platform/upstream/gst-plugins-good.git
14 years agortpg729pay: ptime should is in nanoseconds
Olivier Crête [Tue, 19 Jan 2010 02:15:51 +0000 (21:15 -0500)]
rtpg729pay: ptime should is in nanoseconds

https://bugzilla.gnome.org/show_bug.cgi?id=607403

14 years agowavenc: Post warning if file isnt finished properly
Thiago Santos [Wed, 20 Jan 2010 18:11:15 +0000 (15:11 -0300)]
wavenc: Post warning if file isnt finished properly

When the pipeline is shut down and the file isn't
finished properly, wavenc should post a warning.

Fixes #607440

14 years agomatroskamux: make index size configurable.
Arnout Vandecappelle [Wed, 27 May 2009 11:51:44 +0000 (13:51 +0200)]
matroskamux: make index size configurable.

Added the 'min-index-interval' property to matroskamux,
which determines how much time (nanoseconds) is left
between keyframes stored in the index.

Fixes #583985.

14 years agortph264pay: scale spspps_interval to milliseconds
Wim Taymans [Wed, 20 Jan 2010 15:28:31 +0000 (16:28 +0100)]
rtph264pay: scale spspps_interval to milliseconds

The spspps_interval is kept in seconds. Convert it to milliseconds before
comparing it to another value in milliseconds.

14 years agoqtdemux: always keep media segments within total duration
Mark Nauwelaerts [Wed, 20 Jan 2010 14:18:47 +0000 (15:18 +0100)]
qtdemux: always keep media segments within total duration

... as opposed to only doing so following a seek.

14 years agortph264pay: rename spspps-interval property
Wim Taymans [Wed, 20 Jan 2010 14:44:40 +0000 (15:44 +0100)]
rtph264pay: rename spspps-interval property

Rename the spspps-interval property to config-interval because it is nicer.

14 years agoavidemux: skip RIFF and index in push mode
Wim Taymans [Tue, 19 Jan 2010 17:37:31 +0000 (18:37 +0100)]
avidemux: skip RIFF and index in push mode

When we are in push mode, we can encounter RIFF and idx tags in the data chunk
when we are dealing with ODML files. In these cases, simply skip the chunks and
continue streaming instead of going EOS.

14 years agoavidemux: more DISCONT handling
Wim Taymans [Wed, 20 Jan 2010 10:27:23 +0000 (11:27 +0100)]
avidemux: more DISCONT handling

Add some debug in the DISCONT handling code.
When we receive a DISCONT in push mode, mark all streams as DISCONT.

14 years agoavidemux: reset on flush events
Wim Taymans [Wed, 20 Jan 2010 10:26:34 +0000 (11:26 +0100)]
avidemux: reset on flush events

When we receive a flush event on the sinkpad, reset the EOS state and the
flowreturn of all streams. Also mark the streams with a DISCONT.

14 years agoavidemux: rename some variable
Wim Taymans [Wed, 20 Jan 2010 10:22:04 +0000 (11:22 +0100)]
avidemux: rename some variable

Rename the seek_event variable to seg_event because it really contains the
newsegment event that needs to be pushed.

14 years agoAutomatic update of common submodule
Tim-Philipp Müller [Wed, 20 Jan 2010 00:54:03 +0000 (00:54 +0000)]
Automatic update of common submodule

From 14cec89 to 15d47a6

14 years agortph264pay: Don't set profile-level-id in out caps
Olivier Crête [Mon, 18 Jan 2010 19:49:26 +0000 (14:49 -0500)]
rtph264pay: Don't set profile-level-id in out caps

The profile-level-id represents restrictions on what can be sent, it does not
describe the stream. So it should be reflected in the sink caps of the
payloader, not the src caps.

https://bugzilla.gnome.org/show_bug.cgi?id=607353

14 years agortph264pay: Don't ignore the return value from set_outcaps
Olivier Crête [Mon, 18 Jan 2010 19:41:10 +0000 (14:41 -0500)]
rtph264pay: Don't ignore the return value from set_outcaps

https://bugzilla.gnome.org/show_bug.cgi?id=607353

14 years agodeinterlace: Fix license and copyright headers
Sebastian Dröge [Mon, 18 Jan 2010 16:43:41 +0000 (17:43 +0100)]
deinterlace: Fix license and copyright headers

