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

[Marc Hoffman]

On 7/28/07, Michael Niedermayer <michaelni at gmx.at> wrote:
> On Fri, Jul 27, 2007 at 09:38:53PM -0400, Marc Hoffman wrote:
> > On 7/27/07, Michael Niedermayer <michaelni at gmx.at> wrote:
> > > Hi
> > >
> > > On Fri, Jul 27, 2007 at 05:40:14PM -0400, mmh wrote:
> > > [...]
> >
> > snip --- agreed and will make those changes.
> >
> > > > --- libavcodec/fft.c  (revision 9807)
> > > > +++ libavcodec/fft.c  (working copy)
> > > > @@ -109,7 +109,7 @@
> > > >                  }
> > > >                  nblocks = nblocks >> 1;
> > > >              } while (nblocks != 0);
> > > > -            av_freep(&s->exptab);
> > > > +//            av_freep(&s->exptab);
> > > >          }
> > >
> > > hmm
> >
> >
> > This because I'm testing on a machine that deallocates that array.
> > i.e. X86 and guess what we make that array vanish.  Give me a
> > direction, and I will cast down the code as you command.
>
> ff_fft_init() should call the fixed point init
> also
> the mmx code in it should be split out into a i386/whatever.c
> and allso called from ff_fft_init()
>
> i think that would solve it?

ok a bit of a challenge lots of reworks needed.  But its a good
solution. I will look at it.

Marc