Nonintrusive objective measurement of speech intelligibility: A review of methodology

Objective intelligibility measurement of reverberant vocoded speech for normal-hearing listeners: Towards facilitating the development of speech enhancement algorithms for cochlear implants

Cochlear implant (CI) recipients often struggle to understand speech in reverberant environments. Speech enhancement algorithms could restore speech perception for CI listeners by removing reverberant artifacts from the CI stimulation pattern. Listening studies, either with cochlear-implant recipients or normal-hearing (NH) listeners using a CI acoustic model, provide a benchmark for speech intelligibility improvements conferred by the enhancement algorithm but are costly and time consuming. To reduce the associated costs during algorithm development, speech intelligibility could be estimated offline using objective intelligibility measures. Previous evaluations of objective measures that considered CIs primarily assessed the combined impact of noise and reverberation and employed highly accurate enhancement algorithms. To facilitate the development of enhancement algorithms, we evaluate twelve objective measures in reverberant-only conditions characterized by a gradual reduction of reverberant artifacts, simulating the performance of an enhancement algorithm during development. Measures are validated against the performance of NH listeners using a CI acoustic model. To enhance compatibility with reverberant CI-processed signals, measure performance was assessed after modifying the reference signal and spectral filterbank. Measures leveraging the speech-to-reverberant ratio, cepstral distance and, after modifying the reference or filterbank, envelope correlation are strong predictors of intelligibility for reverberant CI-processed speech.

Syllable-rate-adjusted-modulation (SRAM) predicts clear and conversational speech intelligibility

Introduction

Objectively predicting speech intelligibility is important in both telecommunication and human-machine interaction systems. The classic method relies on signal-to-noise ratios (SNR) to successfully predict speech intelligibility. One exception is clear speech, in which a talker intentionally articulates as if speaking to someone who has hearing loss or is from a different language background. As a result, at the same SNR, clear speech produces higher intelligibility than conversational speech. Despite numerous efforts, no objective metric can successfully predict the clear speech benefit at the sentence level.

Methods

We proposed a Syllable-Rate-Adjusted-Modulation (SRAM) index to predict the intelligibility of clear and conversational speech. The SRAM used as short as 1 s speech and estimated its modulation power above the syllable rate. We compared SRAM with three reference metrics: envelope-regression-based speech transmission index (ER-STI), hearing-aid speech perception index version 2 (HASPI-v2) and short-time objective intelligibility (STOI), and five automatic speech recognition systems: Amazon Transcribe, Microsoft Azure Speech-To-Text, Google Speech-To-Text, wav2vec2 and Whisper.

Results

SRAM outperformed the three reference metrics (ER-STI, HASPI-v2 and STOI) and the five automatic speech recognition systems. Additionally, we demonstrated the important role of syllable rate in predicting speech intelligibility by comparing SRAM with the total modulation power (TMP) that was not adjusted by the syllable rate.

Discussion

SRAM can potentially help understand the characteristics of clear speech, screen speech materials with high intelligibility, and convert conversational speech into clear speech.

Neural tracking measures of speech intelligibility: Manipulating intelligibility while keeping acoustics unchanged

Neural speech tracking has advanced our understanding of how our brains rapidly map an acoustic speech signal onto linguistic representations and ultimately meaning. It remains unclear, however, how speech intelligibility is related to the corresponding neural responses. Many studies addressing this question vary the level of intelligibility by manipulating the acoustic waveform, but this makes it difficult to cleanly disentangle the effects of intelligibility from underlying acoustical confounds. Here, using magnetoencephalography recordings, we study neural measures of speech intelligibility by manipulating intelligibility while keeping the acoustics strictly unchanged. Acoustically identical degraded speech stimuli (three-band noise-vocoded, ~20 s duration) are presented twice, but the second presentation is preceded by the original (nondegraded) version of the speech. This intermediate priming, which generates a "pop-out" percept, substantially improves the intelligibility of the second degraded speech passage. We investigate how intelligibility and acoustical structure affect acoustic and linguistic neural representations using multivariate temporal response functions (mTRFs). As expected, behavioral results confirm that perceived speech clarity is improved by priming. mTRFs analysis reveals that auditory (speech envelope and envelope onset) neural representations are not affected by priming but only by the acoustics of the stimuli (bottom-up driven). Critically, our findings suggest that segmentation of sounds into words emerges with better speech intelligibility, and most strongly at the later (~400 ms latency) word processing stage, in prefrontal cortex, in line with engagement of top-down mechanisms associated with priming. Taken together, our results show that word representations may provide some objective measures of speech comprehension.

