Commit Graph

5487 Commits

Author SHA1 Message Date
Guillaume Poirier
16dae5173b remove alignment correction of the destination pointers in luma_16x6
interpolations, since they are always 16-bytes aligned in practice.
Add asserts to ease narrowing down potential image corructions on exotic plateforms

Based on a patch by Mauricio Alvarez % lokifo A gmail P com %
Original thread:
Date: Jun 26, 2007 1:07 PM
Subject: Re: [FFmpeg-devel] [PATCH] h264 luma interpolation 8x8 for altivec

Originally committed as revision 10238 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-26 21:47:58 +00:00
Ramiro Polla
7bcc1d5b66 CONFIG_7REGS has been renamed to HAVE_7REGS
Originally committed as revision 10237 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-26 16:10:53 +00:00
Michael Niedermayer
90e9e94d0f workaround gcc bug, untested as my gcc is not complaining
Originally committed as revision 10236 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-26 12:34:22 +00:00
Michael Niedermayer
cefa599953 optimize the first vertical lifting step, this also prevents another
overflow, the last known possible overflow

Originally committed as revision 10234 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-26 11:16:23 +00:00
Michael Niedermayer
c90762766a optimize 1st horizontal lifting step
Originally committed as revision 10231 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-26 08:31:55 +00:00
Michael Niedermayer
1104bf2b20 typo
Originally committed as revision 10230 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-26 08:03:23 +00:00
Michael Niedermayer
8b5029295d get rid of totally senseless "m" + read in register we have enough
registers to keep everything in registers

Originally committed as revision 10229 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-26 06:51:51 +00:00
Michael Niedermayer
bc1e78d896 simplify senselessly complex addressing
Originally committed as revision 10228 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-26 02:02:14 +00:00
Michael Niedermayer
25bb359f73 cosmetics
remove brain amputated mmx wrappers around sse2 macros
fix name of ..._sub macro to match ..._add naming

Originally committed as revision 10227 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-26 01:20:09 +00:00
Michael Niedermayer
629750290f avoid overflow in the 3rd lifting step, this now needs mmx2 at minimum
(patch for plain mmx support is welcome ...)

Originally committed as revision 10226 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-26 01:11:02 +00:00
Michael Niedermayer
b696a4c91d avoid an overflow in the 1 horizontal lifting step
Originally committed as revision 10225 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-25 19:04:33 +00:00
Michael Niedermayer
9caa1cccb9 prevent one overflow in the first vertical lifting step
Originally committed as revision 10224 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-25 16:28:45 +00:00
Michael Niedermayer
3e0f7126b5 update mmx code to latest snow changes
note, the code likely can overflow and thus needs some more changes
sse2 updated too but disabled as it is untested

Originally committed as revision 10223 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-25 15:20:56 +00:00
Diego Biurrun
273b0cd3c4 prettyprinting cosmetics
Originally committed as revision 10221 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-25 09:23:17 +00:00
Michael Niedermayer
ff06e0677b cleanup (remove some old experimentation related code)
Originally committed as revision 10220 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-25 03:19:39 +00:00
Michael Niedermayer
b5b88eb3b4 we only have 2 wavelets, the 3rd was just for experimentation ...
Originally committed as revision 10219 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-25 03:04:13 +00:00
Michael Niedermayer
d593e32983 use 16bit IDWT (a SIMD implementation of it should be >2x faster then with
the old 32bit code)
disable mmx/sse2 optimizations as they need a rewrite now

Originally committed as revision 10218 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-25 03:00:51 +00:00
Diego Biurrun
830bf1f20d whitespace/indentation cosmetics
Originally committed as revision 10217 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-24 23:49:11 +00:00
Diego Biurrun
2722c98128 Remove HAVE_ALTIVEC preprocessor directives that are only compiled
when AltiVec is enabled anyway.

Originally committed as revision 10216 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-24 23:01:50 +00:00
Diego Biurrun
f9edc2331d Fix trivial mixed declarations and code warning caused by a double semicolon.
Originally committed as revision 10212 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-24 15:42:50 +00:00
Diego Biurrun
e4506f0c98 Remove SYS_DARWIN preprocessor directive that is just a duplicate of a
gcc-specific directive a few lines below.

Originally committed as revision 10211 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-24 15:38:37 +00:00
Diego Biurrun
4b180bbd73 Replace SYS_DARWIN by the more correct __APPLE_CC__, these preprocessor
directives are aimed at idiosyncracies of Apple's gcc version.

