Multi-talker Babble Research Articles

An algorithm that generalizes to novel noise segments to improve speech intelligibility for hearing-impaired listeners

Machine-learning algorithms to extract speech from background noise hold considerable promise for alleviating limitations associated with hearing impairment. One of the most important considerations for implementing these algorithms into devices such as hearing aids and cochlear implants involves their ability to generalize to conditions not employed during the training stage. A major challenge involves the generalization to novel noise segments. In the current study, sentences were extracted from multi-talker babble and from cafeteria noise using an algorithm that estimates the ideal ratio mask by employing deep neural networks. Importantly, the algorithm was trained on segments of noise and tested using entirely different segments of the same nonstationary noise type. The training set was expanded through a noise-perturbation technique. Substantial benefit was observed for hearing-impaired listeners in both noise types, despite the use of unseen noise segments during the operational stage. Interestingly, normal-hearing listeners displayed benefit in babble but not in cafeteria noise. These results highlight the importance of evaluating these algorithms not only in human subjects, but in members of the actual target population. [Work supported by NIH.]

Segmental contribution to the intelligibility of noise-suppressed speech

Many studies suggested that most speech enhancement algorithms do not improve the subjective intelligibility scores of noise-suppressed speech when presented to normal-hearing listeners. Nevertheless, the reason for lacking intelligibility improvement is still unclear. This study assessed the segmental contribution to the intelligibility of noise-suppressed speech. Mandarin sentences were corrupted by steady-state speech-spectrum shaped noise and multi-talker babble. Stimuli were synthesized by using noise-suppressed speech processed by three types of speech enhancement algorithms and a noise-replacement paradigm. The noise-replacement paradigm preserved selected speech segments (i.e., vowel-only and vowel-plus-consonant-onset), and replaced the rest with noise. Listening experiments were performed to collect the intelligibility scores from normal-hearing listeners. Experimental results showed that the selectively synthesized stimuli (i.e., vowel-only and vowel-plus-consonant-onset) from noise-suppressed speech may be more intelligible than those from noise-corrupted speech. However, this benefit was only observed when the speech signal was corrupted by speech-spectrum shaped noise but not by babble masker. This work suggested that the segmental distortion caused by speech enhancement algorithms may affect the information integration from noise-suppressed speech, and high-level top-down processing may account for the deficiency of the present speech enhancement algorithms for improving subjective intelligibility performance.

Talker variation in the perception of speech in noise

Speech communication commonly occurs in the presence of noise. Research on the perception of speech in noise has largely focused on the perceptual availability of target phonetic information in the presence of various noise maskers (e.g., white noise, speech-shaped noise, temporally modulated noise, and multi-talker babble). Previous work has shown that a glimpsing model and a modified Articular Index model closely approximate overall accuracy of noise-masked speech perception (Cooke, 2006, JASA; Allen, 2005; JASA). Some recent work has focused on variation in confusion patterns across listeners (Silbert, 2012, JASA, 2014, Lab. Phon.) and on difference in perceptual error rates across and within consonant categories (Toscano & Allen, forthcoming, JSLHR). The present work focuses on listener and stimulus-talker variation in the identification of consonants. Eleven listeners identified numerous tokens of each of four consonants (t, d, s, z) in CV syllables produced by 20 talkers (10 male, 10 female) masked ...

Task demands and cognitive abilities impact listening effort for older adult hearing aid users

Understanding speech in background noise requires considerable effort for older adults with hearing loss, even when wearing well-fit hearing aids, but little is known about how effort varies with changes in speech recognition task demands in this population. Following research showing that effort can be indexed by changes in the autonomic pupil dilation response, the current study recorded pupillometry as older adults wearing hearing aids completed the Words-In-Noise test (e.g., Wilson et al., 2007). Using a single loudspeaker at 0 degrees azimuth, participants listened to and repeated monosyllabic words in multi-talker babble at seven signal-to-noise ratios (0 to + 24 dB SNR). A nonlinear relationship between word recognition score and pupil dilation was observed, with maximal effort at moderate levels of speech recognition difficulty. Individual differences in cognitive abilities (e.g., working memory, vocabulary knowledge) modulated this pattern of results. These findings highlight an important dissociation between listening effort and speech recognition task difficulty. Implications for evaluating changes in effort with speech-perception training are discussed. [Work supported, in part, by NIH/NIDCD.]