14 years agov4l2: move G_END_DECLS to the end
Stefan Kost [Mon, 18 Jan 2010 12:57:42 +0000 (14:57 +0200)]
v4l2: move G_END_DECLS to the end

14 years agov4l2: fix bufferpool file names in header comment
Stefan Kost [Mon, 18 Jan 2010 12:55:38 +0000 (14:55 +0200)]
v4l2: fix bufferpool file names in header comment

14 years agoavidemux: avoid some typecasting
Wim Taymans [Fri, 15 Jan 2010 17:15:14 +0000 (18:15 +0100)]
avidemux: avoid some typecasting

14 years agoavidemux: avoid some type checks
Wim Taymans [Fri, 15 Jan 2010 17:13:24 +0000 (18:13 +0100)]
avidemux: avoid some type checks

14 years agoavidemux: fallback to avih duration
Wim Taymans [Fri, 15 Jan 2010 17:09:15 +0000 (18:09 +0100)]
avidemux: fallback to avih duration

when we have not yet parsed the indexes (in push mode, for example) use
the duration as given in the avih header instead of -1.

14 years agoqtdemux: g_free is NULL safe
Thiago Santos [Fri, 15 Jan 2010 16:32:32 +0000 (13:32 -0300)]
qtdemux: g_free is NULL safe

14 years agoqtdemux: use DEMUX errors, instead of DECODE
Thiago Santos [Fri, 15 Jan 2010 16:27:40 +0000 (13:27 -0300)]
qtdemux: use DEMUX errors, instead of DECODE

qtdemux should use DEMUX errors, and not DECODE

Conflicts:

gst/qtdemux/qtdemux.c

14 years agoqtdemux: Minor refactor
Thiago Santos [Thu, 14 Jan 2010 22:16:19 +0000 (19:16 -0300)]
qtdemux: Minor refactor

Replace repeated code with a function call

14 years agoqtdemux: Handle another kind of redirect trak
Thiago Santos [Thu, 14 Jan 2010 20:11:13 +0000 (17:11 -0300)]
qtdemux: Handle another kind of redirect trak

Some traks might contain a redirect rtsp uri inside
hndl atom (which is a dref atom entry). This commit makes qtdemux
post a message when it finds one of these traks and there are
no other traks.

Fixes #597497

14 years agoqtdemux: Post error when reaching EOS without pads
Thiago Santos [Thu, 14 Jan 2010 19:13:08 +0000 (16:13 -0300)]
qtdemux: Post error when reaching EOS without pads

Post an error when EOS is reached and there are no src pads

14 years agoqtdemux: Do not post empty redirect messages
Thiago Santos [Thu, 14 Jan 2010 17:13:50 +0000 (14:13 -0300)]
qtdemux: Do not post empty redirect messages

Some misinterpreted data could result in posting redirect messages
with empty redirect strings. It is better not to post them.

An example is the file on bug #597497

14 years agomatroskademux: polish last buffer end time usage
Mark Nauwelaerts [Thu, 14 Jan 2010 17:19:25 +0000 (18:19 +0100)]
matroskademux: polish last buffer end time usage

That is, reset it upon seek, and note that (rarely) last pushed buffer
time might precede segment start.

14 years agovideomixer: use 'q' constraint instead of 'r'
Stefan Kost [Wed, 13 Jan 2010 14:48:46 +0000 (16:48 +0200)]
videomixer: use 'q' constraint instead of 'r'

This avoids the "bad register name `%dil'" compilation errors on 32bit where
because of 'r' gcc puts the value in a general purpose register and then tries
to access the lower part as %dil/%sil which is not existing on 32bit. 'q' requests
a-d registers

14 years agoavi: add missing include for sscanf
Stefan Kost [Wed, 13 Jan 2010 14:44:58 +0000 (16:44 +0200)]
avi: add missing include for sscanf

14 years agoequalizer: Fix property description for the 3rd band of the 10band equalizer
Sebastian Dröge [Wed, 13 Jan 2010 08:36:03 +0000 (09:36 +0100)]
equalizer: Fix property description for the 3rd band of the 10band equalizer

The frequency is actually 237 Hz, not 227 Hz.

