Children's Speech Recognition Research Articles

Pitch-Normalized Acoustic Features for Robust Children's Speech Recognition

In this letter, the effectiveness of recently reported SMAC (Spectral Moment time–frequency distribution Augmented by low-order Cepstral) features has been evaluated for robust automatic speech recognition (ASR). The SMAC features consist of normalized first central spectral moments appended with low-order cepstral coefficients. These features have been designed for achieving robustness to both additive noise and the pitch variations. We have explored the SMAC features in severe pitch mismatch ASR task, i.e., decoding of children's speech on adults’ speech trained ASR system. In those tasks, the SMAC features are still observed to be sensitive to pitch variations. Toward addressing the same, a simple spectral smoothening approach employing adaptive-cepstral truncation is explored prior to the computation of spectral moments. With the proposed modification, the SMAC features are noted to achieve enhanced pitch robustness without affecting their noise immunity. Furthermore, the effectiveness of the proposed features is explored in three dominant acoustic modeling paradigms and varying data conditions. In all the cases, the proposed features are observed to significantly outperform the existing ones.

A Fast Adaptation Approach for Enhanced Automatic Recognition of Children’s Speech with Mismatched Acoustic Models

This study explores issues in automatic speech recognition (ASR) of children’s speech on acoustic models trained using adults’ speech. For acoustic modeling in ASR, the employed front-end features capture the characteristics of the vocal filter while smoothing out those of the source (excitation). Adults’ and children’s speech differ significantly due to large deviation in the acoustic correlates such as pitch, formants, speaking rate, etc. In the context of children’s speech recognition on mismatched acoustic models, the recognition rates remain highly degraded despite use of the vocal tract length normalization (VTLN) for addressing formant mismatch. For commonly used mel-filterbank-based cepstral features, earlier studies have shown that the acoustic mismatch is exacerbated by insufficient smoothing of pitch harmonics for child speakers. To address this problem, a structured low-rank projection of the test features as well as that of the mean and the covariance parameters of the acoustic models was explored in an earlier work. In this paper, a low-latency adaptation scheme is presented for children’s mismatched ASR. The presented fast adaptation approach exploits the earlier reported low-rank projection technique in order to reduce the computational cost. In the proposed approach, developmental data from the children’s domain is partitioned into separate groups on the basis of their estimated VTLN warp factors. A set of adapted acoustic models is then created by combining the low-rank projection with the model space adaptation technique for each of the warp factors. Given the children’s test utterance, first an appropriate pre-adapted model mean supervector is chosen based on its estimated warp factor. The chosen supervector is then optimally scaled. Consequently, only two parameters are required to be estimated, i.e., a warp factor and a model mean scaling factor. Even with such stringent constraints, the proposed adaptation technique results in a relative improvement of about $$44\%$$ over the VTLN included baseline.

Child-Robot Interactions for Second Language Tutoring to Preschool Children.

In this digital age social robots will increasingly be used for educational purposes, such as second language tutoring. In this perspective article, we propose a number of design features to develop a child-friendly social robot that can effectively support children in second language learning, and we discuss some technical challenges for developing these. The features we propose include choices to develop the robot such that it can act as a peer to motivate the child during second language learning and build trust at the same time, while still being more knowledgeable than the child and scaffolding that knowledge in adult-like manner. We also believe that the first impressions children have about robots are crucial for them to build trust and common ground, which would support child-robot interactions in the long term. We therefore propose a strategy to introduce the robot in a safe way to toddlers. Other features relate to the ability to adapt to individual children’s language proficiency, respond contingently, both temporally and semantically, establish joint attention, use meaningful gestures, provide effective feedback and monitor children’s learning progress. Technical challenges we observe include automatic speech recognition (ASR) for children, reliable object recognition to facilitate semantic contingency and establishing joint attention, and developing human-like gestures with a robot that does not have the same morphology humans have. We briefly discuss an experiment in which we investigate how children respond to different forms of feedback the robot can give.

