Måns Rullgård [Tue, 16 Mar 2010 01:17:00 +0000 (01:17 +0000)]
Move H264 dsputil functions into their own struct
This moves the H264-specific functions from DSPContext to the new
H264DSPContext. The code is made conditional on CONFIG_H264DSP
which is set by the codecs requiring it.
The qpel and chroma MC functions are not moved as these are used by
non-h264 code.
Originally committed as revision 22565 to svn://svn.ffmpeg.org/ffmpeg/trunk
Måns Rullgård [Mon, 15 Mar 2010 19:23:24 +0000 (19:23 +0000)]
Add FATE tests
This adds a "fate" make target which runs the full FATE test suite.
Individual tests can be run with "make fate-$testname".
The location of the FATE test samples must be specified with the
--samples=PATH option to configure.
The tests/fate-update.sh script regenerates the references files and
test list from the online FATE database. These are checked in since
generating them requires non-standard tools.
Originally committed as revision 22552 to svn://svn.ffmpeg.org/ffmpeg/trunk
Martin Storsjö [Mon, 15 Mar 2010 16:36:20 +0000 (16:36 +0000)]
Don't let ff_rtsp_read_reply skip interleaved RTP/TCP packets in rtsp_write_packet.
Skip interleaved packets manually and recheck if there's more to be read.
Originally committed as revision 22548 to svn://svn.ffmpeg.org/ffmpeg/trunk
Martin Storsjö [Mon, 15 Mar 2010 14:20:07 +0000 (14:20 +0000)]
RTSP: Synchronize the start time of the chained RTP muxers
This makes sure that the streams get correctly synchronized when viewed,
previously the streams were out of sync by as much time as it took
between the initialization of the individual muxers.
Originally committed as revision 22545 to svn://svn.ffmpeg.org/ffmpeg/trunk
Micah F. Galizia [Sun, 14 Mar 2010 22:40:16 +0000 (22:40 +0000)]
Move the probe loop from av_open_input_file() into its own method
av_probe_input_buffer() so that it can be reused. Here are a few
differences to the original way things were probed:
- maximum probe buffer size can be specified as a parameter.
- offset within the stream to probe from can be specified as a parameter.
- instead of seeking back to the start each time a probe fails, stream
data is appended to the reallocated buffer. This lowers the amount
of data read from the stream (there is no repetition) and results in
fewer closed and reopened streams (when seeking fails).
New attempt after r22296, which was revert in r22315 due to a FATE
failure.
See the thread:
Subject: [FFmpeg-devel] [PATCH] Move av_open_input_file probe loop to its own method
Date: 2010-03-05 03:23:57 GMT
Patch by Micah F. Galizia printf("%s%s@%s.%s", "micah", "galizia", "gmail", "com").
Originally committed as revision 22532 to svn://svn.ffmpeg.org/ffmpeg/trunk
Stefano Sabatini [Sun, 14 Mar 2010 21:16:29 +0000 (21:16 +0000)]
Change the definition of AVERROR_NUMEXPECTED at the next libavutil
major bump, using an FFmpeg specific error code rather than EDOM,
which has a quite different semantics.
Originally committed as revision 22528 to svn://svn.ffmpeg.org/ffmpeg/trunk
Måns Rullgård [Sun, 14 Mar 2010 17:50:12 +0000 (17:50 +0000)]
Separate DWT from snow and dsputil
This moves the DWT functions from snow.c and dsputil.c to a file of
their own. A new struct, DWTContext, holds the function pointers
previously part of DSPContext.
Originally committed as revision 22522 to svn://svn.ffmpeg.org/ffmpeg/trunk
Reimar Döffinger [Sun, 14 Mar 2010 13:38:00 +0000 (13:38 +0000)]
Avoid creating tiny (possibly only 64 bytes large) audio packets resulting in
huge processing and memory usage overhead for avi files with raw PCM audio.
Originally committed as revision 22521 to svn://svn.ffmpeg.org/ffmpeg/trunk
Stefano Sabatini [Sat, 13 Mar 2010 19:19:40 +0000 (19:19 +0000)]
Make iff.c:decode_init return the value returned by
avctx->get_buffer() in case of error, rather than AVERROR_UNKNOWN
which is deprecated, and mark AVERROR_UNKNOWN for deletion at the next
major bump.
Originally committed as revision 22512 to svn://svn.ffmpeg.org/ffmpeg/trunk
Stefano Sabatini [Sat, 13 Mar 2010 11:27:05 +0000 (11:27 +0000)]
Add AVFilterPicRef.pos field and make libavfilter propagate stream
byte position information, as stored in the pkt.pos, through the
filterchain.
Note that the pos field is added *non* at the end of the
AVFilterPicRef struct, thus breaking ABI compatibility, which is
allowed as the API is still considered non-stable.
Originally committed as revision 22506 to svn://svn.ffmpeg.org/ffmpeg/trunk
Stefano Sabatini [Sat, 13 Mar 2010 10:41:25 +0000 (10:41 +0000)]
Extend the slice filter to make it issue slice height values randomly
choosen between 8 and 32 when the supplied parameter is the string
"random".
This is useful for testing the slice support, but it is not supposed
to be used for other purposes and this interface may change in the
future, thus it is not documented.
The randomization algorithm adopted is the standard Numerical Recipes
LCG.
Originally committed as revision 22505 to svn://svn.ffmpeg.org/ffmpeg/trunk
Diego Biurrun [Fri, 12 Mar 2010 17:29:19 +0000 (17:29 +0000)]
Replace BSD-specific integer types by their POSIX equivalents.
Mark bktr_init function as av_cold.
Tested to work by Michael Kostylev on the NetBSD FATE box.
Originally committed as revision 22487 to svn://svn.ffmpeg.org/ffmpeg/trunk
Diego Biurrun [Fri, 12 Mar 2010 17:28:02 +0000 (17:28 +0000)]
Add _XOPEN_SOURCE=600 to CPPFLAGS on NetBSD.
This make NetBSD compile out of the box at the cost of a hack, but a
locally contained one that is preferable to the one used in the wild.
Currently the NetBSD FATE box adds -D_NETBSD_SOURCE to CPPFLAGS to work
around broken system headers. Since it is unlikely for NetBSD to fix their
headers, it is better to use a standard flag instead of a system-specific one.
As a sideeffect, this allows getting rid of a local _NETBSD_SOURCE definition.
Originally committed as revision 22486 to svn://svn.ffmpeg.org/ffmpeg/trunk
Always reset slice_table.
Previously, the area of a lost slice would be left at the slice number of the previous
frame which could occasionally match the number of the next slice and thus a non existing
slice could have been used for prediction leading to additional decoding errors in otherwise
undamaged slices.
Originally committed as revision 22483 to svn://svn.ffmpeg.org/ffmpeg/trunk
David Conrad [Fri, 12 Mar 2010 05:17:01 +0000 (05:17 +0000)]
oggdec: Pass packets to header() until the stream is done with them
This fixes some old ogm files that had the 3rd vorbis header after a data
packet in another stream. This is invalid in ogg, but this change shouldn't
affect the behaviour of any valid file.
Originally committed as revision 22478 to svn://svn.ffmpeg.org/ffmpeg/trunk