[FFmpeg-devel] Fixpoint FFT optimization, with MDCT and IMDCT wrappers for audio optimization

Marc Hoffman mmhoffm
Sun Jul 29 20:17:38 CEST 2007


On 7/29/07, mmh <mmh at pleasantst.com> wrote:
> mmh writes:
>  >
>  > Hi,
>  >
>  > I could really use some help on the integration aspects of this patch.
>  > The intent of this is to be able to implement optimized audio codecs
>  > for architectures with out floating point operations.  This is part of
>  > the cook fixpoint patch but its a little more general.
>  >
>  > I don't know what to do with s->exptab in fft.c, I need its contents I
>  > guess this table wouldn't be deleted on a machine that will be using
>  > fixpoint arithmetic however for testing I'm not sure what to do other
>  > than code replication.
>  >
>  > Thanks
>  > Marc
>
> I think I got the integration issues out now.  Thanks for all the great input.

cosmetic removed.

+++ libavcodec/fft-test.c	(working copy)
@@ -128,7 +128,6 @@
     }
 }

-
 float frandom(void)
 {
     return (float)((random() & 0xffff) - 32768) / 32768.0;
@@ -153,13 +152,47 @@
     }
 }




More information about the ffmpeg-devel mailing list