Commit Graph

5584 Commits

Author SHA1 Message Date
Diego Biurrun
b146873c94 VIS-specific code should be compiled conditional to HAVE_VIS, not ARCH_SPARC.
Originally committed as revision 10554 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-23 21:02:22 +00:00
Kostya Shishkov
ac14047927 Fix decoding of VMDs representing sprites (Last Dynasty, Woodruff).
Fixes issue 101

Originally committed as revision 10552 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-23 16:11:07 +00:00
Reimar Döffinger
61858a76f2 Apply skip_loop_filter before checking if we can parallelize with the selected
loop filter

Originally committed as revision 10549 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-22 10:28:22 +00:00
Kostya Shishkov
a2085a7e9d Guard against output buffer overflows
Originally committed as revision 10548 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-22 09:21:43 +00:00
Kostya Shishkov
675a0583b4 Do not decode more data than output buffer may hold
Originally committed as revision 10547 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-22 09:09:47 +00:00
Baptiste Coudurier
32bbb99d0a HDYC fourcc, sample hdyc/Test2.avi
Originally committed as revision 10541 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-21 18:23:46 +00:00
Andreas Öman
a27d6e714f Simplify H.264 decode_cabac_mb_cbp_luma(), giving a ~0.5% speedup.
patch by Andreas Öman, andreas olebyn nu
Date: Thu, 20 Sep 2007 12:59:19 +0200
Subject: [FFmpeg-devel] [PATCH] simplify h264's decode_cabac_mb_cbp_luma()

Originally committed as revision 10537 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-21 00:26:31 +00:00
Aurelien Jacobs
dee15eceae allcodecs.h and allformats.h are not used anymore
Originally committed as revision 10536 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-20 20:42:50 +00:00
Aurelien Jacobs
d999cff5a3 remove last dependencies on allformats.h and allcodecs.h
Originally committed as revision 10535 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-20 20:36:15 +00:00
Ivan Kalvachev
ee28dedc86 Remove inline from function that is exported.
Issue noticed by Bernhard Rosenkraenzer.

Originally committed as revision 10529 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-18 18:12:53 +00:00
Marc Hoffman
584a3ac907 blackfin vsad primitives
Originally committed as revision 10521 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-17 13:23:16 +00:00
Marc Hoffman
56261c6150 properly tell the compiler that A0 and A1 are clobbered
Originally committed as revision 10520 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-17 13:17:50 +00:00
Diego Biurrun
4981218420 Remove redundant ARCH_SPARC #ifdef, the file is compiled conditional to
ARCH_SPARC anyway.

Originally committed as revision 10517 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-17 10:14:56 +00:00
Aurelien Jacobs
84a72b1675 remove obsolete and wrong extern declaration
Originally committed as revision 10512 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-16 22:50:12 +00:00
Diego Biurrun
1fe2d0fd1a cosmetics: typo
Originally committed as revision 10507 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-16 16:23:34 +00:00
Andreas Öman
50c21814b9 fix image corruption when with multi-threaded decoding.
dequant-tables were not correctly reinitialized in the slave
contexts when a PPS came with updated matrices.
Patch by Andreas Öman %andreas A olebyn P nu%
Original thread:
date: Sep 16, 2007 6:14 AM
subject: [FFmpeg-devel] Parallelized h264 image corruption bug

Originally committed as revision 10505 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-16 10:37:07 +00:00
Kostya Shishkov
2a898132a2 Enable error resilience in WMV3/VC-1 decoder
Originally committed as revision 10502 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-15 11:06:27 +00:00
Kostya Shishkov
a9d78ac51a HALFPQ should be used only for when PQUANT is selected
Originally committed as revision 10500 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-15 09:24:53 +00:00
Kostya Shishkov
f7d2a43707 In case when any quantizer may occur, HALFPQ should be zero
Originally committed as revision 10499 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-15 09:23:58 +00:00
Justin Ruggles
a135bea507 skip incomplete frames
Originally committed as revision 10497 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-15 02:44:28 +00:00
Justin Ruggles
3df880934e better AC3 header error reporting
Originally committed as revision 10496 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-15 02:41:24 +00:00
Justin Ruggles
d5b7144e1f fix random dithering of zero-bit mantissas
Originally committed as revision 10495 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-15 00:43:40 +00:00
Justin Ruggles
bfcf690c68 cosmetics: rename variables. tbl->tab for consistency with other AC3 tables
Originally committed as revision 10494 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-15 00:19:21 +00:00
Justin Ruggles
396a6ee50e fix decoding of DolbyNet AC3
Originally committed as revision 10493 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-15 00:00:57 +00:00
Kostya Shishkov
f0dd9d4505 Check unp_size for possible overflows too
Originally committed as revision 10490 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-14 06:01:29 +00:00
Diego Biurrun
b3e4d266f4 Pass correct CPU flags to compilation of sparc/simple_idct_vis.o.
based on a patch by Denes Balatoni, dbalatoni interware hu