Speech recognition and listening effort across various conditions in adults with aphasia

Some of the common complaints among adults with aphasia are that it is difficult to understand speech in noise and that listening can be very effortful. Nonetheless, there has been limited research to date on speech recognition in noise or listening effort for adults with aphasia. Stanfield & Richie (2014) investigated speech recognition under various listening conditions for adults with non-fluent aphasia. Participants with aphasia performed better in quiet as compared to noise, as expected, and better in noise that consisted of multi-talker babble as compared to one competing talker. They also showed modest benefit from the addition of visual cues to speech over auditory-only speech recognition. However, subjective reports of listening effort were mixed and did not line up with performance on the various tests of speech recognition. In the present study, the relationship between speech recognition under various listening conditions and an objective measure of listening effort, response time, was examined. Participants’ word and sentence recognition under auditory-only and auditory-visual conditions, in quiet and in two types of noise, will be reported. The potential clinical significance to better understanding listening effort in adults with aphasia will be discussed as well.

Relationship Between Speech Intelligibility and Speech Comprehension in Babble Noise.

The authors investigated the relationship between the intelligibility and comprehension of speech presented in babble noise. Forty participants listened to French imperative sentences (commands for moving objects) in a multitalker babble background for which intensity was experimentally controlled. Participants were instructed to transcribe what they heard and obey the commands in an interactive environment set up for this purpose. The former test provided intelligibility scores and the latter provided comprehension scores. Collected data revealed a globally weak correlation between intelligibility and comprehension scores (r = .35, p < .001). The discrepancy tended to grow as noise level increased. An analysis of standard deviations showed that variability in comprehension scores increased linearly with noise level, whereas higher variability in intelligibility scores was found for moderate noise level conditions. These results support the hypothesis that intelligibility scores are poor predictors of listeners' comprehension in real communication situations. Intelligibility and comprehension scores appear to provide different insights, the first measure being centered on speech signal transfer and the second on communicative performance. Both theoretical and practical implications for the use of speech intelligibility tests as indicators of speakers' performances are discussed.

Cortical activity predicts which older adults recognize speech in noise and when.

Speech recognition in noise can be challenging for older adults and elicits elevated activity throughout a cingulo-opercular network that is hypothesized to monitor and modify behaviors to optimize performance. A word recognition in noise experiment was used to test the hypothesis that cingulo-opercular engagement provides performance benefit for older adults. Healthy older adults (N = 31; 50-81 years of age; mean pure tone thresholds <32 dB HL from 0.25 to 8 kHz, best ear; species: human) performed word recognition in multitalker babble at 2 signal-to-noise ratios (SNR = +3 or +10 dB) during a sparse sampling fMRI experiment. Elevated cingulo-opercular activity was associated with an increased likelihood of correct recognition on the following trial independently of SNR and performance on the preceding trial. The cingulo-opercular effect increased for participants with the best overall performance. These effects were lower for older adults compared with a younger, normal-hearing adult sample (N = 18). Visual cortex activity also predicted trial-level recognition for the older adults, which resulted from discrete decreases in activity before errors and occurred for the oldest adults with the poorest recognition. Participants demonstrating larger visual cortex effects also had reduced fractional anisotropy in an anterior portion of the left inferior frontal-occipital fasciculus, which projects between frontal and occipital regions where activity predicted word recognition. Together, the results indicate that older adults experience performance benefit from elevated cingulo-opercular activity, but not to the same extent as younger adults, and that declines in attentional control can limit word recognition.

Availability of binaural cues for pediatric bilateral cochlear implant recipients.