Fixes bug #606692.

14 years agoaudioamplify: Allow negative amplifications
Kipp Cannon [Wed, 13 Jan 2010 08:22:20 +0000 (09:22 +0100)]
audioamplify: Allow negative amplifications

Fixes bug #606807.

14 years agoapev2mux: Don't call constructors directly, this leads to compiler errors with gcc 4.5
Sebastian Dröge [Wed, 13 Jan 2010 08:17:05 +0000 (09:17 +0100)]
apev2mux: Don't call constructors directly, this leads to compiler errors with gcc 4.5

14 years agoqtdemux: use G_GSIZE_FORMAT for platform independent gsize qualifier
Edward Hervey [Tue, 12 Jan 2010 16:39:05 +0000 (17:39 +0100)]
qtdemux: use G_GSIZE_FORMAT for platform independent gsize qualifier

Fixes build on macosx

14 years agomatroskademux: refactor eos sending when pausing loop
Mark Nauwelaerts [Mon, 11 Jan 2010 18:02:34 +0000 (19:02 +0100)]
matroskademux: refactor eos sending when pausing loop

Also, prevent hanging if no pads yet on which to send eos by
posting a message instead.

14 years agomatroskademux: standardize seek handling
Mark Nauwelaerts [Mon, 11 Jan 2010 16:50:35 +0000 (17:50 +0100)]
matroskademux: standardize seek handling

... which implies fixing some corner cases.

14 years agomatroskamux: use more generic xiphN_streamheader_to_codecdata helper
Mark Nauwelaerts [Mon, 11 Jan 2010 14:14:06 +0000 (15:14 +0100)]
matroskamux: use more generic xiphN_streamheader_to_codecdata helper

14 years agomatroskamux: reflow audio and video setcaps and improve logging
Mark Nauwelaerts [Mon, 11 Jan 2010 16:50:04 +0000 (17:50 +0100)]
matroskamux: reflow audio and video setcaps and improve logging

Also ensure width and height are available as they are mandatory
in matroska specs.

14 years agoqtdemux: fix offset for type 2 mp4a sound sample descriptions.
Michael Smith [Mon, 11 Jan 2010 19:42:43 +0000 (11:42 -0800)]
qtdemux: fix offset for type 2 mp4a sound sample descriptions.

Allows us to correctly find the esds (and thus the codec data) for such
mp4a files.

14 years agortpmp4g(de)pay: Only handle raw aac
Thiago Santos [Mon, 11 Jan 2010 18:45:49 +0000 (15:45 -0300)]
rtpmp4g(de)pay: Only handle raw aac

rtpmp4g(de)pay should only handle raw AAC streams

14 years agovideomixer: Implement basic QoS
Sebastian Dröge [Mon, 11 Jan 2010 17:59:43 +0000 (18:59 +0100)]
videomixer: Implement basic QoS

This drops frames if they're too late anyway before blending and all
that starts but QoS events are not forwarded upstream. In the future
the QoS events should be transformed somehow and forwarded upstream.

14 years agortpmp4a(de)pay: Only accept raw aac
Thiago Santos [Mon, 11 Jan 2010 17:48:26 +0000 (14:48 -0300)]
rtpmp4a(de)pay: Only accept raw aac

rtpmp4a(de)pay should only handle raw aac to conform to the RFC

14 years agovideomixer: Add MMX implementations for I420 and all non-alpha RGB formats
Sebastian Dröge [Mon, 11 Jan 2010 17:35:47 +0000 (18:35 +0100)]
videomixer: Add MMX implementations for I420 and all non-alpha RGB formats

14 years agovideomixer: Refactor processing functions
Sebastian Dröge [Mon, 4 Jan 2010 09:24:45 +0000 (10:24 +0100)]
videomixer: Refactor processing functions

This allows easier plugging of optimized processing functions
in the future, like for SSE or AltiVec.

