Parasitic light simulator

Here are some examples of how auto parasitic light can affect your images.
 
In the first row, the parasitic light is mostly on the opposite side leaving the projected image with a good contrast.
 
In the center row, the black background and small high contrasted objects produce very low pollution and we see almost the same results in the dome.
 
In the last row, our footage is very bright throughout the image, producing a lot of pollution which heavily affects the black level and contrast.

The mathematical concept can be pretty simple to understand with these two general steps.
 
First (1), not all pixels of a dome master are the same size on the screen. Those in the center are a bit stretched and take more surface than those on the outside. So there is compensation of the impact of every pixel by adjusting the relative luminosity.
 
The second step (2) is to sum, for every pixel, the luminosity emitted by every other pixel in the dome. The sum takes into account the relative orientation between each pixel.
 
This calculation assumes the following :
⇒ There is no reflection from the floor or any other surfaces in the dome - only the dome screen.
⇒ The screen is a perfect Lambertian surface.
⇒ It only computes a single pass, which means that it doesn't take into account the pollution caused by the parasitic light itself.
 
But despite this simple explanation, the complexity resides in the optimization of the algorithm which is approximately O(n⁴) growth relative to the resolution. This means that if it take 2 seconds to resolve a 128x128 pixels frame, it will take 32 seconds for 256 pixels, 2h20 for 1k, 36h for a 2k or 24 days for a 4k frame.... I don't think you wish to wait so long. With our optimization, every picture size take approximately the same amount of time to compute.