert

https://bugzilla.gnome.org/show_bug.cgi?id=702036

11 years agoopencv: add foreground/background segmentation element
Miguel Casas-Sanchez [Tue, 11 Jun 2013 12:32:43 +0000 (14:32 +0200)]
opencv: add foreground/background segmentation element

Add an element to the opencv plugin for foregroung/background image
sequence segmentation, using one out of 3 algorithms.

https://bugzilla.gnome.org/show_bug.cgi?id=701421

11 years agowrappercamerabinsrc: Add video-source-filter during construction
Hans de Goede [Mon, 10 Jun 2013 15:00:42 +0000 (17:00 +0200)]
wrappercamerabinsrc: Add video-source-filter during construction

By adding the video-source-filter during construction time, rather then
patching it in later (*), we can greatly reduce the amount of caps involved
in negotation, speeding up pipeline creation.

I wrote this while working on speeding up the startup of cheese. My cheese
has been modified to add a capsfilter, filtering for only the configured
resolution, with that cheese patch + this patch, the pipeline creation time
goes from aprox 1.1 seconds to aprox 350ms. This is with a Logitech 9000
pro camera, which supports lots of different resolutions at many different
framerates per resolution, causing a caps "explosion" if not filtered.

*) Note the code for this is left in, as it is still necessary if the
video-source-filter is changed between a stop + re-start.

https://bugzilla.gnome.org/show_bug.cgi?id=701953

11 years agodashdemux: Make gst-indent happy
Olivier Crête [Mon, 10 Jun 2013 22:00:13 +0000 (18:00 -0400)]
dashdemux: Make gst-indent happy

11 years agowrappercamerabinsrc: Don't unset the video-source property on pipeline start
Hans de Goede [Mon, 10 Jun 2013 08:05:56 +0000 (10:05 +0200)]
wrappercamerabinsrc: Don't unset the video-source property on pipeline start

check_and_replace_src() was setting self->app_vid_src to NULL, which
means that an app setting the video-source property, and then starting,
stopping and re-starting the pipeline (ie to make changes to the
video-source-filter property) would after the restart no longer have
a video-source.

This patch fixes this by making gst_camerabin_setup_default_element return a
ref to the passed in user_element, rather then returning the user_element as
is, so that that ref can be passed on to the bin, and the app_vid_src ref
stays valid.

https://bugzilla.gnome.org/show_bug.cgi?id=701915

11 years agoeglglessink: Remove workaround for a bug in the RPi EGL implementation
Sebastian Dröge [Sat, 8 Jun 2013 12:48:03 +0000 (14:48 +0200)]
eglglessink: Remove workaround for a bug in the RPi EGL implementation

It's fixed in the latest firmware since a few weeks.

11 years agodashdemux: Fix issue with get returning a value for max-buffering-time
Arthur Shipkowski [Mon, 6 May 2013 20:56:01 +0000 (16:56 -0400)]
dashdemux: Fix issue with get returning a value for max-buffering-time

It was not properly divided by GST_SECONDS. Also fix issue with
max-buffering-time being multiplied by GST_SECONDS every time the
property is retrieved.

https://bugzilla.gnome.org/show_bug.cgi?id=700487

11 years agorfb: Link with libgstbase for GstPushSrc
Sebastian Dröge [Fri, 7 Jun 2013 13:18:15 +0000 (15:18 +0200)]
rfb: Link with libgstbase for GstPushSrc

11 years agoandroidmedia: Fix string comparison
Sebastian Dröge [Fri, 7 Jun 2013 12:01:46 +0000 (14:01 +0200)]
androidmedia: Fix string comparison

11 years agompegdemux: accept ID_PRIVATE_STREAM_1 to avoid loosing sync
Alban Browaeys [Thu, 4 Apr 2013 16:22:52 +0000 (18:22 +0200)]
mpegdemux: accept ID_PRIVATE_STREAM_1 to avoid loosing sync

Current fallback to lost_sync seems to impede a delay to restore
sync. Let the parser parse and skip the private stream.
Here it contains the digital camera brand (in 2010 bytes)
and is repeated twice.