Children's Speech Recognition Under Mismatched Condition: A Review

ABSTRACTAutomatic speech recognition (ASR) is a task of converting speech to text. In this article, the ASR system trained using the speech of adult speakers and tested using the speech of child speakers is termed as children's speech recognition under mismatched condition. The mismatch refers to differences in the training and testing conditions due to the differences in the characteristics of speech signals belonging to adults and children. This mismatch results in significant degradation of the recognition performance. Therefore, there are several approaches in the literature to reduce the mismatch and hence improve the performance. The approaches may be broadly viewed of those trying to reduce the mismatch due to vocal tract length and excitation source characteristics. This article will review the works that have been done in this direction. A discussion on the literature critically comments on the existing works and suggests possible directions for future research.

Evaluation on health-related quality of life in deaf children with cochlear implant in China

ObjectivePrevious studies have shown that deaf children benefit considerably from cochlear implants. These improvements are found in areas such as speech perception, speech production, and audiology-verbal performance. Despite the increasing prevalence of cochlear implants in China, few studies have reported on health-related quality of life in children with cochlear implants. The main objective of this study was to explore health-related quality of life on children with cochlear implants in South-west China. Study designA retrospective observational study of 213 CI users in Southwest China between 2010 and 2013. MethodsParticipants were 213 individuals with bilateral severe-to-profound hearing loss who wore unilateral cochlear implants. The Nijmegen Cochlear Implant Questionnaire and Health Utility Index Mark III were used pre-implantation and 1 year post-implantation. Additionally, 1-year postoperative scores for Mandarin speech perception were compared with preoperative scores. ResultsHealth-related quality of life improved post-operation with scores on the Nijmegen Cochlear Implant Questionnaire improving significantly in all subdomains, and the Health Utility Index 3 showing a significant improvement in the utility score and the subdomains of ‘‘hearing,” ‘‘speech,” and “emotion”. Additionally, a significant improvement in speech recognition scores was found. No significant correlation was found between increased in quality of life and speech perception scores. ConclusionHealth-related quality of life and speech recognition in prelingual deaf children significantly improved post-operation. The lack of correlation between quality of life and speech perception suggests that when evaluating performance post-implantation in prelingual deaf children and adolescents, measures of both speech perception and quality of life should be used.

The Effects of Acoustic Bandwidth on Simulated Bimodal Benefit in Children and Adults with Normal Hearing.

The primary purpose of this study was to examine the effect of acoustic bandwidth on bimodal benefit for speech recognition in normal-hearing children with a cochlear implant (CI) simulation in one ear and low-pass filtered stimuli in the contralateral ear. The effect of acoustic bandwidth on bimodal benefit in children was compared with the pattern of adults with normal hearing. Our hypothesis was that children would require a wider acoustic bandwidth than adults to (1) derive bimodal benefit, and (2) obtain asymptotic bimodal benefit. Nineteen children (6 to 12 years) and 10 adults with normal hearing participated in the study. Speech recognition was assessed via recorded sentences presented in a 20-talker babble. The AzBio female-talker sentences were used for the adults and the pediatric AzBio sentences (BabyBio) were used for the children. A CI simulation was presented to the right ear and low-pass filtered stimuli were presented to the left ear with the following cutoff frequencies: 250, 500, 750, 1000, and 1500 Hz. The primary findings were (1) adults achieved higher performance than children when presented with only low-pass filtered acoustic stimuli, (2) adults and children performed similarly in all the simulated CI and bimodal conditions, (3) children gained significant bimodal benefit with the addition of low-pass filtered speech at 250 Hz, and (4) unlike previous studies completed with adult bimodal patients, adults and children with normal hearing gained additional significant bimodal benefit with cutoff frequencies up to 1500 Hz with most of the additional benefit gained with energy below 750 Hz. Acoustic bandwidth effects on simulated bimodal benefit were similar in children and adults with normal hearing. Should the current results generalize to children with CIs, these results suggest pediatric CI recipients may derive significant benefit from minimal acoustic hearing (<250 Hz) in the nonimplanted ear and increasing benefit with broader bandwidth. Knowledge of the effect of acoustic bandwidth on bimodal benefit in children may help direct clinical decisions regarding a second CI, continued bimodal hearing, and even optimizing acoustic amplification for the nonimplanted ear.

