Structured Audio Embeddings for Musical Timbre and Effects

Geodesics and decision boundaries in the Poincare ball model of hyperbolic space — Geodesics and classification regions in the Poincare ball (Fig. 1). Wada et al., DAFx 2025 (CC BY 4.0)

We investigate structure-aware representations for musical timbre and audio effect processing. Hyperbolic embeddings support timbre representations for musical-instrument sound synthesis and order-aware classification of audio effect chains. A separate approach investigates Matryoshka embeddings for estimating audio effect chains; these are distinct representation methods.

Related Publications

Aogu Wada, Tomohiko Nakamura, and Hiroshi Saruwatari, “Hyperbolic embeddings for order-aware classification of audio effect chains,” in Proceedings of International Conference on Digital Audio Effects, Sept. 2025, pp. 396–402.
Futa Nakashima, Tomohiko Nakamura, Norihiro Takamune, Satoru Fukayama, and Hiroshi Saruwatari, “Hyperbolic timbre embedding for musical instrument sound synthesis based on variational autoencoders,” in Proceedings of Asia Pacific Signal and Information Processing Association Annual Summit and Conference, Nov. 2022, pp. 736–743.