Generating Music From Emotion (and other experiments)
summary
Claims that novels and videos can be turned into coherent scores by mapping emotion curves to musical parameters using a word emotion association lexicon and parameter‑mapping composition; prototypes include TransProse (melodic movement tied to emotion splines), an automatic video soundtrack using single‑note scoring, and a factory piece played on the building. Tools include sentiment analysis, basic clustering, machine learning and NVIDIA fast style transfer; a CrowdFlower‑tagged Flickr set of a few thousand images probes visual emotions and reveals usable patterns like evening peaks in UFO sightings.
AI-written summary · about
chapters (4)
related talks
34mModelling emotional nuance of musical performance
41mMusic and Machine Learning
08mThe Aesthetics of Disharmony: Harnessing Sounds and Images for Dynamic Soundscapes Generation
43mMake It Sound
05mInteractive Exploration-Exploitation Balancing for Generative Melody Composition
18mPlaying Fast and Loose with Music Recognition
04mData Sonification: Making Music from the yield curve
54mIntegral scherzo: how machines compose and sing