Multichannel Wiener Filtering for Dual-Microphone Speech Processing in a Car Environment

Abstract

With advancements in automotive electronics and sensors, the sound pick-up using multiple microphones has become feasible for hands-free telephony and voice command in-car applications. Adding a second microphone so that both driver and passenger can have dedicated microphones can help in better sound pickup. However, challenges remain in effectively processing multiple microphone signals due to bandwidth or processing limitations. If a single resulting microphone signal is to be transmitted or processed, adding (mixing) the microphone signals or switching between them can be efficient and cost effective solutions. However, switching between microphone signals can cause sudden changes in the acoustic path between loudspeakers and microphones, leading to performance decay in the acoustic echo canceling module. And mixing the two microphone signals can lead to echoes and notch filtering effects (spectral holes) in the resulting mixed signal. This work explores the use of the Multichannel Wiener Filter algorithm with a two-microphone in-car system to enhance speech quality for driver and passenger voice, i.e., more specifically to mitigate notch-filtering effects caused by echoes and improve background noise reduction. We evaluate its performance under various noise conditions using modern objective metrics such as the Deep Noise Suppression Mean Opinion Score. The effect of head movements of driver/passenger is also investigated. The proposed method is shown to provide significant improvements over a simple mixing of microphone signals.