Real-time Streaming Wave-U-Net with Temporal Convolutions for Multichannel Speech Enhancement
This addresses the problem of low-latency speech enhancement for conferencing applications, but it is incremental as it builds on existing Wave-U-Net architecture.
The paper tackled real-time multichannel speech enhancement by proposing a streaming Wave-U-Net system with temporal convolutions and self-attention, achieving a Real-Time Factor of 0.4 with 40ms chunks.
In this paper, we describe the work that we have done to participate in Task1 of the ConferencingSpeech2021 challenge. This task set a goal to develop the solution for multi-channel speech enhancement in a real-time manner. We propose a novel system for streaming speech enhancement. We employ Wave-U-Net architecture with temporal convolutions in encoder and decoder. We incorporate self-attention in the decoder to apply attention mask retrieved from skip-connection on features from down-blocks. We explore history cache mechanisms that work like hidden states in recurrent networks and implemented them in proposal solution. It helps us to run an inference with chunks length 40ms and Real-Time Factor 0.4 with the same precision.