No abstract

A new, single-ended, i.e. reference-free measure for the prediction of perceived listening effort of noisy speech is presented. It is based on phoneme posterior probabilities (or posteriorgrams) obtained from a deep neural network of an automatic speech recognition system. Additive noisy or other distortions of speech tend to smear the posteriorgrams. The smearing is quantified by a performance measure, which is used as a predictor for the perceived listening effort required to understand the noisy speech. The proposed measure was evaluated using a database obtained from the subjective evaluation of noise reduction algorithms of commercial hearing aids. Listening effort ratings of processed noisy speech samples were gathered from 20 hearing-impaired subjects. Averaged subjective ratings were compared with corresponding predictions computed by the proposed new method, the ITU-T standard P.563 for single-ended speech quality assessment, the American National Standard ANIQUE+ for single-ended speech quality assessment, and a single-ended SNR estimator. The proposed method achieved a good correlation with mean subjective ratings and clearly outperformed the standard speech quality measures and the SNR estimator.

No abstract

In recent years, automatic speech recognition (ASR) systems gradually decreased (and for some tasks closed) the gap between human and automatic speech recognition. However, it is unclear if similar performance implies humans and ASR systems to rely on similar signal cues. In the current study, ASR and HSR are compared using speech material from a matrix sentence test mixed with either a stationary speech-shaped noise (SSN) or amplitude-modulated SSN. Recognition performance of HSR and ASR is measured in term of the speech recognition threshold (SRT), i.e., the signal-to-noise ratio with 50% recognition rate and by comparing psychometric functions. ASR results are obtained with matched-trained DNN-based systems that use FBank features as input and compared to results obtained from eight normal-hearing listeners and two established models of speech intelligibility. For both maskers, HSR and ASR achieve similar SRTs with an average deviation of only 0.4 dB. A relevance propagation algorithm is applied to identify features relevant for ASR. The analysis shows that relevant features coincide either with spectral peaks of the speech signal or with dips of the noise masker, indicating that similar cues are important in HSR and ASR.

No abstract