[MPlayer-dev-eng] [BUG] surround audio plugin broken

[D Richard Felker III]

On Mon, Jan 06, 2003 at 11:09:27AM +0800, Anders Johansson wrote:
> Hi,
> 
> > On Mon, Jan 06, 2003 at 02:48:06AM +0100, Michael Niedermayer wrote:
> > > Hi
> > > 
> > > On Monday 06 January 2003 02:24, D Richard Felker III wrote:
> > > [...]
> > > > > btw IMHO libaf should NOT force float (i mean
> > > > > int->float->filter->float->int) by default.
> > > > > (only if it's requires (the requiested filter cannot operates on
> > > > > integer))
> > > >
> > > > Agree, it's lame. It can be fixed by -af-adv force=1. But the force
> > > > param is only supposed to affect auto-inserted stuff, and right now
> > > > the resample plugin forces things to float even if you load it
> > > > explicitly with -af resample=44100. I'm guessing that in the future
> > > > the force=1 option might do other things that lower quality too, so
> > > > it's very bad to require it to get rid of the float nonsense.
> > > >
> > > > > ppl are already reported slowdowns of 0.90-rc series compared to
> > > maybe we should optimize libaf a bit too ...
> > > it seems there is a modulo in the innermost loop in the resample filter for 
> > > example ...
> > > 
> > > ill try to optimize the libaf stuff tomorrow perhaps ...
> > 
> > Cool. AC3 decoding and resampling to 44100 are the two biggest
> > bottlenecks preventing realtime dvd playback on my box right now! With
> > Fabrice working on the ffmpeg ac3 decoder and libaf optimized, I
> > imagine I'll finally be able to play dvds (without -framedrop).
> 
> Try the linear interpolation it is faster than the filter bank
> resampling.

It does indeed help, reducing sound cpu usage by a couple percent.
However, I expect that an optimized implementation of the filter bank
code could be even faster.

Rich