Effects of interleaved noise on speech recognition in children

Normal-hearing adults can isolate frequency regions containing clean speech from surrounding regions containing noise. However, children have been shown to integrate information over a large number of auditory filters, and so they may not be able to isolate frequency regions as well. To assess children’s level of auditory filter independence, words were filtered into 30 contiguous 1-ERB-width bands. Speech was presented in every other band, for a total of 15 speech bands. Speech-shaped noise was then added to: all 30 contiguous bands, the 15 bands not containing speech (OFF), or the 15 bands containing speech (ON). Three age groups were tested: 10 adults, 9 older children (6–7 yr old), and 9 younger children (5 yr old). Consistent with previous findings involving consonant recognition, adults displayed large performance differences between off- and on-frequency noise (OFF vs. ON). The 6- to 7-yr-old group performed similarly to adults. In contrast, the 5-yr-old group displayed equivalent performance in the OFF and ON conditions. This indicates that, for these younger children, even noise that was mostly non-overlapping in frequency interfered with speech recognition as much as noise that was on frequency. [Work supported by NIH.]

Speech Recognition and Parent Ratings From Auditory Development Questionnaires in Children Who Are Hard of Hearing.

Progress has been made in recent years in the provision of amplification and early intervention for children who are hard of hearing. However, children who use hearing aids (HAs) may have inconsistent access to their auditory environment due to limitations in speech audibility through their HAs or limited HA use. The effects of variability in children's auditory experience on parent-reported auditory skills questionnaires and on speech recognition in quiet and in noise were examined for a large group of children who were followed as part of the Outcomes of Children with Hearing Loss study. Parent ratings on auditory development questionnaires and children's speech recognition were assessed for 306 children who are hard of hearing. Children ranged in age from 12 months to 9 years. Three questionnaires involving parent ratings of auditory skill development and behavior were used, including the LittlEARS Auditory Questionnaire, Parents Evaluation of Oral/Aural Performance in Children rating scale, and an adaptation of the Speech, Spatial, and Qualities of Hearing scale. Speech recognition in quiet was assessed using the Open- and Closed-Set Test, Early Speech Perception test, Lexical Neighborhood Test, and Phonetically Balanced Kindergarten word lists. Speech recognition in noise was assessed using the Computer-Assisted Speech Perception Assessment. Children who are hard of hearing were compared with peers with normal hearing matched for age, maternal educational level, and nonverbal intelligence. The effects of aided audibility, HA use, and language ability on parent responses to auditory development questionnaires and on children's speech recognition were also examined. Children who are hard of hearing had poorer performance than peers with normal hearing on parent ratings of auditory skills and had poorer speech recognition. Significant individual variability among children who are hard of hearing was observed. Children with greater aided audibility through their HAs, more hours of HA use, and better language abilities generally had higher parent ratings of auditory skills and better speech-recognition abilities in quiet and in noise than peers with less audibility, more limited HA use, or poorer language abilities. In addition to the auditory and language factors that were predictive for speech recognition in quiet, phonological working memory was also a positive predictor for word recognition abilities in noise. Children who are hard of hearing continue to experience delays in auditory skill development and speech-recognition abilities compared with peers with normal hearing. However, significant improvements in these domains have occurred in comparison to similar data reported before the adoption of universal newborn hearing screening and early intervention programs for children who are hard of hearing. Increasing the audibility of speech has a direct positive effect on auditory skill development and speech-recognition abilities and also may enhance these skills by improving language abilities in children who are hard of hearing. Greater number of hours of HA use also had a significant positive impact on parent ratings of auditory skills and children's speech recognition.