14 years agoavimux: matroskamux: rename aac's stream-format to raw
Thiago Santos [Mon, 11 Jan 2010 16:26:32 +0000 (13:26 -0300)]
avimux: matroskamux: rename aac's stream-format to raw

AAC's none stream-format has been renamed to raw, rename
on avimux and matroskamux as well

14 years agomatroskamux: Only accept raw aac
Thiago Santos [Mon, 11 Jan 2010 15:07:29 +0000 (12:07 -0300)]
matroskamux: Only accept raw aac

makes matroskamux reject aac streams that are not
in raw format (stream-format=none)

Fixes #598350

14 years agoavimux: Only accept raw aac
Thiago Santos [Mon, 11 Jan 2010 15:08:55 +0000 (12:08 -0300)]
avimux: Only accept raw aac

makes avimux reject aac streams that are not
in raw format (stream-format=none)

Fixes #598350

14 years agoqtdemux: Oops. The gpointer cast is needed because of the const
Robert Swain [Mon, 11 Jan 2010 09:38:10 +0000 (10:38 +0100)]
qtdemux: Oops. The gpointer cast is needed because of the const
qualifiers on the data elements

14 years agoqtdemux: Debug -> info level for a message for benchmarking index parsing
Robert Swain [Mon, 11 Jan 2010 09:17:54 +0000 (10:17 +0100)]
qtdemux: Debug -> info level for a message for benchmarking index parsing

The extra message output at higher levels affects the accuracy of the
benchmark.

14 years agoqtdemux: Don't check for NULL pointers or cast to gpointer as this is
Robert Swain [Mon, 11 Jan 2010 09:05:10 +0000 (10:05 +0100)]
qtdemux: Don't check for NULL pointers or cast to gpointer as this is
not needed

14 years agoqtdemux: Refactor stbl sub-atom freeing. Free when index has been
Robert Swain [Fri, 8 Jan 2010 12:55:05 +0000 (13:55 +0100)]
qtdemux: Refactor stbl sub-atom freeing. Free when index has been
completely parsed.

14 years agoqtdemux: Avoid whitespace commits due to inconsistent GNU indent
Robert Swain [Fri, 8 Jan 2010 13:32:06 +0000 (14:32 +0100)]
qtdemux: Avoid whitespace commits due to inconsistent GNU indent
behaviour

14 years agoqtdemux: remove newline at end of debug statement
Tim-Philipp Müller [Mon, 11 Jan 2010 00:10:34 +0000 (00:10 +0000)]
qtdemux: remove newline at end of debug statement

14 years agomultiudpsink: Compiler warning fixes for Windows
Havard Graff [Fri, 8 Jan 2010 18:26:21 +0000 (19:26 +0100)]
multiudpsink: Compiler warning fixes for Windows

Just simple missing casts

Fixes bug #606438.

14 years agoflacenc: fix seekpoints property copy-and-paste documentation
Mark Nauwelaerts [Fri, 8 Jan 2010 17:04:14 +0000 (18:04 +0100)]
flacenc: fix seekpoints property copy-and-paste documentation

14 years agoflacenc: optionally add a seek table
Mark Nauwelaerts [Wed, 6 Jan 2010 16:06:53 +0000 (17:06 +0100)]
flacenc: optionally add a seek table

API: GstFlacEnc:seekpoints

Fixes #351595.

14 years agoavidemux: Use more glib and be safer
Thiago Santos [Fri, 8 Jan 2010 14:33:02 +0000 (11:33 -0300)]
avidemux: Use more glib and be safer

Be safer on sscanf by limiting string format sizes.
Remove useless parameter and use g_strndup.

14 years agoavidemux: Simplifying code
Thiago Santos [Fri, 8 Jan 2010 13:44:44 +0000 (10:44 -0300)]
avidemux: Simplifying code

Greatly simplify the IDIT chunk handling by using sscanf
instead of 'manually' parsing. Also replaces strncasecmp and
is_alpha/is_digit with glib versions.

14 years agoavidemux: it's feb for february
Thiago Santos [Fri, 8 Jan 2010 13:18:30 +0000 (10:18 -0300)]
avidemux: it's feb for february

Fix typo in last commit.

