Array configuration-agnostic personalized speech enhancement using long-short-term spatial coherence

Personalized speech enhancement (PSE) has been a field of active research for suppression of speech-like interferers, such as competing speakers or television (TV) dialogue. Compared with single-channel approaches, multichannel PSE systems can be more effective in adverse acoustic conditions by leveraging the spatial information in microphone signals. However, the implementation of multichannel PSEs to accommodate a wide range of array topology in household applications can be challenging. To develop an array configuration-agnostic PSE system, we define a spatial feature termed the long-short-term spatial coherence (LSTSC) with a dynamic forgetting factor as the input feature to a convolutional recurrent network to monitor the spatial activity of the target speaker. As another refinement, an equivalent rectangular bandwidth-scaled LSTSC feature can be used to reduce the computational cost. Experiments were conducted to compare the proposed PSE systems, including the complete and the simplified versions with four baselines using unseen room responses and array configurations (geometry and channel count) in the presence of TV noise and competing speakers. The results demonstrated that the proposed multichannel PSE network trained with the LSTSC feature with a dynamic forgetting factor achieves superior enhancement performance without precise knowledge of the array configurations and room responses.