Bilateral implant recipients theoretically have access to binaural cues. Research in postlingually deafened adults with cochlear implants (CIs) indicates minimal evidence for true binaural hearing. Congenitally deafened children who experience spatial hearing with bilateral CIs, however, might perceive binaural cues in the CI signal differently. There is limited research examining binaural hearing in children with CIs, and the few published studies are limited by the use of unrealistic speech stimuli and background noise. The purposes of this study were to (1) replicate our previous study of binaural hearing in postlingually deafened adults with AzBio sentences in prelingually deafened children with the pediatric version of the AzBio sentences, and (2) replicate previous studies of binaural hearing in children with CIs using more open-set sentences and more realistic background noise (i.e., multitalker babble). The study was a within-participant, repeated-measures design. The study sample consisted of 14 children with bilateral CIs with at least 25 mo of listening experience. Speech recognition was assessed using sentences presented in multitalker babble at a fixed signal-to-noise ratio. Test conditions included speech at 0° with noise presented at 0° (S0N0), on the side of the first CI (90° or 270°) (S0N1stCI), and on the side of the second CI (S0N2ndCI) as well as speech presented at 0° with noise presented semidiffusely from eight speakers at 45° intervals. Estimates of summation, head shadow, squelch, and spatial release from masking were calculated. Results of test conditions commonly reported in the literature (S0N0, S0N1stCI, S0N2ndCI) are consistent with results from previous research in adults and children with bilateral CIs, showing minimal summation and squelch but typical head shadow and spatial release from masking. However, bilateral benefit over the better CI with speech at 0° was much larger with semidiffuse noise. Congenitally deafened children with CIs have similar availability of binaural hearing cues to postlingually deafened adults with CIs within the same experimental design. It is possible that the use of realistic listening environments, such as semidiffuse background noise as in Experiment II, would reveal greater binaural hearing benefit for bilateral CI recipients. Future research is needed to determine whether (1) availability of binaural cues for children correlates with interaural time and level differences, (2) different listening environments are more sensitive to binaural hearing benefits, and (3) differences exist between pediatric bilateral recipients receiving implants in the same or sequential surgeries.

Masked sentence recognition assessed at ascending target-to-masker ratios: modest effects of repeating stimuli.

Masked sentence recognition is typically evaluated by presenting a novel stimulus on each trial. As a consequence, experiments calling for replicate estimates in multiple conditions require large corpora of stimuli. The present study evaluated the consequences of repeating sentence-plus-masker pairs at ascending target-to-masker ratios (TMRs). The hypothesis was that performance on each trial would be consistent with the cues available to the listener at the associated TMR, resulting in similar estimates of threshold and slope for procedures using novel versus repeated sentences within an ascending-TMR block of trials. A group of 37 normal-hearing young adults participated. Each listener was tested in the presence of one of three maskers: a multitalker babble, a speech-shaped noise, or an amplitude-modulated speech-shaped noise. There were two data collection procedures, both proceeding in blocks of trials with ascending TMRs. The novel-stimulus procedure used five lists of AzBio sentences, one presented at each of five TMRs, with a novel sentence and masker sample on each trial. The repeated-stimulus procedure used a single list of AzBio sentences, with each sentence presented at multiple TMRs, progressing from low to high; each sentence was paired with a single masker sample, such that only the TMR changed within blocks of repeated stimuli. Listeners completed one run with the novel-stimulus procedure and five runs with the repeated-stimulus procedure. The resulting values of percent correct at each TMR were fitted with a logit function to estimate threshold and psychometric function slope. The novel- and repeated-stimulus procedures resulted in generally similar data patterns. After controlling for effects related to the order in which listeners completed the six data collection runs, mean thresholds were slightly higher (<0.5 dB) for the repeated-stimulus procedure than the novel-stimulus procedure in all three maskers. Function slopes for the multitalker babble and amplitude-modulated noise maskers were slightly shallower using the repeated-stimulus than the novel-stimulus procedure, but slopes were comparable for the speech-shaped noise. The quality of psychometric function fits was significantly better for the repeated-stimulus than the novel-stimulus procedure, even when comparing a single run of the repeated-stimulus procedure (using one list) to a run of the novel-stimulus procedure (using five lists). Repeating sentences at ascending TMRs is an efficient method for estimating thresholds and psychometric function slopes, both in terms of the number of sentences and the number of trials.

How "proficient" is proficient? Bilingual listeners' recognition of English words in noise.

Shi (2011, 2013) obtained sensitivity/specificity measures of bilingual listeners' English and relative proficiency ratings as the predictor of English word recognition in quiet. The current study investigated how relative proficiency predicted word recognition in noise. Forty-two monolingual and 168 bilingual normal-hearing listeners were included. Bilingual listeners rated their proficiency in listening, speaking, and reading in English and in the other language using an 11-point scale. Listeners were presented with 50 English monosyllabic words in quiet at 45 dB HL and in multitalker babble with a signal-to-noise ratio of +6 and 0 dB. Data in quiet confirmed Shi's (2013) finding that relative proficiency with or without dominance predicted well whether bilinguals performed on par with the monolingual norm. Predicting the outcome was difficult for the 2 noise conditions. To identify bilinguals whose performance fell below the normative range, dominance per se or a combination of dominance and average relative proficiency rating yielded the best sensitivity/specificity and summary measures, including Youden's index. Bilinguals' word recognition is more difficult to predict in noise than in quiet; however, proficiency and dominance variables can predict reasonably well whether bilinguals may perform at a monolingual normative level.

