Thanks, it solved most of the artifacts. But I still have the diagonal
dark line. Anyway, it's much better now for any text surface, and on
TV, these artifacts aren't visible of course.

Best Regards,
Raphael

On 12/9/05, Denis Oliver Kropp <[EMAIL PROTECTED]> wrote:
> Raphael Benedet wrote:
> > Hi,
> >
> > When using the HW accelaration provided by DirectFB Unichrome gfx
> > driver, I get some visual artifacts when blitting 32-bits ARGB
> > surfaces.
> >
> > For instance, if I blit a 800X600 ARGB surface loaded from a bitmap on
> > a black primary surface (800X600 too) with all alphas set to 255, I
> > get a diagonal line a little darker going from bottom-left to
> > top-right of the screen.
> > If I blit a surface with rendered white text on it on top of another
> > surface, I can see some lighter pixels in the no-text area appearing.
> >
> > It seems these are rounding errors. It looks like some pixels are
> > blitted with an alpha value of 1-5 while the real value should be 0 or
> > that pixels with an alpha value of 255 are blitted with an alpha value
> > of 250-254. Not disturbing on the TV output, but on a good LCD, you
> > can clearly see them.
> >
> > If I don't use the HW acceleration, the result is perfect of course.
> >
> > I'm using DirectFB 0.9.24 with the viafb driver from "Patcher2K" on
> > top of a 2.6.10 kernel.
> >
> > Am I the only one to get these artifacts?
>
> No, it seems to be a rounding error in the hardware. Try a 32 bit
> primary surface. It also only applies to blending!
>
> --
> Best regards,
>    Denis Oliver Kropp
>
> .------------------------------------------.
> | DirectFB - Hardware accelerated graphics |
> | http://www.directfb.org/                 |
> "------------------------------------------"
>

_______________________________________________
directfb-users mailing list
[email protected]
http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users

Reply via email to