I was wondering why we didn't just directly modify the opacity (or alpha value) of each pixel instead of averaging the samples by their values on the project. Changing the opacity seems much simpler when it comes to implementing it into code, but I wonder why we didn't use this approach.