Hanwu Sun

Investigations into the relationship between measurable speech quality and speech recognition rate for telephony speech

In this paper, an investigation to establish a possible relationship between the performance of a telephony speech recognition system and the method for objective speech quality assessment described in ITU-T Recommendation P.862, known as Perceptual Evaluation of Speech Quality (PESQ), is presented. Experiments using various additive background noises, and at different separations between the microphone and the sound-source have been conducted to establish such a relationship. The preliminary results suggest that telephony speech recognition rates can be mapped to the mean opinion score (MOS) obtained by PESQ using a relatively simple polynomial relationship. This indicates that the PESQ MOS can act as a reliable predictor for the achievable speech recognition rates for telephony-based speech recognition systems.

Performance evaluation of adaptive dual microphone systems

In this paper, the performance of the adaptive noise cancellation method is evaluated on several possible dual microphone system (DMS) configurations. Two groups of DMS are taken into consideration with one consisting of two omnidirectional microphones and another involving directional microphones. The properties of these methods are theoretically analyzed under incoherent, coherent and diffuse noise respectively. To further investigate their achievable noise reduction performance in real situations, a series of experiments in simulated and real office environments are carried out. Some recommendations are given at the end for designing and choosing the suitable methods in real applications.

Theoretical comparisons of dual microphone systems

We compare and evaluate the various noise cancellation schemes available in what we term dual microphone systems (DMS). A DMS is a directional audio-capturing device consisting of two microphones with possibly different directional characteristics: omnidirectional; bidirectional; cardioid. A general structure is proposed to represent coherently the different schemes for the DMS. This is followed by a theoretical derivation of the performance of various DMS configurations under incoherent, coherent and diffuse noise fields. The relationships between the different configurations and some guidelines for designing DMS are also presented.

Development of a compact two-sensor directional audio-capturing device

A directional audio-capturing device is highly desirable for normal communication in a noisy environment, and to enhance the quality of speech signal for the hearing impaired. In this paper, we propose a compact directional audio-capturing device which reduces the background noise in relation to speech signal which originates from a desired direction. This device contains three functional units: (1) a differential microphone formed by two omni-directional microphones connected in an endfire orientation, (2) adaptive beamforming and (3) echo shaping of the residual output signals to further improve the signal-to-noise ratio. The adaptation is based on a control scheme using the differences in the spatial responses between the omni-directional and the differential microphone signals. A post-processing scheme using echo-shaping filter is used to further suppress any residual interferences. A real-time implementation of our proposed scheme has been constructed and its performance measured. The experiments indicate a polar pattern with good interference cancellation, with an average interference suppression of as much as 22 dB. Our experiments also indicate lower speech distortion using our proposed system. In term of speech quality, the improvement is indicated by the mean-opinion-score (MOS) scale. The MOS has been obtained from both subjective tests conducted using ten listeners, as well as using the perceptual evaluation of speech quality tool indicated in ITU-T standards document.

An adaptive microphone array with local acoustic sensitivity

In this paper, a microphone array with 3-D focal zone is proposed. The microphone array consists of one omni-directional and two uni-directional microphones. The microphone array is so constructed that a cross zone is formed such that only the sound within this zone is captured and any interferences outside the zone are effectively cancelled. The proposed framework is flexible in defining the location/size of the closed volume where the sound source of interest is located. Simulations have been carried out to demonstrate the 3-D spatial selectivity as well as the noise cancellation performance. The most important feature which differs from the previous works is that the super volumetric selectivity is realized by strategically use only three microphones, by which the overall apparatus acts as a virtual wireless close-talking microphone with confined position constrained in both distance and directions.

Experimental study of dual microphone systems

In this paper, we compare and evaluate the various noise cancellation schemes available in what we term a dual microphone system. A dual microphone system (DMS) is a composite directional audio-capturing device which consists of two microphones, each microphone having possibly different directional characteristics, e.g., omnidirectional, bidirectional or cardioid. By recasting the various combinations of two microphones for a DMS into a coherent and familiar framework of generalized sidelobe canceller (GSC), we subsequently derive the expected noise reduction of various structures under incoherent, coherent and diffuse noise fields, followed by a series of experiments in a typical office environment. These results are indicative of the achievable reduction of noise in real applications. The relationship and differences between the various methods are also discussed

A robust adaptive cross microphone array

In this paper, a robust adaptive microphone array system which has a cone-shaped directionality pattern is presented. The array consists of two linear microphone arrays, arranged in a cross fashion, and incorporates the CCAF-NCAF robust beamformer scheme of Hoshuyama et al (IEEE Trans. on Signal Proc., vol. 47 no. 10, pp. 2677-2684, 1999). As indicated in the simulation results, the proposed cross microphone array overcomes problem of spatial ambiguity (unlike a conventional linear array), and has the added advantage of three-dimensional interference cancellation. In addition, the system also benefits from the accelerated convergence speed due to a reduced allowable capture region.

Analysis of an Adaptive Filter-bank for Harmonic Measurement and Estimation

This paper presents a statistical analysis of a modified adaptive filter-bank when applied to the estimation of harmonic signals. The application of such a filter-bank for the tracking of the fundamental frequency, estimation of the amplitude and phase of each harmonic component is investigated under various conditions: different A/D quantization errors, the fundamental frequency having small time-varying variations and, finally, mismatches in the filter-bank lengths to the number of the harmonics. The analysis results indicate that this filter-bank is very robust for the estimation of amplitude and phase of harmonic signals.