Speech signals recorded in an enclosed space by microphones at a distance from the speaker are often corrupted by reverberation, which arises from the superposition of many delayed and attenuated copies of the source signal. Because reverberation degrades the signal, removing reverberation would enhance quality. Dereverberation techniques based on acoustic multichannel equalization are known to be sensitive to room impulse response perturbations. In order to increase robustness, several methods have been proposed, as for example, using a shorter reshaping filter length, incorporating regularization, or applying a sparsity-promoting penalty function. This paper focuses on evaluating the performance of these methods for single-source multi-microphone scenarios, using instrumental performance measures as well as using subjective listening tests. By analyzing the correlation between the instrumental and the perceptual results, it is shown that signal-based performance measures are more advantageous than channel-based performance measures to evaluate the perceptual speech quality of signals that were dereverberated by equalization techniques. Furthermore, this analysis also demonstrates the need to develop more reliable instrumental performance measures.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.