Functional neuroanatomy of auditory scene analysis in Alzheimer's disease.

Auditory scene analysis is a demanding computational process that is performed automatically and efficiently by the healthy brain but vulnerable to the neurodegenerative pathology of Alzheimer's disease. Here we assessed the functional neuroanatomy of auditory scene analysis in Alzheimer's disease using the well-known ‘cocktail party effect’ as a model paradigm whereby stored templates for auditory objects (e.g., hearing one's spoken name) are used to segregate auditory ‘foreground’ and ‘background’. Patients with typical amnestic Alzheimer's disease (n = 13) and age-matched healthy individuals (n = 17) underwent functional 3T-MRI using a sparse acquisition protocol with passive listening to auditory stimulus conditions comprising the participant's own name interleaved with or superimposed on multi-talker babble, and spectrally rotated (unrecognisable) analogues of these conditions. Name identification (conditions containing the participant's own name contrasted with spectrally rotated analogues) produced extensive bilateral activation involving superior temporal cortex in both the AD and healthy control groups, with no significant differences between groups. Auditory object segregation (conditions with interleaved name sounds contrasted with superimposed name sounds) produced activation of right posterior superior temporal cortex in both groups, again with no differences between groups. However, the cocktail party effect (interaction of own name identification with auditory object segregation processing) produced activation of right supramarginal gyrus in the AD group that was significantly enhanced compared with the healthy control group. The findings delineate an altered functional neuroanatomical profile of auditory scene analysis in Alzheimer's disease that may constitute a novel computational signature of this neurodegenerative pathology.

An analysis of error patterns in children's backward digit recall in noise.

The purpose of the study was to determine whether perceptual masking or cognitive processing accounts for a decline in working memory performance in the presence of competing speech. The types and patterns of errors made on the backward digit span in quiet and multitalker babble at -5 dB signal-to-noise ratio (SNR) were analyzed. The errors were classified into two categories: item (if digits that were not presented in a list were repeated) and order (if correct digits were repeated but in an incorrect order). Fifty five children with normal hearing were included. All the children were aged between 7 years and 10 years. Repeated measures of analysis of variance (RM-ANOVA) revealed the main effects for error type and digit span length. In terms of listening condition interaction it was found that the order errors occurred more frequently than item errors in the degraded listening condition compared to quiet. In addition, children had more difficulty recalling the correct order of intermediate items, supporting strong primacy and recency effects. Decline in children's working memory performance was not primarily related to perceptual difficulties alone. The majority of errors was related to the maintenance of sequential order information, which suggests that reduced performance in competing speech may result from increased cognitive processing demands in noise.

Raspberry, not a car: context predictability and a phonological advantage in early and late learners' processing of speech in noise.

Second language learners perform worse than native speakers under adverse listening conditions, such as speech in noise (SPIN). No data are available on heritage language speakers’ (early naturalistic interrupted learners’) ability to perceive SPIN. The current study fills this gap and investigates the perception of Russian speech in multi-talker babble noise by the matched groups of high- and low-proficiency heritage speakers (HSs) and late second language learners of Russian who were native speakers of English. The study includes a control group of Russian native speakers. It manipulates the noise level (high and low), and context cloze probability (high and low). The results of the SPIN task are compared to the tasks testing the control of phonology, AXB discrimination and picture-word discrimination, and lexical knowledge, a word translation task, in the same participants. The increased phonological sensitivity of HSs interacted with their ability to rely on top–down processing in sentence integration, use contextual cues, and build expectancies in the high-noise/high-context condition in a bootstrapping fashion. HSs outperformed oral proficiency-matched late second language learners on SPIN task and two tests of phonological sensitivity. The outcomes of the SPIN experiment support both the early naturalistic advantage and the role of proficiency in HSs. HSs’ ability to take advantage of the high-predictability context in the high-noise condition was mitigated by their level of proficiency. Only high-proficiency HSs, but not any other non-native group, took advantage of the high-predictability context that became available with better phonological processing skills in high-noise. The study thus confirms high-proficiency (but not low-proficiency) HSs’ nativelike ability to combine bottom–up and top–down cues in processing SPIN.

