I’ve been playing with shadertoy a bit, and here is a lil demo. It’s probably not the best way to do it, code suggestions welcome.
https://www.shadertoy.com/view/dtlBRM
Here’s most of the code for reference:
void mainImage( out vec4 fragColor, in vec2 fragCoord )
{
// fetch neighbor values from last frame, loop back onto screen if off screen
mat3 n;
n[0][0] = texelFetch(iChannel0, ivec2(mod(vec2(fragCoord) + vec2(-1., -1.), iResolution.xy)), 0).g;
n[1][0] = texelFetch(iChannel0, ivec2(mod(vec2(fragCoord) + vec2(0., -1.), iResolution.xy)), 0).g;
n[2][0] = texelFetch(iChannel0, ivec2(mod(vec2(fragCoord) + vec2(1., -1.), iResolution.xy)), 0).g;
n[0][1] = texelFetch(iChannel0, ivec2(mod(vec2(fragCoord) + vec2(-1., 0.), iResolution.xy)), 0).g;
n[1][1] = texelFetch(iChannel0, ivec2(mod(vec2(fragCoord) + vec2(0., 0.), iResolution.xy)), 0).g;
n[2][1] = texelFetch(iChannel0, ivec2(mod(vec2(fragCoord) + vec2(1., 0.), iResolution.xy)), 0).g;
n[0][2] = texelFetch(iChannel0, ivec2(mod(vec2(fragCoord) + vec2(-1., 1.), iResolution.xy)), 0).g;
n[1][2] = texelFetch(iChannel0, ivec2(mod(vec2(fragCoord) + vec2(0., 1.), iResolution.xy)), 0).g;
n[2][2] = texelFetch(iChannel0, ivec2(mod(vec2(fragCoord) + vec2(1., 1.), iResolution.xy)), 0).g;
// sum of neighbors
float sum = n[0][0] + n[1][0] + n[2][0] +
n[0][1] + n[2][1] +
n[0][2] + n[1][2] + n[2][2];
if(n[1][1] == 0.) {
if(sum == 3.) {
// if dead and has 3 neighbors, come alive
fragColor = vec4(0., 1., 0., 1.);
} else {
// otherwise stay dead
fragColor = vec4(0., 0., 0., 1.);
}
} else {
if(sum == 2. || sum == 3.) {
// if alive and has 2 or 3 neighbors, stay alive
fragColor = vec4(0., 1., 0., 1.);
} else {
// otherwise, die
fragColor = vec4(0., 0., 0., 1.);
}
}
}
Oh that’s pretty rad. I can imagine it being used in a movie as a face for some AI that’s dying or suffering
Interesting. I wonder if there are videos that increase the chance of creating infinite automata?
That’s awesome! What are your thoughts on keeping the colors of your face as they naturally are and as the GoL algorithm acts on the pixels, you blend the colors together?
I was thinking of doing three separate GOL simulations, one on each RGB channel, and letting the colors mix that way into like 6 colors. right now, I clamp the pixel brightness values to 0 or 1, so that’s why it’s black/white, or rather black/green.
deleted by creator
That would be cool, seeing RGB GOL cells moving outward. Maybe instead of 0 or 1 brightness you could just round it for the cell’s purposes, maybe having a const float for flexibility: if it’s greater than
threshold
then it’s “on”.