doronrk / morphopencl Goto Github PK

Usage: Mode 3 is the 'final' version of the project

Change the mode with the following keys
    1   time domain. Tip: try producing vocal white noise, then singing a vowel
    2   frequency domain
    3*  face FFT visualizer. Tip: try looking at the face from the side. 
    4   face melt (bug that looked cool)
    5   split (bug that looked cool)

Notes on Mode 3
    Responds best to white noise and music.
    Use up/down arrow keys to control face ratchetness. Face ratchnetness looks coolest in the range 0.9 to 1.0

Press 'i' to show/hide the instructions when running the visualizer.

Idea behind Mode 3 - Construct a 'brightness' matrix with values that reflect the brightness at a given pixel relative to the entire image - Send a proportional number of 1,000,000 particles to that point in x,y space with additive brightness to reconstruct the image - Project the x,y points onto a cyclinder to give the illusion of a 3-d face - Perturb the z position of the particles at given rows along the face according to energy at corresponding frequency bins.

ideas for improvements: - normalize height of the spectrum according to most recent volume. - add point lighting of the face - better clarity of mapping between precise frequency y-location of face particle perturbance.