Showing 1–1 of 1 results for author: Kaikaus, J

Communication-Cost Aware Microphone Selection For Neural Speech Enhancement with Ad-hoc Microphone Arrays

Authors: Jonah Casebeer, Jamshed Kaikaus, Paris Smaragdis

Abstract: In this paper, we present a method for jointly-learning a microphone selection mechanism and a speech enhancement network for multi-channel speech enhancement with an ad-hoc microphone array. The attention-based microphone selection mechanism is trained to reduce communication-costs through a penalty term which represents a task-performance/ communication-cost trade-off. While working within the t… ▽ More In this paper, we present a method for jointly-learning a microphone selection mechanism and a speech enhancement network for multi-channel speech enhancement with an ad-hoc microphone array. The attention-based microphone selection mechanism is trained to reduce communication-costs through a penalty term which represents a task-performance/ communication-cost trade-off. While working within the trade-off, our method can intelligently stream from more microphones in lower SNR scenes and fewer microphones in higher SNR scenes. We evaluate the model in complex echoic acoustic scenes with moving sources and show that it matches the performance of models that stream from a fixed number of microphones while reducing communication costs. △ Less

Submitted 21 April, 2021; v1 submitted 14 November, 2020; originally announced November 2020.

Comments: 5 pages, 4 figures, ICASSP 2021