14 years agoavidemux: Parse and post IDIT dates
Thiago Santos [Fri, 8 Jan 2010 12:17:22 +0000 (09:17 -0300)]
avidemux: Parse and post IDIT dates

Parses and post date tags contained in IDIT chunks.

Fixes #503582

14 years agoaudiofxbasefirfilter: Add property for not draining the history on kernel changes
Sebastian Dröge [Thu, 7 Jan 2010 16:25:05 +0000 (17:25 +0100)]
audiofxbasefirfilter: Add property for not draining the history on kernel changes

Currently this only works if the kernel size doesn't change, in the future
it will be possible to change the kernel size too without draining
the complete history and without loosing anything.

Partially based on a patch by
Thiago Santos <thiago.sousa.santos@collabora.co.uk>

14 years agortph264pay: remove weird memcmp code
Wim Taymans [Thu, 7 Jan 2010 15:58:55 +0000 (16:58 +0100)]
rtph264pay: remove weird memcmp code

Use plain memcmp for comparing memory instead of the custom buggy one.

Fixes #606198

14 years agolevel: fix typo in 'message' property description
Edward Hervey [Thu, 7 Jan 2010 14:38:36 +0000 (15:38 +0100)]
level: fix typo in 'message' property description

14 years agoflacdec: really use upstream timestamp if there is one
Mark Nauwelaerts [Wed, 6 Jan 2010 13:06:14 +0000 (14:06 +0100)]
flacdec: really use upstream timestamp if there is one

See/fixes #603471.

14 years agortpg728pay: remove unused adapter peek
Wim Taymans [Wed, 6 Jan 2010 12:45:59 +0000 (13:45 +0100)]
rtpg728pay: remove unused adapter peek

14 years agodeinterlace: Improve passthrough tests
Thiago Santos [Tue, 5 Jan 2010 22:00:35 +0000 (19:00 -0300)]
deinterlace: Improve passthrough tests

Improve passthrough tests by forcing more specific
interlaced/deinterlaced caps to be tested

14 years agodeinterlace: Adds some docs to the new tests
Thiago Santos [Tue, 5 Jan 2010 21:22:49 +0000 (18:22 -0300)]
deinterlace: Adds some docs to the new tests

Adds some docs explaining the utility functions of the check
tests of deinterlace

14 years agodeinterlace: Adds tests for passthrough
Thiago Santos [Tue, 5 Jan 2010 21:14:08 +0000 (18:14 -0300)]
deinterlace: Adds tests for passthrough

Adds tests for checking if the element really does
passthrough in disabled mode and in auto (if the input is
not interlaced)

14 years agodeinterlace: Adds tests for caps acceptance
Thiago Santos [Tue, 5 Jan 2010 10:50:51 +0000 (07:50 -0300)]
deinterlace: Adds tests for caps acceptance

Adds check unit tests for deinterlace for validating
caps accepting and the expected caps output on the
other pad

14 years agodeinterlace: Adds basic check test
Thiago Santos [Mon, 4 Jan 2010 16:43:00 +0000 (13:43 -0300)]
deinterlace: Adds basic check test

Adds a basic check test for deinterlace element

14 years agoqtdemux: Add support for wave-style audio in qt.
Michael Smith [Mon, 4 Jan 2010 23:44:28 +0000 (15:44 -0800)]
qtdemux: Add support for wave-style audio in qt.

Uses gstriff to parse the wave headers appropriately. Tested with MS-ADPCM
content.

14 years agotests: Add G.729 RTP payloader/depayloader test
Olivier Crête [Thu, 31 Dec 2009 22:09:03 +0000 (17:09 -0500)]
tests: Add G.729 RTP payloader/depayloader test

https://bugzilla.gnome.org/show_bug.cgi?id=606050

14 years agortpg729pay: Simplify adapter usage
Olivier Crête [Thu, 31 Dec 2009 21:52:30 +0000 (16:52 -0500)]
rtpg729pay: Simplify adapter usage

https://bugzilla.gnome.org/show_bug.cgi?id=606050

14 years agortpg729pay: Support ptime from caps
Olivier Crête [Thu, 31 Dec 2009 21:27:30 +0000 (16:27 -0500)]
rtpg729pay: Support ptime from caps

https://bugzilla.gnome.org/show_bug.cgi?id=606050

14 years agortp: Add maxptime to the README
Olivier Crête [Wed, 2 Dec 2009 14:05:21 +0000 (19:35 +0530)]
rtp: Add maxptime to the README

https://bugzilla.gnome.org/show_bug.cgi?id=606050

14 years agortpg723depay: add G723 depayloader
Wim Taymans [Tue, 5 Jan 2010 18:03:06 +0000 (19:03 +0100)]
rtpg723depay: add G723 depayloader

