@runevision Thanks for the article, helps to understand the difference between the two! Makes me want to see a more accessible introduction to Gabor/Phasor noise, it's a fascinating topic.
@uckersSket@theogramme_dev I agree and I feel like it gave artists more control over the final look. It's harder to art-direct an asset that can change under various conditions (lighting, view angle, etc) vs. one that doesn't. It brings the process closer to that of painting :)
@michal_jares Yup. The same amount of overdraw as a regular particle system, plus every system is effectively rendered twice. So regular lit particles are definitely faster, all things equal :)
@THE2FUN I implemented this as a custom Renderer Feature. Lights render an 'opacity map' similar to traditional shadow mapping. A custom particle shader samples the opacity map to determine shadow attenuation. Technical details can be found in the paper 'Fourier Opacity Mapping' :)
@bitbrain Beautiful work! I've always wondered whether it was worth precomputing the butterfly texture given that FFTs tend to be memory-bound. Did you ever try computing the indices and twiddle factors on the fly?
@SebAaltonen You might be interested in the 'R-Sequence' by Martin Roberts, very similar but requires less operations and possibly more effective :) https://t.co/S5pctVnJ0Z
I wonder if it's worth attempting to port my FFT library to Godot once I'm done with it. I just fear it's too niche and I have doubts that my current Unity implementation will be of much use to anyone anyway. Much to consider ๐ค