Foster Strengths and Circumvent Weaknesses: a Speech Enhancement Framework with Two-branch Collaborative Learning
This work addresses speech enhancement for audio processing applications, presenting an incremental improvement by integrating existing paradigms.
The paper tackles the problem of single-channel speech enhancement by proposing a two-branch framework that combines magnitude-spectrum and complex-spectrum methods to leverage their strengths and overcome weaknesses, achieving superior performance on the TIMIT benchmark.
Recent single-channel speech enhancement methods usually convert waveform to the time-frequency domain and use magnitude/complex spectrum as the optimizing target. However, both magnitude-spectrum-based methods and complex-spectrum-based methods have their respective pros and cons. In this paper, we propose a unified two-branch framework to foster strengths and circumvent weaknesses of different paradigms. The proposed framework could take full advantage of the apparent spectral regularity in magnitude spectrogram and break the bottleneck that magnitude-based methods have suffered. Within each branch, we use collaborative expert block and its variants as substitutes for regular convolution layers. Experiments on TIMIT benchmark demonstrate that our method is superior to existing state-of-the-art ones.