https://bugzilla.gnome.org/show_bug.cgi?id=697283

11 years agoh264parse: Wait until SPS/PPS before outputting any data
Ilya Smelykh [Fri, 7 Jun 2013 10:10:08 +0000 (12:10 +0200)]
h264parse: Wait until SPS/PPS before outputting any data

https://bugzilla.gnome.org/show_bug.cgi?id=646327

11 years agoandroidmedia: Hacks for HuaweiMediaPad which reports wrong color format
Chen Jie [Fri, 7 Jun 2013 06:06:56 +0000 (14:06 +0800)]
androidmedia: Hacks for HuaweiMediaPad which reports wrong color format

COLOR_FormatYCbYCr is reported but it is actually COLOR_TI_FormatYUV420PackedSemiPlanar.

11 years agompegvideoparser: Fix the pixel-aspect-ratio calculation
Sreerenj Balachandran [Wed, 5 Jun 2013 13:16:36 +0000 (16:16 +0300)]
mpegvideoparser: Fix the pixel-aspect-ratio calculation

Ignore the display_extension values if they are greater than the width/height
values provided by seqhdr and calculate the PAR based on the seqhdr values.T
his is what DVD players are doing.

Thanks to "David Schleef <ds@schleef.org>"

https://bugzilla.gnome.org/show_bug.cgi?id=685103

11 years agoBack to development
Sebastian Dröge [Wed, 5 Jun 2013 16:36:40 +0000 (18:36 +0200)]
Back to development

11 years agoRelease 1.1.1
Sebastian Dröge [Wed, 5 Jun 2013 16:10:29 +0000 (18:10 +0200)]
Release 1.1.1

11 years agovdpau: Add some missing headers to dist
Sebastian Dröge [Wed, 5 Jun 2013 16:28:17 +0000 (18:28 +0200)]
vdpau: Add some missing headers to dist

11 years agodash: Add to Makefile.am for make dist
Sebastian Dröge [Wed, 5 Jun 2013 16:01:10 +0000 (18:01 +0200)]
dash: Add to Makefile.am for make dist

11 years agoapplemedia-nonpublic: Add to the Makefile.am for dist
Sebastian Dröge [Wed, 5 Jun 2013 15:49:59 +0000 (17:49 +0200)]
applemedia-nonpublic: Add to the Makefile.am for dist

11 years agoUpdate .po files
Sebastian Dröge [Wed, 5 Jun 2013 14:54:57 +0000 (16:54 +0200)]
Update .po files

11 years agompegtspacketizer: Fix leak
Edward Hervey [Wed, 5 Jun 2013 13:20:52 +0000 (15:20 +0200)]
mpegtspacketizer: Fix leak

Since we were manually setting GValue to the GValueArray, we also need
to specify how many were set, otherwise all values in the array will
be leaked.

11 years agoAutomatic update of common submodule
Sebastian Dröge [Wed, 5 Jun 2013 13:15:06 +0000 (15:15 +0200)]
Automatic update of common submodule

From 098c0d7 to 01a7a46

11 years agompegtspacketizer: Speedup TDT/TOT parsing
Edward Hervey [Wed, 5 Jun 2013 11:39:14 +0000 (13:39 +0200)]
mpegtspacketizer: Speedup TDT/TOT parsing

Use quarks where needed, makes it 2.5 times faster

11 years agompegtspacketizer: Fix string leak
Edward Hervey [Wed, 5 Jun 2013 11:21:29 +0000 (13:21 +0200)]
mpegtspacketizer: Fix string leak

The description/text get copied in the structure.

11 years agompegtspacketizer: Speed up descriptor parsing/packing
Edward Hervey [Wed, 5 Jun 2013 10:49:09 +0000 (12:49 +0200)]
mpegtspacketizer: Speed up descriptor parsing/packing

descriptors are stored as a GValueArray of GString. The downside is
that there is no way to "pass" ownership of a GValue to a GValueArray
which previously resulted in expensive copy/free of the (already expensive)
GString.

Here we estimate first the size of the GValueArray, then create it,
then directly use the GValue of that array.