Intelligibility of emotional speech in younger and older adults.

Little is known about the influence of vocal emotions on speech understanding. Word recognition accuracy for stimuli spoken to portray seven emotions (anger, disgust, fear, sadness, neutral, happiness, and pleasant surprise) was tested in younger and older listeners. Emotions were presented in either mixed (heterogeneous emotions mixed in a list) or blocked (homogeneous emotion blocked in a list) conditions. Three main hypotheses were tested. First, vocal emotion affects word recognition accuracy; specifically, portrayals of fear enhance word recognition accuracy because listeners orient to threatening information and/or distinctive acoustical cues such as high pitch mean and variation. Second, older listeners recognize words less accurately than younger listeners, but the effects of different emotions on intelligibility are similar across age groups. Third, blocking emotions in list results in better word recognition accuracy, especially for older listeners, and reduces the effect of emotion on intelligibility because as listeners develop expectations about vocal emotion, the allocation of processing resources can shift from emotional to lexical processing. Emotion was the within-subjects variable: all participants heard speech stimuli consisting of a carrier phrase followed by a target word spoken by either a younger or an older talker, with an equal number of stimuli portraying each of seven vocal emotions. The speech was presented in multi-talker babble at signal to noise ratios adjusted for each talker and each listener age group. Listener age (younger, older), condition (mixed, blocked), and talker (younger, older) were the main between-subjects variables. Fifty-six students (Mage= 18.3 years) were recruited from an undergraduate psychology course; 56 older adults (Mage= 72.3 years) were recruited from a volunteer pool. All participants had clinically normal pure-tone audiometric thresholds at frequencies ≤3000 Hz. There were significant main effects of emotion, listener age group, and condition on the accuracy of word recognition in noise. Stimuli spoken in a fearful voice were the most intelligible, while those spoken in a sad voice were the least intelligible. Overall, word recognition accuracy was poorer for older than younger adults, but there was no main effect of talker, and the pattern of the effects of different emotions on intelligibility did not differ significantly across age groups. Acoustical analyses helped elucidate the effect of emotion and some intertalker differences. Finally, all participants performed better when emotions were blocked. For both groups, performance improved over repeated presentations of each emotion in both blocked and mixed conditions. These results are the first to demonstrate a relationship between vocal emotion and word recognition accuracy in noise for younger and older listeners. In particular, the enhancement of intelligibility by emotion is greatest for words spoken to portray fear and presented heterogeneously with other emotions. Fear may have a specialized role in orienting attention to words heard in noise. This finding may be an auditory counterpart to the enhanced detection of threat information in visual displays. The effect of vocal emotion on word recognition accuracy is preserved in older listeners with good audiograms and both age groups benefit from blocking and the repetition of emotions.

The effect of increased vocal intensity on interarticulator timing in speakers with Parkinson's disease: A preliminary analysis

PurposeThe purpose of the current study was to investigate the effect of increased vocal intensity on interarticulator timing in individuals with Parkinson's disease (PD). MethodsTen individuals with mild to moderate hypophonia, secondary to PD, were selected for study. Over an 8-week treatment period, multi-talker babble noise was presented monaurally to the individuals with PD during everyday communication contexts to elicit increased vocal intensity (Lombard effect). Outcome measures included sound pressure level (SPL), voice onset time (VOT), VOT ratio, percent voicing, and speech intelligibility. ResultsGroup and individual participant responses to the treatment are reported and discussed. Speakers with PD were shown to significantly increase SPL in response to treatment. Six of the 10 speakers showed improved temporal coordination between the laryngeal and supralaryngeal mechanisms (interarticulator timing) in response to treatment. Four of the 10 speakers, however, showed reduced laryngeal–supralaryngeal timing at the end of treatment. Group speech intelligibility scores were significantly higher post-treatment as compared to pre-treatment. ConclusionsVoice treatment during everyday communication resulted in improved temporal coordination across the laryngeal and supralaryngeal mechanisms for the majority of speakers with PD and made them easier to understand. Further investigations are planned to explore individual differences in response to treatment. The identification of speaker-specific voicing and devoicing strategies is consistent with the heterogeneous nature of PD.Learning outcomes: Readers will be able to: 1. Describe the speech and voice characteristics of individuals with Parkinson's disease. 2. Define the Lombard effect. 3. Describe acoustic measures of voice onset time and percent voicing. 4. Describe the effect of voice treatment on voice onset time and percent voicing in individuals with Parkinson's disease.

