Thinking about scaling code, and how you would render into a higher resolution buffer, but then map that back to a lower resolution pixel grid.
A while ago @Mark_Kriegsman posted an example of how this was done in 1-D, and I understood it and could adapt the code pretty easily. Basically you use n of the most significant bits of your x value as the pixel address, and the remaining m as the brightness of the adjacent pixel.
But my mental model broke down when I tried to move into 2-D. What do the pros call this, and has anyone got some simple example code?
(I think I’ve got the idea in theory, as illustrated—the real pixels with 1 “sub pixel” lit get 1/4 brightness in my example, those with 2, 1/2) but I’m not sure how that translates in practice. is it simply enough to logical AND the “fractional” parts of each x and y pixel address to get the brightness for that pixel?)
What is wrong with the scaling function we discussed here some days ago? It does exactely the job you describe. Give it a try. https://plus.google.com/115124694226931502095/posts/iMKuvhM14Vk
Interpolation for mapping a high-res grid into a low-res one? For what?
Btw. while thinking about an even faster scaling routine I wondered, if the Bresenham line algorythm could be abused for that, as well.
Because I’m not sure if the scaling function you’re pointing to there quite does what I want?
It’s not just lines I want to scale/antialias—it’s points and dots. Actually, I think sprites really. Explain to me what the x1, y1, x2, y2 inputs to your MapNoiseToRectangle() function actually represent?
So I probably understood your post wrong. I still wonder which sense an interpolation has for scaling something DOWN. I would expect it to blur the result. For scaling something UP in my opinion it makes totally sense - to fill the missing points with smooth gradients. Maybe I do not see a relevant detail or I still do not understand the main problem you are going to solve. If so, sorry.
Robert , did you happen to find a way to reproduce a higher resolution matrix into a lower one ?
i fully understand what the intention is and totally agree, a way to “average” the pixel values inside the high res matrix and use the result as low res matrix brightness seams good, is there a better one ? thanks
