[FFmpeg-devel] [PATCH] Common fixed-point ACELP routines (1/3) - math

[Vladimir Voroshilov]

Hi, All

Michael asked me to split my large G.729 patch into smaller part
and care about future sharing code with another ACELP-based codecs.

He pointed me to AMR SOC code and i found several routines very similar to mine.

Unfortunately those code is incomplete and Robert Swain said me that
there was no work on AMR fixed point decoder for a long time.
All other ACELP-based codecs are floating point.

According to above, looks like there is no yet ACELP-based fixed-point decoder
to share code with neither in FFmpeg nor in FFmpeg's SOC repository.

Anyway i looked into floating point code too and tried to extract
routines that can be useful while implementing such codecs sometime in
future.

This patch set contains:
1/3. acelp_math - self-contained
2/3. acelp_filt - selft contained
3/3. acelp_lpc - depends on 1 and 2

This is first patch, containing various math routines like cos(x),
log2(x), pow(14+x)

-- 
Regards,
Vladimir Voroshilov mailto:voroshil at gmail.com
JID: voroshil at gmail.com, voroshil at jabber.ru
ICQ: 95587719
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: acelp_math_18.diff
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20080422/2e958ccb/attachment.asc>