[Mplayer-dev-eng] Matrox dualhead and syncfb.

[Felix Buenemann]

On Dienstag,  1. Mai 2001 21:32 you wrote:
> On Tue, 1 May 2001, Felix Buenemann wrote:
> > On Dienstag,  1. Mai 2001 15:54 you wrote:
> > > On Dienstag,  1. Mai 2001 13:45 you wrote:
> > > > On Tue, 1 May 2001, Felix Buenemann wrote:
> > > > > On Dienstag,  1. Mai 2001 12:23 you wrote:
> > > > > > raw support in order to make it seek correctly. I think that some
> > > > > > kind of list with the timetable of each chapter would be enough.
> > > > > > I've also make some changes to vo_mga and vo_xv, so you can
> > > > > > specify anamorphic format with -16:9 that will scale the image
> > > > > > width by 4/3.
> > > > >
> > > > > Think an -aspect x:y option that changes image size before vo
> > > > > plugins would be a much cleaner solution (would only have to modify
> > > > > d_height and d_width).
> > > >
> > > > 	The only problem is that I think that this aspect resizing should
> > > > be vo-pluing dependant, that is, it should be enable only on plugins
> > > > that could do it fast. And you need someway to get the screen
> > > > dimensions in order to do the right thing while resizing. Although I
> > > > suppose we can assume a 4:3 aspect ratio of the screen, and let
> > > > people decide if the plugin is fast enough for them.
> > >
> > > It's left to the plugin to decide wether it uses d_width and d_height
> > > or only uses height and width values! You should take look at the libvo
> > > API. -aspect only needs to set -x and -y values (passed to vo plugin as
> > > d_width and d_height) and the rest is done by the plugin.
> >
> > I've now added needed code to mplayer.c, but I won't commit until Arpi
> > tells me, or I might have to drink 10l Cola =)
> > I've attached modified mplayer.c and cfg-mplayer.h (replacements for
> > current cvs ver), so you can test it, ie. use -aspect 1.85 for Anamorphic
> > dvd in 1.85:1 format, or 4:3 or 16:9 or whatever (note most dvds say 16:9
> > widescreen and at the same time 1:1.85 where in fact it should be named
> > 1.85:1 cause image is wider then high (leaving away 1: and only typing
> > -aspect 1,85 (same as 1.85) will asume you mean 1.85:1).
>
> 	Uhmm, I think I don't understand correctly what aspect does. What
> I was thinking was that when you set aspect to 16:9 you will get wider
> images, that is a 720x576 original will be converted to a 960x576, so you
> don't get this "everybody-is-taller" effect, but your modification seems
> to do the other way, when I set aspect to 16:9 I get even taller people. I
> suppose that it is just a question of what the aspect is, the monitor
> aspect or the original image aspect, anyway I still think that monitor
> aspect should be determined automatically whenever possible (framebuffer,
> mga, xv, etc.).

I work the other way around, I lower picture height instead of stretching 
image width, so on 16:9 you will have 720x405 (dunno how you come onto 
960x576 fo me the other way around it's 1024x576).
But most anamorphic dvds seems to need -aspect 1,85, ie. 6th Sense DVD I own 
is 1,85:1 and fully streched onto height of 576 and The Matrix is 2,35:1 but 
has Black Bars so 1,85:1 is also correct setting.
Of course in windowed mode it may look wrong if your res is not 4:3 (ie 
1280x1024) currently at least SDL driver doesn't care about that in windowed 
mode but does in fullscreen (-vm), anyway fullscreen function of sdl needs 
rewrite to correctly cope with -zoom -vm and -fs mode, maybe I'll change 
aspect awareness for non 4:3 res for windowed, too.
Quick fix for that would be to lower res to 1280x960 which is 4:3 for 
windowed.
Dunno how much other plugins care about correct aspect. 
>
> 	Regards,
>
> 	- german
>

-- 
Best Regards,
   Felix

_______________________________________________
Mplayer-dev-eng mailing list
Mplayer-dev-eng at lists.sourceforge.net
http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng