A new synthetic media project combines text-to-image and audio generation called Riffusion. Users can describe music in words and hear..