14 years agortpg729depay: remove unused variable
Wim Taymans [Tue, 5 Jan 2010 18:02:39 +0000 (19:02 +0100)]
rtpg729depay: remove unused variable

14 years agortpg723pay: rewrite payloader
Wim Taymans [Tue, 5 Jan 2010 17:33:25 +0000 (18:33 +0100)]
rtpg723pay: rewrite payloader

Handle all 3 packet sizes according to RFC 3551.
Totally untested, we don't have a G723 encoder.

Fixes #605882

14 years agoqtdemux: fix chunk counter
Wim Taymans [Tue, 5 Jan 2010 10:47:20 +0000 (11:47 +0100)]
qtdemux: fix chunk counter

14 years agoqtdemux: more work at reducing loop overhead
Wim Taymans [Mon, 4 Jan 2010 18:44:53 +0000 (19:44 +0100)]
qtdemux: more work at reducing loop overhead

Try to avoid derefs when parsing the index. Save the state into the structures
when we exit the loop instead of for each iteration.

14 years agoqtdemux: cleanups and make duration more accurate
Wim Taymans [Mon, 4 Jan 2010 15:33:30 +0000 (16:33 +0100)]
qtdemux: cleanups and make duration more accurate

Make the QtDemuxSample struct smaller by keeping the duration and the pts_offset
as their 32 bit values.
Make some macros to calculate PTS, DTS and duration of a sample.
Deref the sample index less often by keeping a ref to the sample we're dealing
with.

14 years agoqtdemux: simplify logic to calculate duration
Wim Taymans [Mon, 4 Jan 2010 12:41:18 +0000 (13:41 +0100)]
qtdemux: simplify logic to calculate duration

Since we no longer store the timestamp and duration in nanoseconds, we can now
simply store the duration as-is.

14 years agoqtdemux: Store timestamps in mov format in the index
Robert Swain [Fri, 1 Jan 2010 15:42:57 +0000 (16:42 +0100)]
qtdemux: Store timestamps in mov format in the index

This allows faster building of the index upon seeks so that scaling of
timestamps only occurs when actually needed.

14 years agoqtdemux: make seeking in push mode work
Wim Taymans [Fri, 18 Dec 2009 12:54:46 +0000 (13:54 +0100)]
qtdemux: make seeking in push mode work

Move sample position checks into qtdemux_parse_samples where we can protect it
with a lock.
Refactor and make an qtdemux_ensure_index function.
Rename qtdemux_do_push_seek to qtdemux_seek_offset in order to avoid confusion
with gst_qtdemux_do_push_seek.

14 years agoqtdemux: move error code out of normal flow
Wim Taymans [Fri, 18 Dec 2009 11:44:27 +0000 (12:44 +0100)]
qtdemux: move error code out of normal flow

14 years agoqtdemux: Add push mode seek support for seeking to obtain the moov atom
Robert Swain [Tue, 24 Nov 2009 15:27:26 +0000 (16:27 +0100)]
qtdemux: Add push mode seek support for seeking to obtain the moov atom

14 years agortspsrc: fix on-npt-stop signal warnings for RDT
Wim Taymans [Tue, 5 Jan 2010 11:22:09 +0000 (12:22 +0100)]
rtspsrc: fix on-npt-stop signal warnings for RDT

The RDT manager does not implement this signal so we need to check for it before
trying to connect to it.

