Speech dereverberation is one of the most difficult tasks in acoustic signal processing. Of the various problems involved in this task, this paper highlights "over-whitening," which flattens the characteristics of recovered speech. This distortion sometimes happens when inverse filters are directly calculated from microphone signals. This paper reviews two studies related to this problem. The first study shows the possibility of compensating for such over-whitening to achieve precise speech-dereverberation. The second study presents a new approach for approximating the original speech by removing the effect of late reflections from observed reverberant speech.

It is well known that speech captured in a room by distant microphones suffers from distortions caused by reverberation. These distortions may seriously damage both speech characteristics and intelligibility, and consequently be harmful to many speech applications. To solve this problem, we proposed a dereverberation algorithm based on multi-channel linear prediction. The method is as follows. First we calculate prediction filters that cancel out the room reverberation but also degrade speech characteristics by causing excessive whitening of the speech. Then, we evaluate the prediction-filter degradation to compensate for the excessive whitening. As the reverberation lengthens, the compensation performance becomes worse due to computational accuracy problems. In this paper, we propose a new computation that may improve compensation accuracy when dealing with long reverberation.

The Multiple input-output INverse/filtering Theorem (MINT) proves that N + 1 inverse filters are necessary to precisely control sound at N points in a space, and gives the minimum orders of such filters. In this paper, we propose the Indefinite MINT Filters (IMFs) for adding one or more control points to the above framework without increasing the number of inverse filters. Although the controllability of the new point is not sufficient, that of the other points is still maintained high enough by the principle of the MINT. In a two point sound control (using two inverse filters), the IMFs could reduce the squared error to the desired sound up to - 10 dB at the second point which is not controlled by the MINT.

This paper describes a method for estimating the amplitude characteristics of poles common to multiple room transfer functions from musical audio signals received by multiple microphones. Knowledge of these pole characteristics would make it easier to manipulate audio equalizers, since they correspond to the room resonance. It has been proven that an estimate of the poles can be calculated precisely when a source signal is white. However, if a source signal is colored as in the case of a musical audio signal, the estimate is degraded by the frequency characteristics originally contained in the source signal. In this paper, we consider that an amplitude spectrum of a musical audio signal consists of its envelope and fine structure. We assume that musical pieces can be classified into several categories according to their average amplitude spectral envelopes. Based on this assumption, the amplitude spectral envelope of the musical audio signal can be obtained from prior knowledge of the average amplitude spectral envelope of a musical piece category into which the target piece is classified. On the other hand, the fine structure is identified based on its time variance. By removing both the spectral envelope and the fine structure from the amplitude spectrum estimated with the conventional method, the amplitude characteristics of the acoustical poles can be extracted. Simulation results for 20 popular songs revealed that our method was capable of estimating the amplitude characteristics of the acoustical poles with a spectral distortion of 3.11 dB. In particular, most of the spectral peaks, corresponding to the room resonance modes, were successfully detected.

A speech signal captured by a distant microphone is generally smeared by reverberation, which severely degrades both the speech intelligibility and Automatic Speech Recognition (ASR) performance. Previously, we proposed a single-microphone dereverberation method, named "Harmonicity based dEReverBeration (HERB)." HERB estimates the inverse filter for an unknown room transfer function by utilizing an essential feature of speech, namely harmonic structure. In previous studies, improvements in speech intelligibility was shown solely with spectrograms, and improvements in ASR performance were simply confirmed by matched condition acoustic model. In this paper, we undertook a further investigation of HERB's potential as regards to the above two factors. First, we examined speech intelligibility by means of objective indices. As a result, we found that HERB is capable of improving the speech intelligibility to approximately that of clean speech. Second, since HERB alone could not improve the ASR performance sufficiently, we further analyzed the HERB mechanism with a view to achieving further improvements. Taking the analysis results into account, we proposed an appropriate ASR configuration and conducted experiments. Experimental results confirmed that, if HERB is used with an ASR adaptation scheme such as MLLR and a multicondition acoustic model, it is very effective for improving ASR performance even in unknown severely reverberant environments.