Speeds up total SI parsing by ~30%

11 years agompegtspacketizer: Use gst_value_list_append_and_take_value
Edward Hervey [Wed, 5 Jun 2013 09:21:21 +0000 (11:21 +0200)]
mpegtspacketizer: Use gst_value_list_append_and_take_value

Avoids doing the expensive copy of structures/arrays/...

Speeds up parsing SI by about 50%

11 years agompegtspacketizer: Avoid copying GValueArray
Edward Hervey [Wed, 5 Jun 2013 09:20:49 +0000 (11:20 +0200)]
mpegtspacketizer: Avoid copying GValueArray

Just give the GValueArray to the container GValue.

11 years agocruft: Add signalprocessor directory
Sebastian Dröge [Wed, 29 May 2013 07:25:23 +0000 (09:25 +0200)]
cruft: Add signalprocessor directory

11 years agodocs: removed signalprocessor from docs/libs/Makefile.am
Branko Subasic [Wed, 29 May 2013 06:37:19 +0000 (08:37 +0200)]
docs: removed signalprocessor from docs/libs/Makefile.am

11 years agomidiparse: In pull mode drop SEGMENT, CAPS and STREAM_START events
Sebastian Dröge [Tue, 28 May 2013 13:54:33 +0000 (15:54 +0200)]
midiparse: In pull mode drop SEGMENT, CAPS and STREAM_START events

We create and send our own later from another thread.

11 years agomidiparse: Push stream-start event before anything else
Sebastian Dröge [Tue, 28 May 2013 13:53:07 +0000 (15:53 +0200)]
midiparse: Push stream-start event before anything else

11 years agosignalprocessor: ladspa is not using this anymore
Stefan Sauer [Tue, 28 May 2013 06:08:42 +0000 (08:08 +0200)]
signalprocessor: ladspa is not using this anymore

When we port lv2, we'll go the same route as the new ladspa plugin.

11 years agoladspa: fix typos in launch examples
Stefan Sauer [Tue, 28 May 2013 05:53:37 +0000 (07:53 +0200)]
ladspa: fix typos in launch examples

11 years agoladspa: use the registry cache for plugin details
Stefan Sauer [Mon, 27 May 2013 21:07:16 +0000 (23:07 +0200)]
ladspa: use the registry cache for plugin details

Split the introspection and registration part. This way we only need to open all
plugins when updating the registry. When reading the registry we can register
the elements entierly from the cache.

11 years agoopencv: Add colour image enhancement element based on Retinex algorithm
Miguel Casas-Sanchez [Mon, 27 May 2013 09:20:07 +0000 (11:20 +0200)]
opencv: Add colour image enhancement element based on Retinex algorithm

Add colour image enhancement element based on Retinex algorithm. Two types
exist, namely basic and multiscale; both are described in this article:

Rahman, Zia-ur, Daniel J. Jobson, and Glenn A. Woodell. "Multi-scale retinex
for color image enhancement." Image Processing, 1996. Proceedings.,
International Conference on. Vol. 3. IEEE, 1996

Visually speaking the result looks a bit funny, but is pretty invariable to
lightning changes, which is good for some applications, like image
segmentation.

https://bugzilla.gnome.org/show_bug.cgi?id=700977

11 years agoaudiochannelmix: fill in the author template
Stefan Sauer [Sun, 26 May 2013 12:29:54 +0000 (14:29 +0200)]
audiochannelmix: fill in the author template

The copyright header still contained the author template.

11 years agompegtsdemux: Free memory behind GString if we don't use it
Michael Rubinstein [Sat, 25 May 2013 13:25:56 +0000 (15:25 +0200)]
mpegtsdemux: Free memory behind GString if we don't use it

https://bugzilla.gnome.org/show_bug.cgi?id=700903

11 years agocurlsmtpsink: terminate transfer thread properly
Patricia Muscalu [Thu, 23 May 2013 12:32:07 +0000 (14:32 +0200)]
curlsmtpsink: terminate transfer thread properly

If no EOS has been sent, the curl readfunc callback will
return ABORT. The media file in that case will not be properly
finalized.

https://bugzilla.gnome.org/show_bug.cgi?id=700886

11 years agodirectdrawsink: update window width/height when it's not done in WndProc
Andoni Morales Alastruey [Mon, 8 Apr 2013 14:59:54 +0000 (14:59 +0000)]
directdrawsink: update window width/height when it's not done in WndProc

