Huhtala V, Juvela L, Schlecht SJ (2024)
Publication Type: Journal article
Publication year: 2024
Book Volume: 31
Pages Range: 1169-1173
In recent years, neural network-based black-box modeling of nonlinear audio effects has improved considerably. Present convolutional and recurrent models can model audio effects with long-Term dynamics, but the models require many parameters, thus increasing the processing time. In this letter, we propose KLANN, a Koopman-Linearised Audio Neural Network structure that lifts a one-dimensional signal (mono audio) into a high-dimensional approximately linear state-space representation with nonlinear mapping, and then uses differentiable biquad filters to predict linearly within the lifted state-space. Results show that the proposed models match the high performance of the state-of-The-Art neural models while having a more compact architecture, reducing the number of parameters by tenfold, and having interpretable components.
APA:
Huhtala, V., Juvela, L., & Schlecht, S.J. (2024). KLANN: Linearising Long-Term Dynamics in Nonlinear Audio Effects Using Koopman Networks. IEEE Signal Processing Letters, 31, 1169-1173. https://doi.org/10.1109/LSP.2024.3389465
MLA:
Huhtala, Ville, Lauri Juvela, and Sebastian J. Schlecht. "KLANN: Linearising Long-Term Dynamics in Nonlinear Audio Effects Using Koopman Networks." IEEE Signal Processing Letters 31 (2024): 1169-1173.
BibTeX: Download