On clock synchronization for multi-microphone speech processing in wireless acoustic sensor networks

In wireless acoustic sensor networks (WASNs), clock synchronization is crucial for multi-microphone signal processing, since clock differences between capturing devices will cause signal drift. This in turn severely degrades the performance of multi-microphone signal processing. After a theoretical analysis of the effect of clock synchronization, we evaluate the use of three different clock synchronization algorithms in the context of multi-microphone noise reduction. Our experimental study shows that the achieved precision of clock synchronization enables sufficient accuracy of clock synchronization for the MVDR beamformer in ideal scenarios. However, in practical scenarios with measurement noise on the parameters of interest, time-stamp based clock synchronization algorithms get degraded, while signal based algorithms are still accurate enough for the MVDR beamformer, albeit at a much higher transmission cost.