[vlc-devel] 2.0.0 and resampling

[Laurent Aimar]

On Wed, Feb 29, 2012 at 03:06:02PM +0100, Rémi Denis-Courmont wrote:
> On Tue, 28 Feb 2012 22:13:55 +0100, Jean-Baptiste Kempf <jb at videolan.org>
> wrote:
> > It seems to depend a lot on the platform and the CPU,
> 
> Any floating point-intensive task depends a lot on the compilation flags
> and the architecture. It is certainly much faster with SSE enabled (that is
> to say if the library is compiled with SSE enabled).
> 
> It's definitely going to be slow on ARM, unless NEON is used (not just
> VFP).
> 
> > but in many
> > situation, the audio decoding took more that the video one.
> 
> You don't typically postprocess the video. Scaling is done in hardware in
> that case.

 For CPU with an FPU, I don't really see why that would be so slow. Resampling
is a FIR and you don't need that much taps to have a good enough quality. For
a 48kHz resamling, around 40 mul + 40 add (+ a one integer div + mod) per
output sample should be quite good, and it doesn't seem that much computation
(especially compared to a video decoder).

> It is a design bug that VLC needs to resample all the time.
 Yes, for local playback that could be avoided but I haven't tried to change
that.

-- 
fenrir