11 years agodirectdrwasink: call previous WndProc if any
Andoni Morales Alastruey [Fri, 5 Apr 2013 17:59:41 +0000 (17:59 +0000)]
directdrwasink: call previous WndProc if any

11 years agodirectdrawsink: respect PAR with coordinates
Andoni Morales Alastruey [Fri, 5 Apr 2013 17:26:57 +0000 (17:26 +0000)]
directdrawsink: respect PAR with coordinates

11 years agoapplemedia-nonpublic: fix symbols redefinition for objc files
Andoni Morales Alastruey [Thu, 25 Apr 2013 15:35:50 +0000 (15:35 +0000)]
applemedia-nonpublic: fix symbols redefinition for objc files

11 years agoapplemedia: fix duplicated symbols with applemedia
Andoni Morales Alastruey [Wed, 24 Apr 2013 15:55:12 +0000 (15:55 +0000)]
applemedia: fix duplicated symbols with applemedia

11 years agompegtsbase: Post an error message when EOS'ing without source pads
Edward Hervey [Thu, 23 May 2013 12:57:49 +0000 (14:57 +0200)]
mpegtsbase: Post an error message when EOS'ing without source pads

This ensures we don't "hang" when files can't be decoded, or contain
no valid streams.

11 years agotsdemux: Demote a warning to debug level
Edward Hervey [Thu, 23 May 2013 12:56:40 +0000 (14:56 +0200)]
tsdemux: Demote a warning to debug level

We will get it for virtually every first packets.

11 years agocodecparsers: Actually store mpeg video bitrate value
Edward Hervey [Wed, 1 May 2013 14:37:54 +0000 (16:37 +0200)]
codecparsers: Actually store mpeg video bitrate value

And use the extension also (higher 12 bits)

11 years agoopencv: Add skin color detection element
Miguel Casas-Sanchez [Thu, 23 May 2013 09:04:57 +0000 (11:04 +0200)]
opencv: Add skin color detection element

https://bugzilla.gnome.org/show_bug.cgi?id=700654

11 years agovtdec: simplify caps setting
Andoni Morales Alastruey [Tue, 7 May 2013 19:45:54 +0000 (21:45 +0200)]
vtdec: simplify caps setting

11 years agoapplemedia: fix H264 streams with b-frames
Andoni Morales Alastruey [Fri, 26 Apr 2013 15:47:26 +0000 (17:47 +0200)]
applemedia: fix H264 streams with b-frames

The decoder output frames in DTS order, even with the flag
kVTDecodeFrame_EnableTemporalProcessing. We store a internal
queue of the decoded frames and push them PTS order.

11 years agoapplemedia: fix segfault with dropped frames
Andoni Morales Alastruey [Fri, 26 Apr 2013 15:45:31 +0000 (17:45 +0200)]
applemedia: fix segfault with dropped frames

11 years agoapplemedia: improve usage of the VT API
Andoni Morales Alastruey [Fri, 26 Apr 2013 15:40:00 +0000 (17:40 +0200)]
applemedia: improve usage of the VT API

Add timing information to CV samples and pass the GstBuffer
as extra data in the decode function

11 years agoapplemedia: add support for MPEG-1 too
Andoni Morales Alastruey [Fri, 26 Apr 2013 11:26:03 +0000 (13:26 +0200)]
applemedia: add support for MPEG-1 too

11 years agoapplemedia: add support for MPEG-2 decoding
Andoni Morales Alastruey [Thu, 25 Apr 2013 09:30:07 +0000 (11:30 +0200)]
applemedia: add support for MPEG-2 decoding

11 years agoapplemedia: use the best colorformat on each platform
Andoni Morales Alastruey [Thu, 25 Apr 2013 09:27:13 +0000 (11:27 +0200)]
applemedia: use the best colorformat on each platform

This saves a colorspace conversion before the sink in OS X

11 years agoapplemedia: replace private function with its public variant
Andoni Morales Alastruey [Thu, 25 Apr 2013 09:17:16 +0000 (11:17 +0200)]
applemedia: replace private function with its public variant

FigVideoFormatDescriptionCreateWithSampleDescriptionExtensionAtom
is an un-documented private function which might change its signature
as it already did in the past. Replace it with
CMVideoFormatDescriptionCreate and the also un-documented Extensions
dictionary.

