Cells effect#1589
Conversation
|
This is very fun to use :) A couple small thoughts on parameters:
|
| } | ||
| } | ||
|
|
||
| public static ImmutableHashSet<PointI> CreateControlPoints (RectangleI roi, int pointCount, RandomSeed pointLocationsSeed) |
There was a problem hiding this comment.
I think there could be an argument for putting this in the Sampling class, with a different name. This function is producing randomly sampled points within a region, so it's related to the other function which producing uniformly distributed samples within a pixel
There was a problem hiding this comment.
There was a problem hiding this comment.
Thanks, I think I'm okay with it staying in SpatialPartition for now until there's another use for it in the future
|
@cameronwhite regarding the Regardless, I changed the default to |
|
Thanks, I think |
|
That's great @cameronwhite, but setting it to |
|
In my testing this was working as expected when using Pinta's palette: 
But I think that unit test would be defaulting to the preset "Black and White" gradient, which seems to actually go from white to black? |
|
Great point @cameronwhite it somehow escaped me that the default |
I stumbled upon the term "Worley noise", and found that it's the same basic idea as the Voronoi diagram, except that the color of the location varies depending on the distance to the closest control point.
When I googled it, I immediately mapped its appearance to the "cellular" filter from Jerry Huxtable (http://www.jhlabs.com/ip/filters/CellularFilter.html). I used a similar name, but my implementation is actually completely independent, and based on the Voronoi diagram code we already had.
The various customization options (and very notably the "Color Scheme Edge Behavior" option and the customizable gradients option) allow us to obtain other results that look very different from the default one and from each other too