Matrix sentence intelligibility prediction using an automatic speech recognition system

Schädler, Marc René; Warzybok, Anna; Hochmuth, Sabine; Kollmeier, Birger

doi:10.3109/14992027.2015.1061708

Cited by 51 publications

(50 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, a number of works have attempted to partly or fully predict intelligibility using data-driven methods. One approach in this direction has been to use an Automatic Speech Recognition (ASR) system to transcribe degraded sentences, using the error rate as a measure of intelligibility [48], [49]. Another data-driven approach has been to non-intrusively estimate the output of an intrusive SIP algorithm [50], [51], [52], [53].…”

Section: Introductionmentioning

confidence: 99%

Nonintrusive Speech Intelligibility Prediction Using Convolutional Neural Networks

Andersen

Haan

Jensen

2018

IEEE/ACM Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

Speech Intelligibility Prediction (SIP) algorithms are becoming popular tools within the development and operation of speech processing devices and algorithms. However, many SIP algorithms require knowledge of the underlying clean speech; a signal that is often not available in real-world applications. This has led to increased interest in non-intrusive SIP algorithms, which do not require clean speech to make predictions. In this paper we investigate the use of Convolutional Neural Networks (CNNs) for non-intrusive SIP. To do so, we utilize a CNN architecture that shows similarities to existing SIP algorithms, in terms of computational structure, and which allows for easy and meaningful visualization and interpretation of trained weights. We evaluate this architecture using a large dataset obtained by combining datasets from the literature. The proposed method shows high prediction performance when compared with four existing intrusive and non-intrusive SIP algorithms. This demonstrates the potential of deep learning for speech intelligibility prediction.

show abstract

Section: Introductionmentioning

confidence: 99%

Nonintrusive Speech Intelligibility Prediction Using Convolutional Neural Networks

Andersen

Haan

Jensen

2018

IEEE/ACM Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

show abstract

“…The long-term spectrum of the ICRA5-250 noise was shaped according to the international long-term average speech spectrum (Wagener, Brand, and Kollmeier 2006). Previous studies showed that the long-term spectrum of the SSN derived from the Oldenburg sentence test and the spectrum of the ICRA1 noise (stationary noise with the same long-term spectrum as ICRA5-250) resulted in the same SRTs (Hochmuth et al 2015;Wagener and Brand 2005). We thus assume that any differences in SRT observed between SSN and ICRA5-250 cannot be explained on the basis of spectral differences between the two maskers.…”

Section: Speech Recognitionmentioning

confidence: 99%

“…The LP and HP condition were generated by filtering the speech and noise signals using a 1024 th order FFT filter with cut-off frequencies of 1.5 and 1.65 kHz, respec-tively. The cut-off frequencies of the filters were based on dividing the speech signals in two parts of comparable importance, considering the SII band importance function for speech in noise (ANSI 1997;Hochmuth et al 2015). The sum of the SII octave band importance weights is 0.457 for frequencies 0.25, 0.5, and 1 kHz (roughly corresponding to the LP condition in this study) and 0.543 for frequencies 2, 4, and 8 kHz (roughly corresponding to the HP condition in this study).…”

Section: Speech Recognitionmentioning

confidence: 99%

“…Hence, supra-threshold deficits need to be explored and included in the SII model. Another modelling approach which, aside from including elevated pure tone threshold, accounts for supra-threshold deficits associated with reduced intensity discrimination or frequency selectivity was proposed by Schädler et al (2015;. This type of model, or more biophysical models accounting for auditory-nerve deficits (e.g., Zilany, Bruce, and Carney 2014;Verhulst, Altoè, and Vasilkov 2018), can be used to examine the influence of elevated pure tone thresholds as well as supra-threshold deficits on speech intelligibility in stationary and modulated noise in future studies.…”

Section: Sii Predictionsmentioning

confidence: 99%

See 1 more Smart Citation

Contributions of Low- and High-Frequency Sensorineural Hearing Deficits to Speech Intelligibility in Noise

Verhulst

Warzybok

2018

Preprint

Self Cite

View full text Add to dashboard Cite

This study investigates the role of hearing sensitivity, age and supra-threshold hearing to speech recognition in quiet, stationary, and modulated noise. Three participant groups were tested: young normal hearing (yNH), older normalhearing (oNH), and older hearing-impaired (oHI) listeners. The relative importance of near and supra-threshold hearing deficits was assessed in different frequency regions by measuring speech reception thresholds (SRTs) using broad- . CC-BY-NC-ND 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint . http://dx.doi.org/10.1101/358127 doi: bioRxiv preprint first posted online Jun. 29, 2018; present for BB, LP and HP conditions and related to the modulated (not stationary) SRT performance. (iv) Sensitivity to temporal-envelopes is important for HF masking release and SRTs in modulated noise, whereas spectral resolution is a better predictor of LF speech recognition in modulated noise. This study identified the role of near and supra-threshold cues in the perception of amplitude-modulation and speech-in-noise, and can steer the development of hearing-aid algorithms targeted to mediate supra-threshold hearing deficits.

show abstract

“…

A framework for simulating auditory discrimination experiments, based on an approach from Sch€ adler, Warzybok, Hochmuth, and Kollmeier [(2015). Int.

…”

mentioning

confidence: 99%

A simulation framework for auditory discrimination experiments: Revealing the importance of across-frequency processing in speech perception

Schädler

Warzybok

Ewert

et al. 2016

The Journal of the Acoustical Society of America

Self Cite

View full text Add to dashboard Cite

A framework for simulating auditory discrimination experiments, based on an approach from Schädler, Warzybok, Hochmuth, and Kollmeier [(2015). Int. J. Audiol. 54, 100-107] which was originally designed to predict speech recognition thresholds, is extended to also predict psychoacoustic thresholds. The proposed framework is used to assess the suitability of different auditory-inspired feature sets for a range of auditory discrimination experiments that included psychoacoustic as well as speech recognition experiments in noise. The considered experiments were 2 kHz tone-in-broadband-noise simultaneous masking depending on the tone length, spectral masking with simultaneously presented tone signals and narrow-band noise maskers, and German Matrix sentence test reception threshold in stationary and modulated noise. The employed feature sets included spectro-temporal Gabor filter bank features, Mel-frequency cepstral coefficients, logarithmically scaled Mel-spectrograms, and the internal representation of the Perception Model from Dau, Kollmeier, and Kohlrausch [(1997). J. Acoust. Soc. Am. 102(5), 2892-2905]. The proposed framework was successfully employed to simulate all experiments with a common parameter set and obtain objective thresholds with less assumptions compared to traditional modeling approaches. Depending on the feature set, the simulated reference-free thresholds were found to agree with-and hence to predict-empirical data from the literature. Across-frequency processing was found to be crucial to accurately model the lower speech reception threshold in modulated noise conditions than in stationary noise conditions.

show abstract

Matrix sentence intelligibility prediction using an automatic speech recognition system

Cited by 51 publications

References 29 publications

Nonintrusive Speech Intelligibility Prediction Using Convolutional Neural Networks

Nonintrusive Speech Intelligibility Prediction Using Convolutional Neural Networks

Contributions of Low- and High-Frequency Sensorineural Hearing Deficits to Speech Intelligibility in Noise

A simulation framework for auditory discrimination experiments: Revealing the importance of across-frequency processing in speech perception

Contact Info

Product

Resources

About