Ferruccio Busoni : While others have worked with AI image synthesis to surpass photorealism, I use to make images from music. Large Language Models behave unpredictably when prompted with text that it cannot associate easily with its training. Words like ‘counterpoint’, ‘melody’, ‘rhythmic vitality’ make models such as DALL-E and Stable Diffusion reach for anything that might by some chance generate images. Add ‘in the style of Stravinsky’ or ‘inspired by music of Bach’ and the models react in riotously unexpected, chaotic ways. But their training always ensures coherent visual syntheses. In this series, I combine select synthograms with parts of different images, applying layer modes, transparency, masking, lighting, cloning and erasure. The results simulate a synaesthesia, but are not literal (if that were even possible) representations of a composer’s music. In the end, these are visual celebrations of my experience and love of the rich, layered and challenging world of formal or classical music. https://platform.wise.art/author/TomAng/ https://www.instagram.com/tomangphoto/