[MPlayer-cvslog] r30455 - trunk/libaf/af.h

Reimar Döffinger Reimar.Doeffinger at gmx.de
Sun Jan 31 02:43:20 CET 2010


On Sun, Jan 31, 2010 at 02:26:28AM +0200, Uoti Urpala wrote:
> On Sat, 2010-01-30 at 11:09 +0100, reimar wrote:
> > Author: reimar
> > Date: Sat Jan 30 11:09:14 2010
> > New Revision: 30455
> > 
> > Log:
> > Make the code match the documentation by making higher-quality but slower
> > audio filtering the default.
> > This mostly means lavcresample being the default instead of plain "resample".
> 
> Didn't your own earlier commit (imported to svn as r15951) already do
> this? As far as I can tell this new commit only breaks selecting the
> "automatic" mode with "-af-adv 0", making it behave the same as "-af-adv
> 1". I doubt anyone really used that feature anyway, but now it's further
> from documentation.

Hm, probably. I did some testing but only looked at CPU usage and that obviously
was a bad idea.
However I'm not sure about "further from the documentation", the previous code
used the "fast" path when SSE or 3DNOW are available and the "slow" otherwise,
which in addition to being more a x86 vs. not test nowadays just seems like the wrong
way around...
I guess all this needs a bit more work, but unless I find someone with some real
issues with it to help fix it I just updated the documentation.


More information about the MPlayer-cvslog mailing list