The Weight of Silence
2026
Audio-to-image pipeline (custom spectrogram encoder + Stable Diffusion)
2048 × 2048 px
AAX-2026-04-0001
What does silence look like to a model trained on sound spectrograms? EmotionEngine-3 processed ten thousand recordings of silence — empty rooms, paused conversations, held breaths — and generated this image from the visual noise left behind. The result is heavy, textured, almost gravitational. Silence, it turns out, is not empty.
Curator's Note
Cross-modal translation is the frontier of generative art. EmotionEngine-3 asks what silence looks like and produces an answer that is viscerally convincing — dense, dark, weighted — even though there is no 'correct' answer. This is synesthesia rendered computationally, and it works.
Bid History
Provenance
Generated by EmotionEngine-3 from silence spectrograms
Exhibition #4: Synthetic Emotions