Commit Graph

51 Commits

Author SHA1 Message Date
Mans Rullgard
5da7494dc5 ARM optimised vp56_rac_get_prob()
Approximately 3% faster on Cortex-A8.

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 7da48fd011)
2011-02-13 00:52:51 +01:00
Jason Garrett-Glaser
05c04cdf54 VP5/6/8: ~7% faster arithmetic decoding
Grab from the bitstream in 16-bit chunks instead of 8-bit chunks.
TODO: grab in 32-bit chunks on 64-bit systems.

Originally committed as revision 24783 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-08-12 01:11:32 +00:00
Jason Garrett-Glaser
905ef0d064 VP5/6/8: eliminate CABAC dependency
Create a custom table for VP5/6/8's renorm to avoid depending on H.264's.
Saves one instruction in the arithmetic decoder as well.

Originally committed as revision 24701 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-08-04 23:04:05 +00:00
Jason Garrett-Glaser
61ee75bbf8 Silence unused function warnings in vp56.h
Originally committed as revision 24684 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-08-03 11:35:41 +00:00
Jason Garrett-Glaser
cd2769c142 VP5/6/8: tweak some arithcoder inlining
Always inline the arithmetic coder, except in the case of header-parsing stuff,
in which case don't inline it at all to save code size.

Originally committed as revision 24677 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-08-03 08:06:08 +00:00
Alex Converse
52fa20542b Reindent after last commit.
Originally committed as revision 24562 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-07-28 05:38:30 +00:00
Alex Converse
d950497083 ff_prefix non static vp56 functions.
Originally committed as revision 24561 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-07-28 05:36:33 +00:00
Eli Friedman
3611e7a309 Inline asm for VP56 arith coder
This is a lot more reliable to get cmov rather than trying to trick gcc into
generating it, useful since it's 2% faster overall.

Patch by Eli Friedman <eli.friedman at gmail>

Originally committed as revision 24471 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-07-23 21:46:30 +00:00
David Conrad
fe1b5d974a Decode DCT tokens by branching to a different code path for each branch
on the huffman tree, instead of traversing the tree in a while loop.

Based on the similar optimization in libvpx's detokenize.c

10% faster at normal bitrates, and 30% faster for high-bitrate intra-only

Originally committed as revision 24468 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-07-23 21:46:17 +00:00
David Conrad
5474ec2ac8 Move renormalization of the VP56 arith decoder to before decoding a bit
No difference at the moment, but allows a future branchy variant
of vp56_rac_get_prob to be significantly faster

Originally committed as revision 24467 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-07-23 21:46:14 +00:00
David Conrad
b3d755ec8b Split renorm of vp56 arith decoder to its own function
Originally committed as revision 24466 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-07-23 21:46:08 +00:00
David Conrad
24675b8093 vp56's arith decoder's code_word is only 16 bits, no need for unsigned long
Originally committed as revision 24465 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-07-23 21:46:01 +00:00
Jason Garrett-Glaser
702e8d3376 Much faster VP8 mv and mode prediction
Originally committed as revision 24412 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-07-22 04:26:41 +00:00
Jason Garrett-Glaser
d229ae2b62 Convert vp56_mv to 16-bit.
Saves nothing except a bit of memory/cache now, but will allow future
optimizations.

Originally committed as revision 24411 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-07-22 03:33:29 +00:00
Aurelien Jacobs
6eabaef4b8 cosmetic: improve comment breaking at 80 cols
Originally committed as revision 23938 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-07-01 16:28:38 +00:00
Jason Garrett-Glaser
4148855ee4 Eliminate another redundant instruction in vp56/8 arithcoder
Necessary because of this GCC bug:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44474
To do this, convert some, but not all (!) of the variables in VP56RangeCoder
into local variables.
If we convert c->high into a local variable, gcc gets the stupids and refuses
to use a conditional move for the unpredictable main branch.

TODO: dispense with this bullshit and write an asm version.

