git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21196 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21195 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21194 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21193 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21192 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Up to 6% faster overall on i7, no change on A8.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21191 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

GCC should be able to do this, but doesn't.  Slightly faster code.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21190 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Simplify cur_band_type, group_len, and coef/offset calculations.  This
makes the code easier to read and slightly faster.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21189 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

The codebooks each consist of small number of values repeated in
groups of 2 or 4.  Storing the codebooks as a packed list of 2- or
4-bit indexes into a table reduces their size substantially (from 7.5k
to 1.5k), resulting in less cache pressure.

For the band types with sign bits in the bitstream, storing the number
and position of non-zero codebook values using a few bits avoids
multiple get_bits() calls and floating-point comparisons which gcc
handles miserably.

Some float/int type punning also avoids gcc brain damage.

Overall speedup 20-35% on Cortex-A8, 20% on Core i7.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21188 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21187 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Patch by Tomas Härdin: $(name) punto hardin chez codemill dot se


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21186 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21185 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

the file for future 16k mode decoder code.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21184 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Two of these are in fact constant size, so use the constant instead of
a variable in the declarations.  The remaining one is small enough
that always using the maximum size is acceptable.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21183 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21182 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

not slower according to benchmarks.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21181 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Yes i did compile&test, no svq3.c was not recompiled.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21180 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21179 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Seems to speed the code up a little...
The placement of many generic functions between h264.c and h264.h is still open
Currently they are a little randomly placed between them.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21178 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

which it inserts the new filter.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21177 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

filter cannot convert between input and output formats.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21176 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Patch by myself and Yuriy Kaminskiy [yumkam mail ru]


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21175 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21174 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

called once per MB in worst case and doesnt seem to benefit from static inline.
Actually the code might be a hair faster now (0.1% according to my benchmark but
this could be random noise)


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21173 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21172 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21171 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

in the context to increase compatibility.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21170 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

no speedloss meassured, also its really not touching anything that is speed relevant.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21169 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21168 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21167 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

<$firstname.$lastname()codemill,se>.



git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21166 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

No speedloss meassured (its slightly faster here but that may be random fluctuations)


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21165 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21164 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Patch by Martin Storsjo (martin AT martin DOT st)


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21163 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

instead.  Avoid a crash on MinGW.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21162 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Patch by Sergiy (gmail(piratfm))



git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21161 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

This fixes issue streaming from Red5 server.
Patch by Art Clarke (aclarke@`echo xyzzy|sed s/y/u/|sed s/y/le/|tr z g`.com)



git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21160 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

No meassureable speed difference on pentium dual & cathedral sample.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21159 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21158 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

functions called more than per mb are moved into the header, scan8 is also
as it must be known at compiletime.
The code after this patch duplicates h264data.h, this has been done to minimize
the changes in this step and allow more fine grained benchmarking.
Speedwise this is 1% faster on my pentium dual core with diegos cursed cathedral
sample.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21157 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b