Originally committed as revision 10486 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-13 09:13:56 +00:00
Kostya Shishkov
e938637b2c Add checks on input/output buffers size for some audio decoders
Originally committed as revision 10485 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-13 05:59:58 +00:00
Kostya Shishkov
bf4a1f17ee Monkey Audio decoder
Originally committed as revision 10484 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-13 03:22:47 +00:00
Michael Niedermayer
f612801465 fix timestamps used for ratecontrol
these were wrong (in pts vs dts sense) when b frames were in use
they were also wrong if the average framerate was smaller than 1/timebase
resulting in totally wrong final bitrate

Originally committed as revision 10477 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-11 22:46:41 +00:00
Michael Niedermayer
b9b7472eb5 cosmetic
Originally committed as revision 10465 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-10 00:31:27 +00:00
Michael Niedermayer
8db1372814 make changing spatial_decomposition_count per frame work
Originally committed as revision 10464 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 18:28:36 +00:00
Reimar Döffinger
ba430e0cfa dsputil must be initialized before calling rtjpeg init.
Originally committed as revision 10463 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 18:14:11 +00:00
Michael Niedermayer
dbd6a1fee8 do not ignore return of decode_header()
Originally committed as revision 10461 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 17:01:11 +00:00
Michael Niedermayer
b85bf991f6 do not force the halfpel filter coeffs to be retransmitted on every frame
if always_reset is set

Originally committed as revision 10460 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 16:54:00 +00:00
Michael Niedermayer
e559c30a20 allow changing of the spatial_decomposition_count and quantization tables
per frame (this is untested as the encoder does not use it yet)

Originally committed as revision 10458 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 15:06:46 +00:00
Michael Niedermayer
4f90f33ab4 perform init after reading the values needed for init
fixes decoding of files with spatial_decomposition_count!=5

Originally committed as revision 10457 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 13:47:25 +00:00
Michael Niedermayer
081c9d1d7b unneeded and impossible during init ...
Originally committed as revision 10455 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 12:46:25 +00:00
Michael Niedermayer
32272f7e95 unused
Originally committed as revision 10454 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 12:36:10 +00:00
Michael Niedermayer
014f9d233f allocate slice buffer after parsing the header containing needed parameters
argh i really wish i had never accepted these slice patches ...

Originally committed as revision 10453 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 12:33:34 +00:00
Michael Niedermayer
ba8964db60 remove ()
Originally committed as revision 10452 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 12:25:12 +00:00
Michael Niedermayer
61806e9fa7 spatial_decomposition_count != 5 bugfix
Originally committed as revision 10451 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 12:24:45 +00:00
Michael Niedermayer
fa731ccd4e scaling obmc tables under #if 0 also by 4 (yes they where forgotten apparently)
this also requires a few entries to be ++ and -- to avoid a max of 256

Originally committed as revision 10450 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 12:22:11 +00:00
Reimar Döffinger
1551e8e87f Properly handle nuv file with changing resolution
Originally committed as revision 10448 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 10:20:47 +00:00
Reimar Döffinger
fd065d0593 Make sure rtjpeg quality is a valid value
Originally committed as revision 10444 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 09:16:47 +00:00
Reimar Döffinger
6cf7e81f8c Fix extreme stupidity: buffer size check must of course be before buffer use!
Originally committed as revision 10443 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 09:00:12 +00:00
Reimar Döffinger
048905a484 10l, add check forgotten in last commit
Originally committed as revision 10442 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 08:52:53 +00:00
Reimar Döffinger
4f8a91c961 First ugly and slow attempt to fix nuv files with extra frameheader
and per-frame rtjpeg quality

Originally committed as revision 10441 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 08:42:49 +00:00
Reimar Döffinger
c612b00d7e Simplify nuv: factor out LZO decompression
Originally committed as revision 10440 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-09 08:22:07 +00:00
Michael Niedermayer
61d6e445b1 rename HTAPS -> HTAPS_MAX (later is correct after the previous change)
Originally committed as revision 10437 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-08 14:52:21 +00:00
Michael Niedermayer
7d7f57d939 store halfpel filter coefficients in the header as well as the
flag for diagonal interpolation

the primary reason for this change is that previously MC up to 1/4 pel
matched H.264 exactly and that increases the risk of stumbling over
patents

secondly this allows 0.10 db or more quality gain by choosing a longer
filter and the filter could also be chosen optimally for each frame
though that of course would cause speed loss at the decoder and encoder
side ...

Originally committed as revision 10436 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-08 14:51:13 +00:00