[MPlayer-dev-eng] [PATCH] video filter: aspect correction
Christian Ohm
chr.ohm at gmx.net
Wed Dec 18 00:44:38 CET 2002
On Tuesday, 17 December 2002 at 14:42, D Richard Felker III wrote:
> On Tue, Dec 17, 2002 at 10:44:38AM +0100, Diego Biurrun wrote:
> > On Tue, Dec 17, 2002 at 01:30:35AM +0100, Christian Ohm wrote:
> > > the name is a bit long, feel free to change it if you can think of
> > > a better one. (see filter.diff and vf_aspectcorrection.c).
> >
> > What about vf_aspect (and aspect=f)? After all it sets the aspect to a
> > value you specify, doesn't it? The correction is just what you use it
> > for...
>
> I would love to have such a filter. -aspect, -x, -y, ... just don't
> work well when you have a filter chain that's alterring the image
> size. I'd like to be able to set the final desired aspect and not
> worry about having to compute the initial aspect every time I play a
that was the reason i wrote this.
> movie. For example, playing a 400x300 movie to tv out (640x240) with
> g200, I'd like to use:
>
> -vop aspect=4:3,format=yuy2,scale=400:240
i currently use
'vop=expand=672:454:-1:-1:1,scale=672:-2,aspectcorrection=0.92,pp=de/lb' for
my tv out. this isn't fullscreen, because not the whole area is shown on the
tv. the 672 and 0.92 are trial and error with a 4:3 movie until it filled
the screen perfectly.
with my patch you could add a 'aspectcorrection=0.5' to your vop line. i
think this should work in your case.
> I'd be happy to implement vf_aspect.c if it's a desirable feature. I
> just figured before that I was probably the only person who wanted it
> and that someone would complain that it's useless.
hm, look at my original mail and the attached patch.
> BTW, it might be even better to have vf_dsize.c (display size) that
> treats its argument as an aspect if it just gets one number or as
> width and height if it gets two. This would require passing aspect as
> 4/3 rather than 4:3 but that's probably no problem. Or they could be
> two separate filters.
>
> BTW2, I could use some help getting slices passing through this if I
> write it...
>
> Rich
--
Most of your faults are not your fault.
More information about the MPlayer-dev-eng
mailing list