11 years agoapplemedia: remove the unneeded buffer factory
Andoni Morales Alastruey [Wed, 24 Apr 2013 20:22:28 +0000 (22:22 +0200)]
applemedia: remove the unneeded buffer factory

11 years agoapplemedia: don't use the dynamic API for public frameworks
Andoni Morales Alastruey [Wed, 24 Apr 2013 20:15:01 +0000 (22:15 +0200)]
applemedia: don't use the dynamic API for public frameworks

Public frameworks don't need to build the API dynamically, we instead
use the framework directly.
The exception is for VideoToolbox which went public in the 10.8 SDK,
but it's still private in older version of the SDK and iOS. This allow
building the plugin against SDK's where it's not a public framework.

11 years agoapplemedia: move plugins using private frameworks to applemedia-nonpublic
Andoni Morales Alastruey [Wed, 24 Apr 2013 16:53:11 +0000 (18:53 +0200)]
applemedia: move plugins using private frameworks to applemedia-nonpublic

11 years agoandroidmedia: Fix indention
Sebastian Dröge [Mon, 20 May 2013 11:24:07 +0000 (13:24 +0200)]
androidmedia: Fix indention

11 years agoandroidmedia: fix a minor error in COLOR_FormatYUV420Planar converting
Chen Jie [Fri, 17 May 2013 12:17:31 +0000 (20:17 +0800)]
androidmedia: fix a minor error in COLOR_FormatYUV420Planar converting

https://bugzilla.gnome.org/show_bug.cgi?id=700521

11 years agoelement-maker: Add transform functions to audiofilter
David Schleef [Sat, 18 May 2013 21:01:22 +0000 (14:01 -0700)]
element-maker: Add transform functions to audiofilter

11 years agoaudiofx: Add plugin, add audiochannelmix
David Schleef [Sat, 18 May 2013 20:57:11 +0000 (13:57 -0700)]
audiofx: Add plugin, add audiochannelmix

11 years agoflitetestsrc: audio format is "S16" not "s16"
David Schleef [Sat, 18 May 2013 20:56:18 +0000 (13:56 -0700)]
flitetestsrc: audio format is "S16" not "s16"

11 years agovideosignal: fix CFLAGS order
Tim-Philipp Müller [Sat, 18 May 2013 10:58:14 +0000 (11:58 +0100)]
videosignal: fix CFLAGS order

11 years agodashdemux: send stream start event
Thiago Santos [Sat, 18 May 2013 00:23:55 +0000 (21:23 -0300)]
dashdemux: send stream start event

Removes warnings about strict sticky events ordering

11 years agomssdemux: fix wma pro caps creation
Thiago Santos [Fri, 17 May 2013 15:08:10 +0000 (12:08 -0300)]
mssdemux: fix wma pro caps creation

WmaPro is actually wmaversion 3, and can also be found by the
WMAP fourcc.

Some manifests also contain the block_align field as "PacketSize"
in the audio track description, the libav decoders require it
to be present in caps.

Fixes #699921

11 years agomssdemux: fix assertions when a stream caps can't be created
Thiago Santos [Fri, 17 May 2013 14:37:30 +0000 (11:37 -0300)]
mssdemux: fix assertions when a stream caps can't be created

Fixes two assertions related to cleanup for a pad that couldn't
be exposed because the caps couldn't be identified from the
Manifest

Fixes #699921

11 years agoeagl: fix handling of surface dimension changes
Thiago Santos [Wed, 15 May 2013 14:49:22 +0000 (11:49 -0300)]
eagl: fix handling of surface dimension changes

Detect when the eagl surface changed its dimension (when the user rotates
the device for example) and adapt the egl internals to draw to that,
preventing that ios resizes the image again when drawing.

This is particularly harmful when eagl would scale down a image
to draw and the ios screen would scale it back up because the
surface is now bigger than when the element was configured.