Nyaaori/rockbox
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
23752 commits 17 branches 142 tags 291 MiB
Commit graph

17 commits

Author SHA1 Message Date
Andree Buschmann
28bc321dbb Remove CLIP_TO_15 from codeclib. Remove tabs. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@24834 a1c6a512-1295-4272-9138-f99709370657
 2010-02-21 22:05:48 +00:00
Andree Buschmann
d67e25d3b9 Further work on cook codec. Rounding is not needed when using a large fract part in the internal sample representation. Move quantization array to iram. Beautification of mdct post processing. Speed up of 0.2 MHz on PP5022. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@24821 a1c6a512-1295-4272-9138-f99709370657
 2010-02-21 21:18:05 +00:00
Andree Buschmann
3d6faa08bf Optimization for cook codec. Rework sample output to be able to use highly optimized dsp routines. Moved some functions to iram. Speeds up codec by 1.3 MHz on PP5022. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@24815 a1c6a512-1295-4272-9138-f99709370657
 2010-02-21 19:47:05 +00:00
Dave Hooper
42774d3128 Merge from branches/mdctexp - faster ifft+imdct in codec lib 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@24712 a1c6a512-1295-4272-9138-f99709370657
 2010-02-17 00:49:53 +00:00
Nils Wallménius
b9d7f98f8c Move av_log2 function and asociated table to the codec lib, remove 3 identical implementations, always have LUT in iram (gives a *tiny* speedup on coldfire), make the clz based function return the same value as the non clz based function for 0 input to be safe (slows down flac ~2% on the gigabeat S) 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@23858 a1c6a512-1295-4272-9138-f99709370657
 2009-12-05 16:47:43 +00:00
Mohamed Tarek
c956059ec5 Modify libcook to use rockbox's mdct library in its standalone test program. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@22297 a1c6a512-1295-4272-9138-f99709370657
 2009-08-13 20:26:31 +00:00
Mohamed Tarek
c393b1dc5c Fix compilation of the standalone test program. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@22064 a1c6a512-1295-4272-9138-f99709370657
 2009-07-27 11:28:14 +00:00
Dave Hooper
a8d1cfdec8 Approx 10% speedup in cook on files tested: Remove some inner loops in favour of memcpy/memset/vect_add calls; remove multiplication from index arithmetic in loops in favour of pointer arithmetic; make use of the MULT31, MULT31_SHIFT15 and CLIP_TO_15 implementations from codelib instead of having their own implementations in cook 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@22055 a1c6a512-1295-4272-9138-f99709370657
 2009-07-26 19:06:36 +00:00
Michael Giacomelli
debfd39882 Commited the wrong thing, fix zeroing in cook. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@21989 a1c6a512-1295-4272-9138-f99709370657
 2009-07-21 03:47:41 +00:00
Michael Giacomelli
c062259216 Rearrange loop to avoid one branch per iteration. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@21988 a1c6a512-1295-4272-9138-f99709370657
 2009-07-21 03:40:53 +00:00
Michael Giacomelli
6539b535ad Optimize overlap_math by only doing shifting if theres gain, and moving the check for sign outside of the for loop. 3% speedup on PP5024. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@21940 a1c6a512-1295-4272-9138-f99709370657
 2009-07-18 00:24:54 +00:00
Michael Giacomelli
b957f7214b Rearrange the MDCT library lookup tables so that codecs can access them. Access them in cook. Remove old cook trig tables. Replace old 32x16 bit fixed point multiply code with faster ASM ARM/Coldfire versions using full 32 bit precision. Improves performance signficantly on PP5024 (~200% realtime for a 64k file) while saving a lot of memory and improving accuracy. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@21939 a1c6a512-1295-4272-9138-f99709370657
 2009-07-17 23:53:40 +00:00
Mohamed Tarek
20f8666b89 Remove a call to av_clip() which limits the PCM output of the decoder to 16-bit. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@21911 a1c6a512-1295-4272-9138-f99709370657
 2009-07-16 23:09:28 +00:00
Mohamed Tarek
e184ef1027 Adding support for rm playback. Only cook codec is supported for now and no seeking. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@21695 a1c6a512-1295-4272-9138-f99709370657
 2009-07-06 22:40:45 +00:00
Mohamed Tarek
49ba646d57 -Remove all dynamic allocations, hence remove cook_decode_close() which was basically 
needed for freeing allocated memory.
-Remove any ffmpeg-specific attributes (av_const,av_always_inline .. etc.).
-Move some math functions to cook_fixpoint.h - libavutil/common.h is no longer 
needed.
-Remove libavutil/mem.[c/h], libavutil/common.h and libavutil/internal.h.
-Fix a warning in cookdata_fixpoint.h.


git-svn-id: svn://svn.rockbox.org/rockbox/trunk@20922 a1c6a512-1295-4272-9138-f99709370657
 2009-05-12 20:50:35 +00:00
Mohamed Tarek
0cb3ad58b3 -Took cook_random() from an old cook.c revision [6 mar 2007], which is a 
simple RNG, to eliminate the need for lfg.c/h.
-Remove lfg.c/h and md5.c/h as they are no longer needed and use the 
more common form of my name in docs/COMMITTERS.


git-svn-id: svn://svn.rockbox.org/rockbox/trunk@20915 a1c6a512-1295-4272-9138-f99709370657
 2009-05-11 23:35:50 +00:00
Dave Chapman
fc28cb4ed5 Patch by Mohamed Tarek from FS#10182 - convert codec to fixed-point using patches submitted to the ffmpeg mailing list in 2007 by Ian Braithwaite. 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@20901 a1c6a512-1295-4272-9138-f99709370657
 2009-05-10 22:26:02 +00:00