Pitch adaptive MFCC features for improving children’s mismatched ASR

A pitch normalization algorithm is proposed for addressing the pitch mismatch between adults' and children's speech for children's automatic speech recognition (ASR). Motivated by the appearance of pitch-dependent distortions in the smoothed mel spectral envelope for high-pitched children's speech, the algorithm modifies the mel filterbank during MFCC feature extraction to improve ASR performance. Relative improvements of 16 % and 9 % are obtained over the corresponding baseline in children's mismatched ASR performance on a connected-digit recognition task and a continuous speech recognition task. The improvements obtained in ASR performance with the proposed pitch normalization algorithm are also found to be additive to that obtained with existing speaker normalization techniques, VTLN and CMLLR.

Evaluation of wideband frequency responses and non-linear frequency compression for children with mild to moderate high-frequency hearing loss

Objective: To evaluate wideband amplification and non-linear frequency compression (NLFC) as a means to improve speech recognition for children with mild/moderate hearing loss. Design: Randomized within-subject design with repeated measures across test conditions. Study sample: Eleven children with mild to moderate hearing loss were evaluated with: (1) Phonak BTE without NLFC, (2) Phonak BTE with NLFC, and (3) Oticon BTE with wideband response extending to 8000 Hz. Results: Use of NLFC provided better detection and recognition of high-frequency stimuli (e.g. /sh/ and /s/). No difference in performance between conditions was observed for speech recognition when measured with the University of Western Ontario (UWO) plurals test and the UWO distinctive features difference test. Finally, there were no differences between conditions on the BKB-SIN test. Conclusions: Children with mild to moderate hearing loss have good access to high-frequency phonemes presented at fixed levels (e.g. 50 to 60 dBA) with both wideband and NLFC technology. Similarly, sentence recognition in noise was similar with wideband and NLFC. Adaptive test procedures that probe performance at lower input levels showed small but significant improvements in the detection and recognition of the phonemes /s/ and /sh/ with NLFC condition when compared to the NLFC Off and wideband conditions.

Evaluation of the benefits of binaural hearing on the telephone for children with hearing loss.

There is a paucity of published studies examining how children with hearing loss understand speech over the telephone. Previous studies on adults with hearing aids have suggested that adults with bilateral hearing aids experience significant difficulty recognizing speech on the telephone when listening with one ear, but the provision of telephone input to both ears substantially improved speech understanding. The objectives of this study were to measure speech recognition in quiet and in noise for a group of older children with hearing loss over the telephone and to evaluate the effects of binaural hearing (e.g., DuoPhone) on speech recognition over the telephone. A cross-sectional, repeated-measures design was used in this study. A total of 14 children, ages 6-14 yr, participated in the study. Participants were obtained using convenience sampling from a nonprofit clinic population. Speech recognition in quiet and in noise with binaural versus monaural telephone input was compared in pediatric participants. Monosyllabic word recognition was assessed in quiet and classroom noise set at 50 dBA in conditions with monaural and binaural (DuoPhone) telephone input. The children's speech recognition in quiet and in noise was significantly better with binaural telephone input relative to monaural telephone input. To obtain optimal performance on the telephone, the following considerations may apply: (1) use of amplification with binaural streaming capabilities (e.g., DuoPhone), (2) counseling of family and children on how to best use the telephone, (3) provision of telecoil with microphone attenuation for improved signal-to-noise ratio, and (4) use of probe tube measures to verify the appropriateness of the telephone programs.

Evaluation of wideband frequency responses and nonlinear frequency compression for children with cookie-bite audiometric configurations.

