Rick Brewster

Alex pointed me to the paper that gpufilter is based on: https://hhoppe.com/recursive.pdf (referring back to the post above ...) Is it possible to split these calculations into two passes, i.e....

Once we have an integral table, the remainder of what you describe is very doable. From a data transfer standpoint, it's pretty similar to a convolution matrix shader. For generating...

For anti-aliasing, I agree that we would treat them separately. The pixel shader would take 2 input arrays (via resource texture) - a 1D array of `float4` for each of...

So from what I understand, we don't actually need a full integral table. The full integral table is problematic to calculate in Direct2D because it means that calculating its bottom-right...

For a rect that spans all 4 tiles (expanded to 5x5 tiles for legibility), we'd have: I + H - G + F - E + D - C -...

Also I think the shader can take 4 sets of RLE data: - rectangles - vertical strips - horizontal strips - pixels Each of these has a slightly different implementation...

@JimBobSquarePants I recently spent a lot of time implementing XMP support for Paint.NET earlier this year. Happy to consult and provide sample code. This stuff is a beast, it took...

In Paint.NET, I only bubble up XMP as the raw XML. If a plugin wants to handle XMP they have to parse the XML themselves. The main -- frankly only...

> Also would be nice to support DDS files Paint.NET's DDS support is now sourced from @0xC0000054 's DdsFileTypePlus plugin, https://github.com/0xC0000054/pdn-ddsfiletype-plus . This could probably be adapted into something for...

Oh -- I don't need help, I was just replying to the earlier comment. Thanks though!