Non-native listeners' recognition of high-variability speech using PRESTO.

Natural variability in speech is a significant challenge to robust successful spoken word recognition. In everyday listening environments, listeners must quickly adapt and adjust to multiple sources of variability in both the signal and listening environments. High-variability speech may be particularly difficult to understand for non-native listeners, who have less experience with the second language (L2) phonological system and less detailed knowledge of sociolinguistic variation of the L2. The purpose of this study was to investigate the effects of high-variability sentences on non-native speech recognition and to explore the underlying sources of individual differences in speech recognition abilities of non-native listeners. Participants completed two sentence recognition tasks involving high-variability and low-variability sentences. They also completed a battery of behavioral tasks and self-report questionnaires designed to assess their indexical processing skills, vocabulary knowledge, and several core neurocognitive abilities. Native speakers of Mandarin (n = 25) living in the United States recruited from the Indiana University community participated in the current study. A native comparison group consisted of scores obtained from native speakers of English (n = 21) in the Indiana University community taken from an earlier study. Speech recognition in high-variability listening conditions was assessed with a sentence recognition task using sentences from PRESTO (Perceptually Robust English Sentence Test Open-Set) mixed in 6-talker multitalker babble. Speech recognition in low-variability listening conditions was assessed using sentences from HINT (Hearing In Noise Test) mixed in 6-talker multitalker babble. Indexical processing skills were measured using a talker discrimination task, a gender discrimination task, and a forced-choice regional dialect categorization task. Vocabulary knowledge was assessed with the WordFam word familiarity test, and executive functioning was assessed with the BRIEF-A (Behavioral Rating Inventory of Executive Function - Adult Version) self-report questionnaire. Scores from the non-native listeners on behavioral tasks and self-report questionnaires were compared with scores obtained from native listeners tested in a previous study and were examined for individual differences. Non-native keyword recognition scores were significantly lower on PRESTO sentences than on HINT sentences. Non-native listeners' keyword recognition scores were also lower than native listeners' scores on both sentence recognition tasks. Differences in performance on the sentence recognition tasks between non-native and native listeners were larger on PRESTO than on HINT, although group differences varied by signal-to-noise ratio. The non-native and native groups also differed in the ability to categorize talkers by region of origin and in vocabulary knowledge. Individual non-native word recognition accuracy on PRESTO sentences in multitalker babble at more favorable signal-to-noise ratios was found to be related to several BRIEF-A subscales and composite scores. However, non-native performance on PRESTO was not related to regional dialect categorization, talker and gender discrimination, or vocabulary knowledge. High-variability sentences in multitalker babble were particularly challenging for non-native listeners. Difficulty under high-variability testing conditions was related to lack of experience with the L2, especially L2 sociolinguistic information, compared with native listeners. Individual differences among the non-native listeners were related to weaknesses in core neurocognitive abilities affecting behavioral control in everyday life.

English vowel identification in quiet and noise: effects of listeners' native language background.

Purpose: To investigate the effect of listener's native language (L1) and the types of noise on English vowel identification in noise.Method: Identification of 12 English vowels was measured in quiet and in long-term speech-shaped noise and multi-talker babble (MTB) noise for English- (EN), Chinese- (CN) and Korean-native (KN) listeners at various signal-to-noise ratios (SNRs).Results: Compared to non-native listeners, EN listeners performed significantly better in quiet and in noise. Vowel identification in long-term speech-shaped noise and in MTB noise was similar between CN and KN listeners. This is different from our previous study in which KN listeners performed better than CN listeners in English sentence recognition in MTB noise.Discussion: Results from the current study suggest that depending on speech materials, the effect of non-native listeners' L1 on speech perception in noise may be different. That is, in the perception of speech materials with little linguistic cues like isolated vowels, the characteristics of non-native listener's native language may not play a significant role. On the other hand, in the perception of running speech in which listeners need to use more linguistic cues (e.g., acoustic-phonetic, semantic, and prosodic cues), the non-native listener's native language background might result in a different masking effect.