Previous research has suggested that use of nonlinear frequency compression (NLFC) can improve audibility for high-frequency sounds and speech recognition of children with moderate to profound high-frequency hearing loss. Furthermore, previous studies have generally found no detriment associated with the use of NLFC. However, there have been no published studies examining the effect of NLFC on the performance of children with cookie-bite audiometric configurations. For this configuration of hearing loss, frequency-lowering processing will likely move high-frequency sounds to a lower frequency range at which a greater degree of hearing loss exists. The purpose of this study was to evaluate and compare the effects of wideband amplification and NLFC on high-frequency audibility and speech recognition of children with cookie-bite audiometric configurations. This study consisted of a within-participant design with repeated measures across test conditions. Seven children, ages 6-13 yr, with cookie-bite audiometric configurations and normal hearing or mild hearing loss at 6000 and 8000 Hz, were recruited. Participants were fitted with Phonak Nios S H2O III behind-the-ear hearing aids and Oticon Safari 300 behind-the-ear hearing aids. The participants were evaluated after three 4-to 6-wk intervals: (1) Phonak Nios S H2O III without NLFC, (2) Phonak Nios S H2O III with NLFC, and (3) Oticon Safari 300 with wideband frequency response extending to 8000 Hz. The order in which each technology was used was counterbalanced across participants. High-frequency audibility was evaluated by assessing aided thresholds (dB SPL) for warble tones and the high-frequency phonemes /sh/ and /s/. Speech recognition in quiet was measured with the University of Western Ontario (UWO) Plurals Test, the UWO Distinctive Features Difference (DFD) Test, and the Phoneme Perception Test vowel-consonant-vowel nonsense syllable test. Sentence recognition in noise was evaluated with the Bamford-Kowal-Bench Speech-In-Noise (BKB-SIN) Test. Repeated-measures analyses of variance were used to analyze the data collected in this study. The results across the three different conditions were compared. No difference in performance across conditions was observed for detection of high-frequency warble tones and the speech sounds /sh/ and /s/. No significant difference was seen across conditions for speech recognition in quiet when measured with the UWO Plurals Test, the UWO-DFD Test, and the Phoneme Perception Test vowel-consonant-vowel nonsense syllable test. Finally, there were also no differences across conditions on the BKB-SIN Test. These results suggest that NLFC does not degrade or improve audibility for and recognition of high-frequency speech sounds as well as sentence recognition in noise when compared with wideband amplification for children with cookie-bite audiometric configurations.

Fuzzy-based discriminative feature representation for children's speech recognition

Automatic recognition of the speech of children is a challenging topic in computer-based speech recognition systems. Conventional feature extraction method namely Mel-frequency cepstral coefficient (MFCC) is not efficient for children's speech recognition. This paper proposes a novel fuzzy-based discriminative feature representation to address the recognition of Malay vowels uttered by children. Considering the age-dependent variational acoustical speech parameters, performance of the automatic speech recognition (ASR) systems degrades in recognition of children's speech. To solve this problem, this study addresses representation of relevant and discriminative features for children's speech recognition. The addressed methods include extraction of MFCC with narrower filter bank followed by a fuzzy-based feature selection method. The proposed feature selection provides relevant, discriminative, and complementary features. For this purpose, conflicting objective functions for measuring the goodness of the features have to be fulfilled. To this end, fuzzy formulation of the problem and fuzzy aggregation of the objectives are used to address uncertainties involved with the problem.The proposed method can diminish the dimensionality without compromising the speech recognition rate. To assess the capability of the proposed method, the study analyzed six Malay vowels from the recording of 360 children, ages 7 to 12. Upon extracting the features, two well-known classification methods, namely, MLP and HMM, were employed for the speech recognition task. Optimal parameter adjustment was performed for each classifier to adapt them for the experiments. The experiments were conducted based on a speaker-independent manner. The proposed method performed better than the conventional MFCC and a number of conventional feature selection methods in the children speech recognition task. The fuzzy-based feature selection allowed the flexible selection of the MFCCs with the best discriminative ability to enhance the difference between the vowel classes.

Evidence indicates that cochlear implantation in children with auditory neuropathy spectrum disorder results in favorable outcomes; however, the quality of evidence is weak

This review provides a summary and appraisal commentary on the treatment review by Humphriss, R., Hall, A., Maddocks, J., Macleod, J., Sawaya, K., & Midgley, E. (2013). Does cochlear implantation improve speech recognition in children with auditory neuropathy spectrum disorder? A systematic review. International Journal of Audiology, 52, 442–454.Source of funding and declaration of interests: Authors (RH, JM, and EM) declared association within a cochlear implant program in the United Kingdom.

