A comparative intelligibility study of single-microphone noise reduction algorithms

TLDR

The study corrupted IEEE sentences and consonants with babble, car, street, and train noise at 0 and 5 dB SNR, processed them with eight speech‑enhancement methods spanning spectral subtraction, sub‑space, statistical‑model, and Wiener classes, and evaluated the enhanced speech with normal‑hearing listeners. Except for one noise condition, none of the eight algorithms significantly improved speech intelligibility or the place‑feature score, and the algorithms that excelled in overall quality did not perform best in intelligibility; notably, the sub‑space method, previously the poorest in quality, preserved intelligibility, indicating that improving place and manner feature scores is essential for intelligibility gains.

Abstract

The evaluation of intelligibility of noise reduction algorithms is reported. IEEE sentences and consonants were corrupted by four types of noise including babble, car, street and train at two signal-to-noise ratio levels (0 and 5dB), and then processed by eight speech enhancement methods encompassing four classes of algorithms: spectral subtractive, sub-space, statistical model based and Wiener-type algorithms. The enhanced speech was presented to normal-hearing listeners for identification. With the exception of a single noise condition, no algorithm produced significant improvements in speech intelligibility. Information transmission analysis of the consonant confusion matrices indicated that no algorithm improved significantly the place feature score, significantly, which is critically important for speech recognition. The algorithms which were found in previous studies to perform the best in terms of overall quality, were not the same algorithms that performed the best in terms of speech intelligibility. The subspace algorithm, for instance, was previously found to perform the worst in terms of overall quality, but performed well in the present study in terms of preserving speech intelligibility. Overall, the analysis of consonant confusion matrices suggests that in order for noise reduction algorithms to improve speech intelligibility, they need to improve the place and manner feature scores.