Reproduce “Expressive Guitar Synthesis via Controllable Latent Flow Matching” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hsueh-Wei Fu, Yen-Tung Yeh, Yi-Hsuan Yang
The authors present a guitar synthesis system that predicts editable per-string pitch and amplitude curves, then renders them as audio. A latent flow-matching synthesizer generates sound from these controls. On GuitarSet, reference-velocity results show 26.52 cents voiced-frame pitch error versus 33.53 for a matched baseline, and envelope correlation of 0.88 versus 0.63. Listening scores lead synthesized systems on all four criteria, but only with reference velocity.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.