Auditory and auditory-visual intelligibility of speech in fluctuating maskers for normal-hearing and hearing-impaired listeners

Bernstein, Jeffrey; Grant, Ken W.

doi:10.1121/1.3110132

Cited by 200 publications

(280 citation statements)

References 60 publications

Supporting

Mentioning

239

Contrasting

Order By: Relevance

“…Alternatively, it has been proposed that benefit of listening in the dips of modulated noise decreases with increasing signal-to-noise ratio (SNR). As speech perception in noise is often tested at equated intelligibility levels, hearing-impaired participants are usually listening at higher SNRs that does not allow listening in the dips (Bernstein & Grant 2009). Lunner and Sundewall-Thorén (2007) showed in older adults using hearing aids that WMC accounted for about 40% of the variance in speech recognition in modulated noise.…”

Section: Noisementioning

confidence: 99%

Exploring Cognitive Spare Capacity : Executive Processing of Degraded Speech

Mishra¹

2014

View full text Add to dashboard Cite

Cognitive resources, specifically working memory capacity are used for listening to speech, especially in noise. Cognitive resources are limited, and if listeners allocate a greater share of these resources to recovering the input signal in noise, fewer resources are available for interpreting and encoding its linguistic content. Although the importance of CSC for individual success in communicative situations has been acknowledged, this concept has not hitherto been explored experimentally. In this thesis, a CSC test (CSCT) was developed and administered to young adults with normal hearing and older adults with age-related hearing loss. CSCT required executive processing of speech at different memory loads with and without visual cues in different noise conditions. A free recall task using the same material was administered for comparison purposes and a battery of cognitive tests was administered to understand the relation between CSC and established cognitive concepts. The aims of the thesis were to investigate how CSC is influenced by 1) different executive demands and memory loads; 2) background noise; 3) visual cues; 4) aging and concomitant hearing loss. The results showed that 1) CSC was sensitive to memory load, and updating demands reduced CSC more than inhibition demands; 2) CSC was reduced in background noise compared to quiet; 3) visual cues enhanced CSC especially in noise; 4) CSC was reduced with ageing and concomitant hearing loss especially when visual cues were absent, memory demands were increased and background noise was speech-like. The main finding of this thesis was that visual cues enhanced CSC for older individuals with hearing loss, specifically in adverse listening conditions. This demonstrates the importance of audiovisual testing in audiological assessment. Further, specific cognitive resources depleted during listening in noise were at least partially compensated by other cognitive functions. This thesis is the first step towards a theoretical understanding of CSC and in future, tests of CSC may play a crucial role in planning rehabilitation of persons with hearing loss.

show abstract

Section: Noisementioning

confidence: 99%

Exploring Cognitive Spare Capacity : Executive Processing of Degraded Speech

Mishra¹

2014

View full text Add to dashboard Cite

show abstract

“…Listeners with normal hearing (NH) can take advantage of gaps in these fluctuating maskers. They are able to ''listen in the dips'' of temporally varying noise to extract information about the speech signal, thereby experiencing improvement in speech recognition (e.g., Bernstein & Grant, 2009;Festen & Plomp, 1990;Jin & Nelson, 2006). Such performance improvement in the presence of fluctuating compared to steady-state noise conditions is known as masking release.…”

mentioning

confidence: 99%

Masking Release and Modulation Interference in Cochlear Implant and Simulation Listeners

Jin

Nie

2013

Am J Audiol

View full text Add to dashboard Cite

Purpose: To examine the effects of temporal and spectral interference of masking noise on sentence recognition for listeners with cochlear implants (CI) and normal-hearing persons listening to vocoded signals that simulate signals processed through a CI (NH-Sim). Method: NH-Sim and CI listeners participated in the experiments using speech and noise that were processed by bandpass filters. Depending on the experimental condition, the spectra of the maskers relative to that of speech were set to be completely embedded with, partially overlapping, or completely separate from, the speech. The maskers were either steady or amplitude modulated and were presented at +10 dB signal-to-noise ratio. Results: NH-Sim listeners experienced progressively more masking as the masker became more spectrally overlapping with speech, whereas CI listeners experienced masking even when the masker was spectrally remote from the speech signal. Both the NH-Sim and CI listeners experienced significant modulation interference when noise was modulated at a syllabic rate (4 Hz), suggesting that listeners may experience both modulation interference and masking release. Thus, modulated noise has mixed and counteracting effects on speech perception. Conclusion: When the NH-Sim and CI listeners with poor spectral resolution were tested using syllabic-like rates of modulated noise, they tended to integrate or confuse the noise with the speech, causing an increase in speech errors. Optional training programs might be useful for CI listeners who show more difficulty understanding speech in noise.Key Words: cochlear implants, hearing loss, speech perception T ypical environmental noises such as background conversations are temporally varying in frequency and amplitude. Listeners with normal hearing (NH) can take advantage of gaps in these fluctuating maskers. They are able to ''listen in the dips'' of temporally varying noise to extract information about the speech signal, thereby experiencing improvement in speech recognition (e.g., Bernstein & Grant, 2009;Festen & Plomp, 1990;Jin & Nelson, 2006). Such performance improvement in the presence of fluctuating compared to steady-state noise conditions is known as masking release. Previous studies have reported that NH listeners' speech recognition scores could improve by as much as 80 percentage points when noise was modulated versus steady (Jin & Nelson, 2006). However, significant masking release reduction or no masking release has been found in cochlear implant (CI) users or in NH listeners identifying vocoded speech that simulates speech processed by a CI device (NH-Sim; Fu & Nogaki, 2004;Kwon, Perry, Wilhelm, & Healy 2012;Nelson & Jin, 2004;Nelson, Jin, Carney, & Nelson, 2003;Qin & Oxenham, 2003;Stickney, Zeng, Litovsky, & Assmann, 2004). For example, Nelson and colleagues (Nelson & Jin, 2004;Nelson et al., 2003) compared the performance of three listener groups (NH, CI, and NH-Sim) for sentence recognition in the presence of different masking noises, including steady-state noise and gated noise m...