Masked speech recognition in school-age children and adults: Effects of age, masker type, and context

This study evaluated the influences of age, masker type, and context on masked speech recognition. A repeated-measures design compared the speech recognition thresholds of two age groups of children (5–7 and 9–13 years) and a group of adults (19–33 years) in a continuous speech-shaped noise or a two-talker speech masker. Target stimuli were disyllabic words that were familiar to young children. Masker level was fixed at 60 dB SPL, and signal level was adapted to estimate the SNR required for 70.7% correct performance. Each listener completed testing in each masker condition using both an open-set task requiring a verbal response, and a 4AFC closed-set task requiring a picture-pointing response. Consistent with previous studies, and regardless of response context, larger and more prolonged child-adult differences were observed in the two-talker compared to the speech-shaped noise masker. Poorer performance was observed for all three age groups using the open-set compared to the closed-set context. This performance gap was similar across the three age groups in the speech-shaped noise masker, but a developmental effect was observed in the two-talker masker. Specifically, the decrement in performance using the open-set compared to the closed-set procedure increased with age in the two-talker masker.

The Effects of Digital Signal Processing Features on Children's Speech Recognition and Loudness Perception

The purpose of this study was to determine the effects of hearing instruments set to Desired Sensation Level version 5 (DSL v5) hearing instrument prescription algorithm targets and equipped with directional microphones and digital noise reduction (DNR) on children's sentence recognition in noise performance and loudness perception in a classroom environment. Ten children (ages 8-17 years) with stable, congenital sensorineural hearing losses participated in the study. Participants were fitted bilaterally with behind-the-ear hearing instruments set to DSL v5 prescriptive targets. Sentence recognition in noise was evaluated using the Bamford-Kowal-Bench Speech in Noise Test (Niquette et al., 2003). Loudness perception was evaluated using a modified version of the Contour Test of Loudness Perception (Cox, Alexander, Taylor, & Gray, 1997). Children's sentence recognition in noise performance was significantly better when using directional microphones alone or in combination with DNR than when using omnidirectional microphones alone or in combination with DNR. Children's loudness ratings for sounds above 72 dB SPL were lowest when fitted with the DSL v5 Noise prescription combined with directional microphones. DNR use showed no effect on loudness ratings. Use of the DSL v5 Noise prescription with a directional microphone improved sentence recognition in noise performance and reduced loudness perception ratings for loud sounds relative to a typical clinical reference fitting with the DSL v5 Quiet prescription with no digital signal processing features enabled. Potential clinical strategies are discussed.

Does cochlear implantation improve speech recognition in children with auditory neuropathy spectrum disorder? A systematic review

Objective: Cochlear implantation (CI) is a standard treatment for severe-profound sensorineural hearing loss (SNHL). However, consensus has yet to be reached on its effectiveness for hearing loss caused by auditory neuropathy spectrum disorder (ANSD). This review aims to summarize and synthesize current evidence of the effectiveness of CI in improving speech recognition in children with ANSD. Design: Systematic review. Study sample: A total of 27 studies from an initial selection of 237. Results: All selected studies were observational in design, including case studies, cohort studies, and comparisons between children with ANSD and SNHL. Most children with ANSD achieved open-set speech recognition with their CI. Speech recognition ability was found to be equivalent in CI users (who previously performed poorly with hearing aids) and hearing-aid users. Outcomes following CI generally appeared similar in children with ANSD and SNHL. Assessment of study quality, however, suggested substantial methodological concerns, particularly in relation to issues of bias and confounding, limiting the robustness of any conclusions around effectiveness. Conclusions: Currently available evidence is compatible with favourable outcomes from CI in children with ANSD. However, this evidence is weak. Stronger evidence is needed to support cost-effective clinical policy and practice in this area.

