Last video for today, I promise:

Last video for today, I promise: Parallax scrolling with 3 layer masks
(this time on a Teensy 3.1 because on an AVR the framerate broke down to 20 fps)
The video and the live effect as well need a bit distance for the best illusion while viewing.
It must be 20 years ago now that I tryed to program a parallax scrolling the last time. This time it looks better.
Except the flickering edges I am happy with it. But hey, it is still just 8 bit math I am using.
http://www.youtube.com/watch?v=lWT5u1XoPF4

oh man, that’s very very nice. Since the noise stuff is just X/Y you could totally play with zooming in and out.

Also, this makes me want to get the 4d versions of the noise functions working - because then you could have X/Y and Z planes for multi-layers like this, and move all three of those forward in “Time” :slight_smile:

(I have them sketched out, just haven’t gotten to them yet).

Also, I should put some work into performance optimizing the arm versions of the noise code. Maybe consider making 32-bit versions available as well :slight_smile:

Hell yeah! The noise has so much potential! Basically you insired me to do all that with a short video you posted, I guess a noise moving along z with a black/red blend palette. I could never forget it. THANK YOU for the inspiration and the amazing tools you provide again and again!

What is going on here? I’m baffled about how to make this. Can you post code?

It is basically very simple: there are 3 noise layers (blue, pink, orange) with a different scale moving linear along y and wobbling along x. All 3 noises are not mapped completely - just a constrained range (between 130 and 160 I think) results in a mapped pixel. It draws a kind of a noise mesh - noise with holes in it. That way I draw the blue layer first, then the middle pink one and last the front orange one. So you see through the holes the layers behind. Makes my explanation sense? If you still want to see the code I have to (and would) clean it up…

So are the noise “images” static matrices and you’re just translating them around (in And out of the frame) or are you always generating new noise? I haven’t used noise much, but it guess it’s surprising to me that it makes these “webs;” I would think it would be much more stochastic.

I recommend to play with the noise example first. The 3 underlaying noises are not static, they are moving along x and y with static z and scale. All 3 of them are calculated new for every frame. The web results only from the colormapping. But instead of mapping “black” I do nothing which means, what was there before is not touched and stays visible. If you constrain the used noise values just to one side you get noise spots - you draw only the holes while the web is not showed.