Color-/Lineshifts under Windows?

[Hauke Heibel]

Ok, seriously - sorry for all the noise. Of course it should be

if ((val >> 24) != 0x00)

and all dots are gone. There is just a tiny gray border left around
the non-zero area. Again, I assume this comes from the blending
settings cairo is using.

- Hauke

On Wed, Aug 11, 2010 at 7:20 PM, Hauke Heibel
<hauke.heibel at googlemail.com> wrote:
> I comitted a fix. You can find it here:
>
> http://github.com/hauke76/OpenSlide/commits/
>
> There is an issue left with the alpha channel. I am not sure where it
> is coming from. Using this code in test.c
>
>    if ((val >> 24) == 0xFF)
>    {
>      putc((val >> 16) & 0xFF, f); // R
>      putc((val >> 8) & 0xFF, f);  // G
>      putc((val >> 0) & 0xFF, f);  // B
>    }
>    else
>    {
>      putc(0xFF, f);
>      putc(0xFF, f);
>      putc(0xFF, f);
>    }
>
> I am still seeing a regular pattern of white points. Maybe that is a
> cairo bug?? I am not sure.
>
> After some more testing I also understand now, that cairo seems to
> really blend the image data when sub-pixel transformations are
> applied. The only consequence is that the level 9 image looks a bit
> more blurry than SlideAC's version. On the other hand side,
> downscaling can be seen as blurring and then dropping every second
> pixel, right? So this might be expected and acceptable.
>
> Regards,
> - Hauke
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: lvl9.jpg
Type: image/jpeg
Size: 6978 bytes
Desc: not available
Url : http://lists.andrew.cmu.edu/pipermail/openslide-users/attachments/20100811/c5e01e25/attachment.jpg