[Ffmpeg-devel] Fixed vs. Floating Point AAC

Måns Rullgård mru
Wed Mar 8 22:10:35 CET 2006


Rich Felker <dalias at aerifal.cx> writes:

> On Wed, Mar 08, 2006 at 04:54:39PM -0000, M?ns Rullg?rd wrote:
>> 
>> Rich Felker said:
>> > On Wed, Mar 08, 2006 at 01:01:47PM +0100, Michael Niedermayer wrote:
>> >> it should support both, fixed point is absolutely needed for embeded systems
>> >> (and richs K6), and floating point is more accurate and faster on modern
>> >> cpus, and floating point is much easier to implement
>> >
>> > Would you like to provide benches showing it's faster? Testing of
>> > libvorbis versus tremor has shown that the integer implementation
>> > (tremor) is always faster, even on the latest cpus. This is without
>> > any vectorization (except possible float vectorization by gcc?); with
>> > mmx/3dnow/etc the difference should be even more drastic.
>> 
>> libvorbis faster than tremor on Alpha.
>
> Including the 'low precision' mode in tremor (which makes no audible
> difference, i.e. you won't be able to distinguish them with
> doubleblind testing)?? IIRC it's between 1.5 and 2x as fast as the
> normal mode in tremor.

It's a long time since I compared them, so I don't remember exactly
what I tried.

-- 
M?ns Rullg?rd
mru at inprovide.com





More information about the ffmpeg-devel mailing list