Generative models enable an entirely new type of content creation where the creative process is a shared interaction between man and machine. Ever since the advent of high-resolution GANs I’ve been exploring their potential to create rich, immersive soundscapes that function as a kind of visual meditation exercise. This piece uses a StyleGAN2 model trained on thousands of existing paintings and uses audio feature extraction to manipulate the latent space of the generative model to create an interpolation that is in perfect sync with the music. Model trained by @pbaylies, music by @kuplasound. More information at https://lionbridge.ai/articles/neural-synesthesia-when-art-meets-gans/