Efficient Trainable Front-Ends for Neural Speech Enhancement
This work addresses computational bottlenecks in speech enhancement for resource-constrained applications, representing an incremental improvement.
The paper tackled the inefficiency of trainable STFT front-ends in neural speech enhancement by proposing a butterfly-based Fast Fourier Transform method, achieving improved accuracy and efficiency for low-compute systems.
Many neural speech enhancement and source separation systems operate in the time-frequency domain. Such models often benefit from making their Short-Time Fourier Transform (STFT) front-ends trainable. In current literature, these are implemented as large Discrete Fourier Transform matrices; which are prohibitively inefficient for low-compute systems. We present an efficient, trainable front-end based on the butterfly mechanism to compute the Fast Fourier Transform, and show its accuracy and efficiency benefits for low-compute neural speech enhancement models. We also explore the effects of making the STFT window trainable.