The GAN Color Organ: Generate visual music using onset detection and HistoGAN

The three-centuries long tradition of unifying color and sound has led countless innovators and artists to create their own version of the color organ. The GAN Color Organ aims to fuse this tradition with AI-generated visuals by integrating audio feature extraction, generative adversarial networks (GAN), and color modification into one complete algorithm. Using this program, artists could create generative GAN-based music visualizations and explore various music and color relationships. The result of this study shed light the audio-visual correspondences between pitch and color as well as note onset and latent space noise vector interpolation.

​

Keywords:

Onset Detection; ReHistoGAN; Music Visualization; Color Organ