한국어 유아 음성인식을 위한 수정된 Mel 주파수 캡스트럼

본 논문에서는 한국어에서 유아 대상의 음성인식 향상을 위한 새로운 특징추출 알고리즘을 제안한다. 제안하는 특징추출 알고리즘은 세 가지 방법을 통합한 기법이다. 첫째 성도의 길이가 성인에 비해 짧은 유아의 음향적 특징을 보완하기 위한 방법으로 성도정규화 방법을 사용한다. 둘째 성인의 음성과 비교했을 때 높은 스펙트럼 영역에 집중되어 있는 유아의 음향적 특징을 보완하기 위해 균일한 대역폭을 사용하는 방법이다. 마지막으로 실시간 환경에서의 잡음에 강건한 음성인식기 개발을 위해 스무딩 필터를 사용하여 보완하는 방법이다. 세 가지 방법을 통해 제안하는 특징추출 기법은 실험을 통해 유아의 음성인식 성능 향상에 도움을 준다는 것을 확인했다. This paper proposes a new feature extraction algorithm to improve children's speech recognition in Korean. The proposed feature extraction algorithm combines three methods. The first method is on the vocal tract length normalization to compensate acoustic features because the vocal tract length in children is shorter than in adults. The second method is to use the uniform bandwidth because children's voice is centered on high spectral regions. Finally, the proposed algorithm uses a smoothing filter for a robust speech recognizer in real environments. This paper shows the new feature extraction algorithm improves the children's speech recognition performance.

Children weight dynamic spectral structure more than adults: Evidence from equivalent signals

Earlier work using sine-wave and noise-vocoded signals suggests that dynamic spectral structure plays a greater role in speech recognition for children than adults [Nittrouer and Lowenstein. (2010). J. Acoust. Soc. Am. 127, 1624-1635], but questions arise concerning whether outcomes can be compared because sine waves and wide noise bands are different in nature. The current study addressed that question using narrow noise bands for both signals, and applying a difference ratio to index the contribution made by dynamic spectral structure. Results replicated earlier findings, supporting the idea that dynamic spectral structure plays a critical role in speech recognition, especially for children.

Effects of Reverberation on Speech Recognition in Stationary and Modulated Noise by School-Aged Children and Young Adults

The purpose of this study was to determine how combinations of reverberation and noise, typical of environments in many elementary school classrooms, affect normal-hearing school-aged children's speech recognition in stationary and amplitude-modulated noise, and to compare their performance with that of normal-hearing young adults. In addition, the magnitude of release from masking in the modulated noise relative to that in stationary noise was compared across age groups in nonreverberant and reverberant listening conditions. Last, for all noise and reverberation combinations the degree of change in predicted performance at 70% correct was obtained for all age groups using a best-fit cubic polynomial. Bamford-Kowal-Bench sentences and noise were convolved with binaural room impulse responses representing nonreverberant and reverberant environments to create test materials representative of both audiology clinics and school classroom environments. Speech recognition of 48 school-aged children and 12 adults was measured in speech-shaped and amplitude-modulated speech-shaped noise, in the following three virtual listening environments: nonreverberant, reverberant at 2 m, and reverberant at 6 m. Speech recognition decreased in the reverberant conditions and with decreasing age. Release from masking in modulated noise relative to stationary noise decreased with age and was reduced by reverberation. In the nonreverberant condition, participants showed similar amounts of masking release across ages. The slopes of performance-intensity functions increased with age, with the exception of the nonreverberant modulated masker condition. The slopes were steeper in the stationary masker conditions, where they also decreased with reverberation and distance. In the presence of a modulated masker, the slopes did not differ between the two reverberant conditions. The results of this study reveal systematic developmental changes in speech recognition in noisy and reverberant environments for elementary-school-aged children. The overall pattern suggests that younger children require better acoustic conditions to achieve sentence recognition equivalent to their older peers and adults. In addition, this is the first study to report a reduction of masking release in children as a result of reverberation. Results support the importance of minimizing noise and reverberation in classrooms, and highlight the need to incorporate noise and reverberation into audiological speech-recognition testing to improve predictions of performance in the real world.