The Sound Sketchpad: Expressively Combining Large and Diverse Audio Collections
summary
An algorithm-in-the-loop audio–graphical tool that turns a sung or whistled guide plus drawn paths into collaged music, extracting time-varying features (pitch default) and searching up to 100 candidates from large audio collections. Graphical controls map to energy density weight variance and to audio-processing functions; two reconstruction modes are implemented: sparse (greedy selection minimizing Euclidean distance) and dense (simulated annealing optimizing multi-feature KL divergence, with a density threshold to drop low-amplitude sounds), producing mixable collages that scale beyond DAW workflows.
AI-written summary · about
chapters (6)
related talks
20mFlexible visualization of sound collections
09mComposing And Improvising Using Sound Content-Based Descriptive Filtering
25mCombining Collaborative & Content Filtering in a Recommendation System for a Web-based DAW
16mDemo: Functional Sound Design
08mThe Aesthetics of Disharmony: Harnessing Sounds and Images for Dynamic Soundscapes Generation
43mMake It Sound
15mSeeing Sound
17mParticipatory musical improvisations with Playsound.space