[vlc-devel] [RFC PATCH] Direct rendering spu placement
thomas at gllm.fr
Mon Jan 5 18:38:17 CET 2015
On Wed, Dec 31, 2014, at 19:04, Rémi Denis-Courmont wrote:
> Le 2014-12-29 20:46, Thomas Guillem a écrit :
> > Hello, I need your opinions about this patch/question.
> > I have an issue when displaying an HD video on a display of
> > 1200x1920.
> > Here is the example: with a 1916x1076 video on a 1200x1920 display.
> > The place
> > returned by vout_display_PlacePicture is 1200x673, which is good, but
> > it's
> > smaller than the video size, so it's not used.
> > So, do you think it's safe to remove the following check ?
> > (fmt_spu.i_width * fmt_spu.i_height < place.width * place.height)"
> git-blame points at:
> commit 4a49754d943560fe79bc42f107d8ce566ea24898
> Author: Laurent Aimar <fenrir at videolan.org>
> Date: Thu Apr 14 22:45:34 2011 +0200
> Improved subtitiles/OSD quality when the rendering is done by the
> vout display module.
> In this case, the subpictures are rendered either at the display
> size or
> the video size (the biggest one).
> > I don't see any problems with having a spu size smaller than the
> > video size.
> The code thrives for the best quality option. In my opinion, the video
> size should always be used because that's the only clean and race-free
> option - even though that will reduce quality on low resolution videos.
> I fail to see any advantage to your patch.
The subpicture will be downscaled to the display size. So, why not using
the display size directly ? It'll be the same quality, no ?
My patch was made for direct spu rendering in android.
I use the display size and not the video size to setup an android
If the display is portrait, video size will be used for spu placement,
and the spu region will be outside of the video.
I can also fix my issue by patching the android window vout.
> Rémi Denis-Courmont
> vlc-devel mailing list
> To unsubscribe or modify your subscription options:
More information about the vlc-devel