Originally committed as revision 10210 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-24 15:18:41 +00:00
Diego Biurrun
b1312aa24d Merge __MWERKS__ preprocessor directives.
Originally committed as revision 10208 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-24 15:08:46 +00:00
Diego Biurrun
8008a04325 Move Apple gcc AltiVec vector declaration syntax to libavutil.
Originally committed as revision 10207 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-24 15:04:00 +00:00
Diego Biurrun
08f87a9c99 Replace SYS_DARWIN by the more correct __APPLE_CC__, these preprocessor
directives are aimed at idiosyncracies of Apple's gcc version.

Originally committed as revision 10206 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-24 14:57:46 +00:00
Michael Niedermayer
be693f956e int->uint8_t
Originally committed as revision 10205 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-24 14:18:09 +00:00
Michael Niedermayer
50464a1e83 int->int8_t (no need for them to be int ...)
Originally committed as revision 10204 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-24 14:02:02 +00:00
Reimar Döffinger
b1ec601f7f Add request_channels member to AVCodecContext so we now have a proper
way to tell the decoder how many output channels we would like.

Originally committed as revision 10199 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-23 20:28:28 +00:00
Michael Niedermayer
b538791bc9 use more bits on the encoder side
negligible increase in quality

Originally committed as revision 10195 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-23 10:49:14 +00:00
Luca Abeni
e02f7490da Initialize the "stc" variable before using it in ff_find_start_code()
Originally committed as revision 10193 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-23 07:02:27 +00:00
Diego Biurrun
9d16f87ffd Rename CONFIG_DARWIN to SYS_DARWIN, it is not configurable (in FFmpeg).
Originally committed as revision 10190 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-22 22:46:39 +00:00
Michael Niedermayer
81717747ab simplify
Originally committed as revision 10186 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-22 17:26:49 +00:00
Benoit Fouet
e2d0eeecf7 right filename in doxygen comment
Originally committed as revision 10185 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-22 15:03:14 +00:00
Diego Biurrun
e9bfae45bb cosmetics: alphabetical order
Originally committed as revision 10184 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-22 13:19:10 +00:00
Diego Biurrun
d55388d028 Simplify preprocessor directives.
Originally committed as revision 10181 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-22 13:04:10 +00:00
Marc Hoffman
18b8b43b4f make clean bfin
Originally committed as revision 10180 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-22 12:54:49 +00:00
Diego Biurrun
57b499c7da Identifiers starting with underscores are reserved.
Originally committed as revision 10179 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-22 12:41:16 +00:00
Aurelien Jacobs
5ab657079d parsers and bsfs declarations are not meant to be part of public API
Originally committed as revision 10173 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-21 18:24:08 +00:00
Michael Niedermayer
ce611a27be Change rounding of the horizontal DWT to match the vertical one.
This allows some simplifications and optimizations and should
not have any effect on quality.

Originally committed as revision 10172 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-21 16:29:40 +00:00
Aurelien Jacobs
289a723ce3 asv1.c directly uses Put/GetBitContext, so it should include bitstream.h
Originally committed as revision 10169 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-21 15:48:08 +00:00
Michael Niedermayer
30cd3e66bd remove code which become unused by the previous changes
Originally committed as revision 10166 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-21 00:05:30 +00:00
Michael Niedermayer
72dee89b5d Simplify and optimize the 4th vertical lifting step of the SSE2 code (untested)
This also reduces the needed headroom in that step by 1 bit

Originally committed as revision 10165 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-21 00:03:18 +00:00
Michael Niedermayer
d0dae46a90 Simplify and optimize the 4th vertical lifting step of the MMX code
This also reduces the needed headroom in that step by 1 bit

Originally committed as revision 10164 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-21 00:02:19 +00:00
Michael Niedermayer
1ffbbef217 Simplify and speedup code, reduce needed headroom by 2 bits in the 3rd
vertical lifting step of the SSE2 code (untested)

Originally committed as revision 10163 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-20 23:59:20 +00:00
Michael Niedermayer
4bf1790421 simplify, speedup and reduce needed headroom by 2 bits in the 3rd
vertical lifting step

Originally committed as revision 10162 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-20 23:54:49 +00:00
Michael Niedermayer
dd30437bbe replace <<1 by add for SSE2 (untested)
Originally committed as revision 10161 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-20 23:11:05 +00:00
Michael Niedermayer
7e665a3943 replace <<1 by add
Originally committed as revision 10160 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-20 23:09:39 +00:00
Michael Niedermayer
eee649d3b5 slightly change horizontal lift3 so it needs 1 bit less headroom
Originally committed as revision 10159 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-20 23:02:59 +00:00
Michael Niedermayer
be3b22f9ca remove idiotc double subtraction from the sse2 code (untested, no sse2 here)
Originally committed as revision 10158 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-20 22:41:47 +00:00
Michael Niedermayer
8510c519e7 fixme note
Originally committed as revision 10157 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-08-20 22:34:42 +00:00