show abstract

“…This suggests that the selective compression did not provide masking release Additional analysis was conducted to assess whether the use of selective compression improved performance relative to that obtained with the unprocessed stimuli, which were processed (for all phonetic segments) using the log-mapping function. Two-way ANOVA, run on the scores obtained in the SSN conditions, indicated significant effect (F [1,6] As shown in Figure 1.3, selective compression improved performance by nearly 15 percentage points at 5 dB SNR and by approximately 10-15 percentage points at 10 dB SNR. The improvement in performance was found to be consistent for both types of maskers.…”

Section: Resultsmentioning

confidence: 99%

“…Two-way ANOVA with repeated measures was used to assess the effect of masker type. The control noisy stimuli (shown in Figure 5.1 with blue bars) showed no significant effect of masker type (F [1,6] = 0.036, p = 0.89). No significant interaction (F [1,6] =2.1, p = 0.176) was found between SNR level and masker type.…”

Section: Resultsmentioning

confidence: 99%

“…The control noisy stimuli (shown in Figure 5.1 with blue bars) showed no significant effect of masker type (F [1,6] = 0.036, p = 0.89). No significant interaction (F [1,6] =2.1, p = 0.176) was found between SNR level and masker type. Performance with steady noise was comparable to that attained by the two-talker masker, consistent with findings reported in other cochlear implant studies Figure 1.1 Mean speech recognition scores as a function of SNR level for the var-ious masker (TT=two-talker and SSN=steady noise) and channel conditions.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

To Improve Speech Recognition in Noise for Cochlear Implant Users

Goyal¹

2013

IOSR-JEEE

View full text Add to dashboard Cite

Abstract:The hypothesis that when listening to speech in fluctuating maskers, CI users can not fuse the pieces of the message over temporal gaps because they are not able to perceive reliably the acoustic landmarks introduced by obstruent consonants (e.g., stops I. The Clinical Program 1.1Experiment 1: Masking Releases for Cochlear Implant UsersMethods Subjects A total of seven postlingually deafened Clarion CII implant users participated in this experiment. All subjects had at least 3 years of experience with their implant devices. Most subjects visited our lab two times. The biographical data for each subject are given in Table 1.1. StimuliThe speech material consisted of sentences taken from the IEEE database . All sentences were produced by a male speaker. The sentences were recorded in a sound-proof booth in our lab at a 25 kHz sampling rate. Two types of maskers were used. The first was speech-shaped noise, which is continuous (steadystate) and had the same long-term spectrum as the test sentences in the IEEE corpus. The second masker was a two-talker competing speech (female) recorded in our lab. Two long sentences, produced by a female talker, were used from the IEEE database. This was done to ensure that the target signal was always shorter (in duration) than the masker. The IEEE sentences were manually segmented into two broad phonetic classes: (a) the obstruent consonants which included the stops, fricatives and affricates, and (b) The sonorant sounds which included vowels, semivowels and nasals. More detailed description can be found in [12]. (C) Signal Processing Signals were first processed through a pre-emphasis filter (2000 Hz cutoff), with a 3 dB/octave roll off, and then band pass filtered into 16 channels (according to subjects' condition) using sixth-order Butterworth filters. Logarithmic filter spacing was used to allocate the channels across a 300-5500 Hz bandwidth. The envelope of the signal was extracted by full-wave rectification and low-pass filtering (second-order Butterworth) with a 400 Hz cutoff frequency. The envelopes in each channel were log compressed to the subject's electrical dynamic range. The speech stimuli were generated using the above algorithm in two different conditions. In the first control condition, the corrupted speech stimuli were left unaltered. That is, the obstruent consonants remained corrupted by the maskers (baseline). In the second

show abstract

Auditory and auditory-visual intelligibility of speech in fluctuating maskers for normal-hearing and hearing-impaired listeners

Cited by 200 publications

References 60 publications

Exploring Cognitive Spare Capacity : Executive Processing of Degraded Speech

Exploring Cognitive Spare Capacity : Executive Processing of Degraded Speech

Masking Release and Modulation Interference in Cochlear Implant and Simulation Listeners

To Improve Speech Recognition in Noise for Cochlear Implant Users

Contact Info

Product

Resources

About