ac3dec: Move center&surround mix level tables to parser.
That way all mix levels as exported by the parser
will have the same meaning.
Previously the 3bit center mix level for eac3 was
used to index in a 4 entry table leading to out of array reads.
this change removes the table and offsets the ac3 variable by 4
so it matches the meanings for eac3 except the reserved case.
The reserved case is then explicitly handled.
Idea-by: Justin Ruggles <justin.ruggles@gmail.com> Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Ray Simard [Sat, 21 Jan 2012 02:19:46 +0000 (18:19 -0800)]
deshake: variable used uninitialized
Sometimes the scan finds nothing that qualifies for addition to
the array and pos is zero after the loops. The code forces pos to
1 and the array is then processed as if it had one valid element in it,
producing some amusing but not very useful results.
I don't see the rationale for this. If pos is zero coming out of the
loops, the only appropriate thing to do is set t->angle to zero. The
attached patch does that. It's worked properly in several tests so far.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
* qatar/master:
Add a tool that uses avio to read and write, doing a plain copy of data
ARM: fix build with FFT enabled and MDCT disabled
lavf: force single-threaded decoding in avformat_find_stream_info
avidec: migrate last of lavf from FF_ER_* to AV_EF_*
avserver: fix build after the next bump.
Janne Grunau [Wed, 18 Jan 2012 19:32:32 +0000 (20:32 +0100)]
lavf: force single-threaded decoding in avformat_find_stream_info
The H.264 decoder needs SPS and PPS for initialization during
multi-threaded decoding. When probed single-threaded SPS and PPS are
copied to extradata and are available for proper initialization of
the decoder before the first frame is decoded.
* qatar/master:
mpeg12: check for available bits to avoid an infinite loop
fate: add some shorthands to run groups of tests
fate: Give some tests more sensible names.
cosmetics: Rename ffsink to avsink.
pp: bump major in git master to match what debian contains.
The ABI differences are negligible, but its easier for all if
all distros have libpostproc HEAD under the same soname and
debian bumped soname without consulting upstream, so as silly as
it is following this is probably the least pain for all.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
lavc: call ff_init_buffer_info() from default reget buffer.
This is needed in case the get_buffer() callback doesnt set
width/height.
Ideally all decoders would make calls through some wraper
to the callbacks and that wraper would call ff_init_buffer_info()
But until thats done, the default reget buffer must call this
itself as it needs the values for the changed size check later.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Stefano Sabatini [Tue, 17 Jan 2012 23:22:17 +0000 (00:22 +0100)]
ffprobe: use safer type for index variable in value_string()
There is no guarantee that the casted double which is assigned to the
variable will be contained in an int (also if it is almost sure for most
non-alien architectures).
Stefano Sabatini [Tue, 17 Jan 2012 23:01:07 +0000 (00:01 +0100)]
ffprobe: fix printing of unit values which cannot be contained in an int
Use long long int to contain such values instead of an int, which is
required to contain at least 64 bits, so it is guaranteed to contain also
int64_t values, which are used by some fields.
Reimar Döffinger [Mon, 16 Jan 2012 23:10:17 +0000 (00:10 +0100)]
ffmpeg: check tcgetattr result.
This fixes parallel FATE (make fate -j4) failing under valgrind with:
Syscall param ioctl(TCSET{S,SW,SF}) points to uninitialised byte(s)
at 0x5D98B23: tcsetattr (tcsetattr.c:88)
by 0x43D66C: term_init (ffmpeg.c:734)
by 0x43CD8D: main (ffmpeg.c:5071)
Address 0x7fefffdd0 is on thread 1's stack
Uninitialised value was created by a stack allocation
at 0x43D5B0: term_init (ffmpeg.c:716)
Reimar Döffinger [Sun, 15 Jan 2012 13:26:55 +0000 (14:26 +0100)]
Add and use av_fast_padded_malloc.
The same as av_fast_malloc but uses av_mallocz and keeps extra
always-0 padding.
This does not mean the memory will be 0-initialized after each call,
but actually only after each growth of the buffer.
However this makes sure that
a) all data anywhere in the buffer is always initialized
b) the padding is always 0
c) the user does not have to bother with adding the padding themselves
Fixes another valgrind warning about use of uninitialized data,
this time with fate-vsynth1-jpegls.
Stefano Sabatini [Sun, 15 Jan 2012 22:05:07 +0000 (23:05 +0100)]
tools: add ffeval tool
The tool is useful for testing the internal arithmetic evaluation engine
(indeed I plan to use it in FATE), and provides a handy calculator when
you can't rely on bc ;-).
* qatar/master:
rv34: add NEON rv34_idct_add
rv34: 1-pass inter MB reconstruction
add SMJPEG muxer
avformat: split out common SMJPEG code
pictordec: Use bytestream2 functions
avconv: use avcodec_encode_audio2()
pcmenc: use AVCodec.encode2()
avcodec: bump minor version and add APIChanges for the new audio encoding API
avcodec: Add avcodec_encode_audio2() as replacement for avcodec_encode_audio()
avcodec: add a public function, avcodec_fill_audio_frame().
rv34: Intra 16x16 handling
rv34: Inter/intra MB code split
Reimar Döffinger [Sun, 11 Dec 2011 15:02:30 +0000 (16:02 +0100)]
Rework VP8 decoder frame skipping code.
This reworks the frame skipping code such that the reference
buffers are still updated according to the header.
However it also ensures that the current frame will not end
up in any reference buffer.
Also fixes a hang with frame-multithreading, probably because
get_buffer was already called and would have reset the progress,
however the frame could remain in framep due to the missing update
(or it could be assigned to next_framep and a skip_frame skip would
then write it into framep - there might be even more failure modes).
Sample might become available at samples/nsv/vp8.nsv