I’m wondering if there’s a way to do something like this. The basic idea is if the first b&w mask overlaps any piece of the second b&w mask, we render that piece of the original image brighter than the rest. The idea is to make any brick that is partially overlapped appear to be lit up.
This is like those LED walls that convert a video feed into low-res LED output. Except in this case the array of “LEDs” is not laid out so neatly.
(Ideally the brightness of each brick will be proportional to the amount of overlap, but that’s not demonstrated in the above image)
The Trautner help patch says: “The Mask Image is supposed to be a 8 bit grayscale bitmap. In the bitmap you define areas of interest via shades of gray. For now you can therefore distinguish between movement in 256 areas.”
My mask has 1000-2000 areas (the image above is just a small piece).
you can also do that with pipet node, grid 100x100 (1000 values) is not realy hard to get. i’ve would convert your luma mask to a mesh with subsets, then get a center of each subset, then put thouse cords in to pipet, then from the pipet values you can get an color of the point under subset, so you can use color from that point to color of subset.
you need to check it first with grid like the one you show on top, so you have a clue how subsets work so you don’t mess anything.
Thanks guys I ended up sort-of combining your two ideas. First I created a mask where each region was a different color (the colors are automatically generated by a photoshop script after i create the one-color mask). Each color corresponds to an integer index (for example RGB 0,0,5 is index 5). Then I created a Pipet-esque Dynamic plugin that generates a color spread, where for each item a color represents a coordinate (for example when R=0.5 and G=0.5 the coordinate refers to the center of the image). The result of the plugin is cached so it doesn’t actually add any overhead. The color spread essentially maps each region of the led mask to the pixel location smack dab in the middle of the region. The color spread feeds a dynamic texture which goes into the Pixel Shader along with the other textures. The pixel shader does all the work and only requires 3 tex lookups.