Originally committed as revision 23924 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-06-30 23:59:27 +00:00
Jason Garrett-Glaser
36d6b545a1 CMOV-ify vp56 arithcoder
This incantation causes gcc 4.3 to generate cmov on x86, a vastly better option
than a completely unpredictable branch.
Hopefully this carries over to newer versions and other CPUs with conditionals.
~5 cycles saved per call on a Core i7.

Originally committed as revision 23921 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-06-30 23:18:47 +00:00
Jason Garrett-Glaser
2e6ed48d6e Optimize vp56 arithmetic decoder
Negate "bits" to eliminate a negate in cache refilling.

Originally committed as revision 23920 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-06-30 23:15:25 +00:00
Stefan Gehrer
bce3bd1ced renormalize VP5/6/7/8 range coder without loop
Originally committed as revision 23915 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-06-30 22:05:29 +00:00
David Conrad
3b636f21da Native VP8 decoder.
Patch by David Conrad <lessen42 gmail com> and myself.

Originally committed as revision 23719 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-06-22 19:24:09 +00:00
Ronald S. Bultje
c6ef6e14cf Change a / 256 into a >> 8.
Originally committed as revision 23718 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-06-22 19:19:13 +00:00
Måns Rullgård
5e1ba34bbb VP56: move vp56_edge_filter to new VP56DSPContext
Using macro templates allows the vp[56]_adjust functions to be
inlined instead of called through function pointers.  The new
function pointers enable optimised implementations of the filters.

4% faster VP6 decoding on Cortex-A8.

Originally committed as revision 22992 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-04-30 21:30:22 +00:00
Diego Biurrun
ba87f0801d Remove explicit filename from Doxygen @file commands.
Passing an explicit filename to this command is only necessary if the
documentation in the @file block refers to a file different from the
one the block resides in.

Originally committed as revision 22921 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-04-20 14:45:34 +00:00
Måns Rullgård
84dc2d8afa Remove DECLARE_ALIGNED_{8,16} macros
These macros are redundant.  All uses are replaced with the generic
DECLARE_ALIGNED macro instead.

Originally committed as revision 22233 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-03-06 14:24:59 +00:00
Reimar Döffinger
800274f871 Make VP5 and VP6 decoders output a qscale table to allow for more automatic
post-processing, and add a new FF_QSCALE_TYPE_VP56 for this.

Originally committed as revision 21529 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-01-29 19:11:47 +00:00
Måns Rullgård
c67278098d Move array specifiers outside DECLARE_ALIGNED() invocations
Originally committed as revision 21377 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-01-22 03:25:11 +00:00
Aurelien Jacobs
7576516a7e vp56dec: ensure range coder won't read past the end of input buffer
Originally committed as revision 19348 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-05 15:23:42 +00:00
Stefano Sabatini
9106a698e7 Rename bitstream.h to get_bits.h.
Originally committed as revision 18494 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-04-13 16:20:26 +00:00
Aurelien Jacobs
012f930855 fix vp5/vp6 decoding by using new prototype for decode function
Originally committed as revision 18352 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-04-07 18:24:47 +00:00
Diego Biurrun
bad5537e2c Use full internal pathname in doxygen @file directives.
Otherwise doxygen complains about ambiguous filenames when files exist
under the same name in different subdirectories.

Originally committed as revision 16912 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-02-01 02:00:19 +00:00
Aurelien Jacobs
d887151d8c rename Vp56Model to VP56Model for consistency
Originally committed as revision 16073 to svn://svn.ffmpeg.org/ffmpeg/trunk
2008-12-12 00:51:53 +00:00
Aurelien Jacobs
76025d91da cosmetics: vertical alignment
Originally committed as revision 16069 to svn://svn.ffmpeg.org/ffmpeg/trunk
2008-12-11 23:00:58 +00:00
Aurelien Jacobs
3d52bca6bf avoid POSIX reserved _t suffix
Originally committed as revision 16068 to svn://svn.ffmpeg.org/ffmpeg/trunk
2008-12-11 22:56:19 +00:00
Stefano Sabatini
987903826b Globally rename the header inclusion guard names.
Consistently apply this rule: the guard name is obtained from the
filename by stripping the leading "lib", converting '/' and '.'  to
'_' and uppercasing the resulting name. Guard names in the root
directory have to be prefixed by "FFMPEG_".

