What Do Neural Networks Listen To? Exploring The Crucial Bands In Speech Enhancement Using Sinc-convolution

Abstract

This study introduces a reformed Sinc-convolution (Sincconv) framework tailored for the encoder component of deep networks for speech enhancement (SE). The reformed Sincconv, based on parametrized sinc functions as band-pass filters, offers notable advantages in terms of training efficiency, filter diversity, and interpretability. The reformed Sinc-conv is evaluated in conjunction with various SE models, showcasing its ability to boost SE performance. Furthermore, the reformed Sincconv provides valuable insights into the specific frequency components that are prioritized in an SE scenario. This opens up a new direction of SE research and improving our knowledge of their operating dynamics.

Stats

• citations1
• S2 citations—
• github stars0
• HF likes0
• heat score2.26
• arxiv keyho2024what

Related papers

• Speech And Speaker Recognition From Raw Waveform With Sincnet (2018)0.00
• SICRN: Advancing Speech Enhancement Through State Space Model And Inplace Convolution Techniques (2024)7.81
• Speaker Recognition From Raw Waveform With Sincnet (2018)20.65
• Complex Spectral Mapping With Attention Based Convolution Recurrent Neural Network For Speech Enhancement (2021)0.00
• Dmf-net: A Decoupling-style Multi-band Fusion Model For Full-band Speech Enhancement (2022)7.16
• Curricular Sincnet: Towards Robust Deep Speaker Recognition By Emphasizing Hard Samples In Latent Space (2021)4.52
• Towards Advanced Speech Signal Processing: A Statistical Perspective On Convolution-based Architectures And Its Applications (2024)2.26
• Audio-visual Speech Enhancement Using Multimodal Deep Convolutional Neural Networks (2017)17.39