Deep Generative Fixed-filter Active Noise Control
This work addresses dynamic noise handling for applications like audio systems or industrial settings, but it appears incremental as it builds on existing fixed-filter methods with a generative twist.
The paper tackles the problem of dynamic noise reduction in active noise control by proposing a generative fixed-filter method that overcomes the limited performance of pre-trained filters, achieving effective noise reduction for various real-recorded noises as demonstrated in simulations.
Due to the slow convergence and poor tracking ability, conventional LMS-based adaptive algorithms are less capable of handling dynamic noises. Selective fixed-filter active noise control (SFANC) can significantly reduce response time by selecting appropriate pre-trained control filters for different noises. Nonetheless, the limited number of pre-trained control filters may affect noise reduction performance, especially when the incoming noise differs much from the initial noises during pre-training. Therefore, a generative fixed-filter active noise control (GFANC) method is proposed in this paper to overcome the limitation. Based on deep learning and a perfect-reconstruction filter bank, the GFANC method only requires a few prior data (one pre-trained broadband control filter) to automatically generate suitable control filters for various noises. The efficacy of the GFANC method is demonstrated by numerical simulations on real-recorded noises.