Speaker's voice as a memory cue

Speaker's voice occupies a central role as the cornerstone of auditory social interaction. Here, we review the evidence suggesting that speaker's voice constitutes an integral context cue in auditory memory. Investigation into the nature of voice representation as a memory cue is essential to understanding auditory memory and the neural correlates which underlie it. Evidence from behavioral and electrophysiological studies suggest that while specific voice reinstatement (i.e., same speaker) often appears to facilitate word memory even without attention to voice at study, the presence of a partial benefit of similar voices between study and test is less clear. In terms of explicit memory experiments utilizing unfamiliar voices, encoding methods appear to play a pivotal role. Voice congruency effects have been found when voice is specifically attended at study (i.e., when relatively shallow, perceptual encoding takes place). These behavioral findings coincide with neural indices of memory performance such as the parietal old/new recollection effect and the late right frontal effect. The former distinguishes between correctly identified old words and correctly identified new words, and reflects voice congruency only when voice is attended at study. Characterization of the latter likely depends upon voice memory, rather than word memory. There is also evidence to suggest that voice effects can be found in implicit memory paradigms. However, the presence of voice effects appears to depend greatly on the task employed. Using a word identification task, perceptual similarity between study and test conditions is, like for explicit memory tests, crucial. In addition, the type of noise employed appears to have a differential effect. While voice effects have been observed when white noise is used at both study and test, using multi-talker babble does not confer the same results. In terms of neuroimaging research modulations, characterization of an implicit memory effect reflective of voice congruency is currently lacking.

Ideal time-frequency masking algorithms lead to different speech intelligibility and quality in normal-hearing and cochlear implant listeners.

Hearing impaired listeners using cochlear implants (CIs) suffer from a decrease in speech intelligibility (SI) in adverse listening conditions. Time-frequency masks are often applied to perform noise suppression in an attempt to increase SI. Two important masks are the so-called ideal binary mask (IBM) with its binary weights and the ideal Wiener filter (IWF) with its continuous weights. It is unclear which of the masks has the highest potential for SI and speech quality enhancement in CI users. In this study, both approaches for SI and quality enhancement were compared. The investigations were conducted in normal-hearing (NH) subjects listening to noise vocoder CI simulations and in CI users. The potential for SI improvement was assessed in a sentence recognition task with ideal mask estimates in multitalker babble and with an interfering talker. The robustness of the approaches was evaluated with simulated estimation errors. CI users assessed the speech quality in a preference rating. The IWF outperformed the IBM in NH listeners. In contrast, no significant difference was obtained in CI users. Estimation errors degraded SI in CI users for both approaches. In terms of quality, the IWF outperformed, slightly, the IBM processed signals. The outcomes of this study suggest that the mask pattern is not that crucial for CIs. Results of speech enhancement algorithms obtained with NH subjects listening to vocoded or normally processed stimuli do not translate to CI users. This outcome means that the effect of new strategies has to be quantified with the user group considered.

How difficult is difficult? Speech perception in noise in the elderly hearing impaired.

Perception of speech in the presence of competing multitalker noise is difficult for most individuals with sensory hearing loss, and in particular, for the elderly hearing impaired. Elderly people frequently report that these difficulties are poorly compensated for by hearing aids, albeit the algorithms and technologies aiming to improve speech perception in noise. The aim of the current study was therefore to assess competing speech signals processing by measuring the amount of signal to noise ratio (SNR) loss experienced by elderly hearing impaired individuals and their performance in dichotic listening tests. Speech in multitalker babble noise and dichotic listening were assessed in older hearing impaired individuals and in young normal-hearing adults. The average dichotic scores in the elderly group were substantially and significantly lower compared with the scores of the younger group with a significant right ear advantage (higher accuracy in reporting words arriving at the right ear than words arriving at the left ear) in the elderly group. Speech identification in noise in the older group was significantly poorer than in the younger group (SNR loss: 10 dB). The poor performance in the elderly hearing impaired group highlights the difficulties faced by hearing impaired older adults in demanding listening environments. Our results suggest that it may be unreasonable to expect that elderly hearing impaired individuals may fully understand speech in multitalker environments even when hearing aids are used.