14 years agov4l2src: fix memory leak in new uri handler code
Tim-Philipp Müller [Tue, 5 Jan 2010 09:47:00 +0000 (09:47 +0000)]
v4l2src: fix memory leak in new uri handler code

Don't leak a string everytime get_uri() is called and a device
has been set. There's a limited number of devices, so just
intern the string instead of doing more elaborate housekeeping
and storing it in the instance struct or so.

14 years agoavimux: fix typo in warning message
Stefan Kost [Fri, 1 Jan 2010 12:10:49 +0000 (14:10 +0200)]
avimux: fix typo in warning message

14 years agoshout2send: Add 'public' property
Robert Weidlich [Mon, 4 Jan 2010 12:28:36 +0000 (09:28 -0300)]
shout2send: Add 'public' property

Adds a property to set 'public' flag on libshout, making
the stream listed on the server's stream directory.

Fixes #605269

14 years agoqtdemux: Add tags for average and maximum bitrate
Arun Raghavan [Wed, 30 Dec 2009 08:44:55 +0000 (14:14 +0530)]
qtdemux: Add tags for average and maximum bitrate

Fixes #599300.

14 years agoaudiofxbasefirfilter: do not try to alloc really large buffers
Thiago Santos [Sat, 26 Dec 2009 19:59:14 +0000 (16:59 -0300)]
audiofxbasefirfilter: do not try to alloc really large buffers

When nsamples_out is larger than nsamples_in, using unsigned
ints lead to a overflow and the resulting value is wrong and
way too large for allocating a buffer. Use signed integers
and returning immediatelly when that happens.

14 years agovideomixer: optimize blend code some more
Wim Taymans [Fri, 25 Dec 2009 11:38:35 +0000 (12:38 +0100)]
videomixer: optimize blend code some more

Use more efficient formula that uses less multiplies.
Reduce the amount of scalar code, use MMX to calculate the desired
alpha value.
Unroll and handle 2 pixels in one iteration for improved pairing.

14 years agovideomixer: scale and clamp
Wim Taymans [Thu, 24 Dec 2009 21:59:09 +0000 (22:59 +0100)]
videomixer: scale and clamp

Scale and clamp to the max alpha values.

14 years agoalpha: scale and clamp alpha to its full extend
Wim Taymans [Thu, 24 Dec 2009 21:50:31 +0000 (22:50 +0100)]
alpha: scale and clamp alpha to its full extend

Convert the alpha value to 0->255 when setting and to 0->256 when using as
a scaling factor. This makes sure we can reach the full opacity value of 0xff in
all cases.

14 years agortspsrc: fix some comments, remove property check
Wim Taymans [Thu, 24 Dec 2009 21:23:01 +0000 (22:23 +0100)]
rtspsrc: fix some comments, remove property check

Fix some comments, clarify some FIXMEs
Remove the on-ntp-stop signal check now that the jitterbuffer is in
-good and we know that it supports this signal.

14 years agovideomixer: some trivial cleanups
Wim Taymans [Thu, 24 Dec 2009 19:27:57 +0000 (20:27 +0100)]
videomixer: some trivial cleanups

14 years agortspsrc: Parse all rtpinfo entries
Thiago Santos [Thu, 24 Dec 2009 20:04:28 +0000 (17:04 -0300)]
rtspsrc: Parse all rtpinfo entries

Do not forget to parse all rtp-info entries, instead of
parsing the first one only.

Fixes #605222

14 years agoqtdemux: perf tag should map to GST_TAG_ARTIST
Thiago Santos [Tue, 22 Dec 2009 15:44:50 +0000 (12:44 -0300)]
qtdemux: perf tag should map to GST_TAG_ARTIST

14 years agointerleave: fix weird indentation
Wim Taymans [Thu, 24 Dec 2009 16:03:02 +0000 (17:03 +0100)]
interleave: fix weird indentation

14 years agortph263ppay: use faster _adapter_copy() whem possible
Wim Taymans [Thu, 24 Dec 2009 16:01:54 +0000 (17:01 +0100)]
rtph263ppay: use faster _adapter_copy() whem possible

14 years agotests: use right type when passing vararg value
Wim Taymans [Thu, 24 Dec 2009 16:01:15 +0000 (17:01 +0100)]
tests: use right type when passing vararg value