Originally committed as revision 15120 to svn://svn.ffmpeg.org/ffmpeg/trunk
2008-08-31 07:39:47 +00:00
Aurelien Jacobs
d37f007d54 remove #include "mpegvideo.h" where it is not needed
Originally committed as revision 12321 to svn://svn.ffmpeg.org/ffmpeg/trunk
2008-03-05 00:07:53 +00:00
Michael Niedermayer
7e5cdb5b1b const
Originally committed as revision 11791 to svn://svn.ffmpeg.org/ffmpeg/trunk
2008-02-01 16:16:40 +00:00
Diego Biurrun
5b21bdabe4 Add FFMPEG_ prefix to all multiple inclusion guards.
Originally committed as revision 10765 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-10-17 09:37:46 +00:00
Aurelien Jacobs
f28b10483a add support for VP6 with huffman encoded blocks
closes issue 104

Originally committed as revision 10737 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-10-14 22:30:59 +00:00
Aurelien Jacobs
91fc2cf150 add a new vp6a codec (add alpha plan support to vp6)
Originally committed as revision 10574 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-25 19:12:46 +00:00
Aurelien Jacobs
247df384ec move all model related tables into their own struct
Originally committed as revision 10571 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-24 22:29:11 +00:00
Aurelien Jacobs
1457516f70 some simplifications and uniformisation
Originally committed as revision 10568 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-09-24 20:43:47 +00:00
Aurelien Jacobs
977f6a9049 remove wrong const qualifier
Originally committed as revision 9795 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-07-25 19:34:31 +00:00
Diego Biurrun
e5a389a1b7 license header consistency cosmetics
Originally committed as revision 9484 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-07-05 10:40:25 +00:00
Ramiro Polla
2c124cb65c Use AV_xx throughout libavcodec
Originally committed as revision 9169 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-06-02 01:41:07 +00:00
Aurelien Jacobs
704a2881bc Use pointers to avoid copying AVFrame.
This ensure get_buffer() won't be confused with an AVFrame pointer
which looks like it wasn't released.

Originally committed as revision 8124 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-02-25 15:33:38 +00:00
Aurelien Jacobs
dd9b86354c add support for another variant of vp6
with block coeffs coded separatly from other parts of the frame

Originally committed as revision 7484 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-01-14 18:17:15 +00:00
Aurelien Jacobs
9110a0e32d Add support for VP60 and VP61.
Originally committed as revision 7329 to svn://svn.ffmpeg.org/ffmpeg/trunk
2006-12-17 22:00:12 +00:00
Steve L'Homme
33400850cf Don't include inttypes.h/stdint.h directly, there is an abstraction in
libavutil/common.h for platforms with no inttypes.h.
patch by Steve Lhomme, slhomme divxcorp com

Originally committed as revision 6790 to svn://svn.ffmpeg.org/ffmpeg/trunk
2006-10-24 18:57:43 +00:00
Diego Biurrun
b78e7197a8 Change license headers to say 'FFmpeg' instead of 'this program/this library'
and fix GPL/LGPL version mismatches.

Originally committed as revision 6577 to svn://svn.ffmpeg.org/ffmpeg/trunk
2006-10-07 15:30:46 +00:00
Aurelien Jacobs
4fde50604d rename vector to vect to avoid clash with Apple gcc
Originally committed as revision 6331 to svn://svn.ffmpeg.org/ffmpeg/trunk
2006-09-25 18:54:18 +00:00