On Mon, 2 Apr 2018 20:40:42 -0400
Søren Sandmann wrote:
> These changes are not particularly invasive, but I don't think making them
> is possible without some understanding of how pixman's rendering pipeline
> actually works. I could elaborate on that subject and have
Here are some images:
- with no dithering (current pixman):
http://ssp.impulsetrain.com/upload/gradient-none.png
- with the proposed patch applied:
http://ssp.impulsetrain.com/upload/gradient-patch.png
- with proper dithering:
Hi,
If I were still maintainer of pixman, I would definitely stand firm that
the proposed patch is not the right approach. For one, it doesn't actually
have the desired effect on 16 bit destinations. The attached patch (to be
applied on top of the dithering patch) modifies demos/linear-gradient.c
Hi Søren,
I'd really like to see this change available for Cairo. Like you
remarked in your referenced post, dithering has a huge improvement for
gradients on 16 bit destinations. I'm glad to hear you agree the
general idea is in-scope for pixman. The improvements to banding are
almost