Neural Tracking Measures of Speech Intelligibility: Manipulating Intelligibility while Keeping Acoustics Unchanged

Neural speech tracking has advanced our understanding of how our brains rapidly map an acoustic speech signal onto linguistic representations and ultimately meaning. It remains unclear, however, how speech intelligibility is related to the corresponding neural responses. Many studies addressing this question vary the level of intelligibility by manipulating the acoustic waveform, but this makes it difficult to cleanly disentangle effects of intelligibility from underlying acoustical confounds. Here, using magnetoencephalography (MEG) recordings, we study neural measures of speech intelligibility by manipulating intelligibility while keeping the acoustics strictly unchanged. Acoustically identical degraded speech stimuli (three-band noise vocoded, ~20 s duration) are presented twice, but the second presentation is preceded by the original (non-degraded) version of the speech. This intermediate priming, which generates a 'pop-out' percept, substantially improves the intelligibility of the second degraded speech passage. We investigate how intelligibility and acoustical structure affects acoustic and linguistic neural representations using multivariate Temporal Response Functions (mTRFs). As expected, behavioral results confirm that perceived speech clarity is improved by priming. TRF analysis reveals that auditory (speech envelope and envelope onset) neural representations are not affected by priming, but only by the acoustics of the stimuli (bottom-up driven). Critically, our findings suggest that segmentation of sounds into words emerges with better speech intelligibility, and most strongly at the later (~400 ms latency) word processing stage, in prefrontal cortex (PFC), in line with engagement of top-down mechanisms associated with priming. Taken together, our results show that word representations may provide some objective measures of speech comprehension.

EEG-based Objective Quality Assessment of Interrupted Speech

As an important dimension of speech degradation, interruption destroys the acoustic waveform information of a speech signal and seriously affects the quality rating in speech perception. Existing studies assessing the quality of interrupted speech mainly focus on subjective methods. This work used event-related potentials (ERPs) to objectively investigate the effects of two interruption factors (i.e., interruption duty cycle and interruption rate) on the perceived quality of the periodically interrupted speech. An oddball-paradigm based ERP experiment was designed, and the periodically interrupted stimuli were presented to participants in the ERP experiment. Meanwhile, participants were instructed to provide their subjective quality ratings of interrupted speech under the same interruption conditions. The experimental results showed that the degree of both interruption duty cycle and interruption rate affected the peak latency of mismatch negativity (MMN) of the collected ERP response, with a smaller MMN peak latency under conditions with a smaller interruption duty cycle or a larger interruption rate. These findings are largely consistent with the subjective quality rating results, suggesting that the peak latency of MMN response can be potentially used as an objective quality indicator of the periodically interrupted speech.

ASR-based speech intelligibility prediction: A review

Various types of methods and approaches are available to predict the intelligibility of speech signals, but many of these still suffer from two major problems: first, their required prior knowledge, which itself could limit the applicability and lower the objectivity of the method, and second, a low generalization capacity, e.g. across noise types, degradation conditions, and speech material. Automatic speech recognition (ASR) has been suggested as a machine-learning-based component of speech intelligibility prediction (SIP), aiming to ameliorate the shortcomings of other SIP methods. Since their first introduction, ASR-based SIP approaches have been developing at an increasingly rapid pace, were deployed in a range of contexts, and have shown promising performance in many scenarios. Our article provides an overview of this body of research. The main differences between competing methods are highlighted and their benefits are explained next to their limitations. We conclude with an outlook on future work and new related directions.