Classification of indexical and segmental features of human speech using low- and high-frequency energya).

The high-frequency region (above 4-5 kHz) of the speech spectrum has received substantial research attention over the previous decade, with a host of studies documenting the presence of important and useful information in this region. The purpose of the current experiment was to compare the presence of indexical and segmental information in the low- and high-frequency region of speech (below and above 4 kHz) and to determine the extent to which information from these regions can be used in a machine learning framework to correctly classify indexical and segmental aspects of the speech signal. Naturally produced vowel segments produced by ten male and ten female talkers were used as input to a temporal dictionary ensemble classification model in unfiltered, low-pass filtered (below 4 kHz), and high-pass filtered (above 4 kHz) conditions. Classification performance in the unfiltered and low-pass filtered conditions was approximately 90% or better for vowel categorization, talker sex, and individual talker identity tasks. Classification performance for high-pass filtered signals composed of energy above 4 kHz was well above chance for the same tasks. For several classification tasks (i.e., talker sex and talker identity), high-pass filtering had minimal effect on classification performance, suggesting the preservation of indexical information above 4 kHz.

Electrophysiological Evidence of Early Cortical Sensitivity to Human Conspecific Mimic Voice as a Distinct Category of Natural Sound.

Purpose From an anthropological perspective of hominin communication, the human auditory system likely evolved to enable special sensitivity to sounds produced by the vocal tracts of human conspecifics whether attended or passively heard. While numerous electrophysiological studies have used stereotypical human-produced verbal (speech voice and singing voice) and nonverbal vocalizations to identify human voice-sensitive responses, controversy remains as to when (and where) processing of acoustic signal attributes characteristic of "human voiceness" per se initiate in the brain. Method To explore this, we used animal vocalizations and human-mimicked versions of those calls ("mimic voice") to examine late auditory evoked potential responses in humans. Results Here, we revealed an N1b component (96-120 ms poststimulus) during a nonattending listening condition showing significantly greater magnitude in response to mimics, beginning as early as primary auditory cortices, preceding the time window reported in previous studies that revealed species-specific vocalization processing initiating in the range of 147-219 ms. During a sound discrimination task, a P600 (500-700 ms poststimulus) component showed specificity for accurate discrimination of human mimic voice. Distinct acoustic signal attributes and features of the stimuli were used in a classifier model, which could distinguish most human from animal voice comparably to behavioral data-though none of these single features could adequately distinguish human voiceness. Conclusions These results provide novel ideas for algorithms used in neuromimetic hearing aids, as well as direct electrophysiological support for a neurocognitive model of natural sound processing that informs both neurodevelopmental and anthropological models regarding the establishment of auditory communication systems in humans. Supplemental Material https://doi.org/10.23641/asha.12903839.

Effects of signal bandwidth on listening effort in young- and middle-aged adults.

OBJECTIVE: The purpose of this study was to assess the effects of signal bandwidth, noise and signal predictability on listening effort among younger and middle-aged individuals with normal hearing, where YNH = young normal hearing and ONH = older normal hearing. It was of interest to determine if a reduction in high-frequency energy would differentially influence performance between the two groups. DESIGN: A mixed-model design was used, where listening effort, word identification and effort ratings served as the within-subject factors and listener group served as the between-subject factor. STUDY SAMPLE: Twenty listeners aged 18-25 years (YNH group) and 20 listeners aged 40-55 years (ONH group) participated in the experiment. RESULTS: Results showed significantly poorer word recall and significantly higher perceived effort among the ONH group. Increasing the signal bandwidth from 2 to 8 kHz significantly improved word recall and decreased listening effort ratings. Effort ratings from both groups matched word recall performance in that when word recall was lower perceived effort was higher and when word recall was higher perceived effort was lower. CONCLUSIONS: In general, providing additional high-frequency energy up to 8 kHz reduced listening effort among both groups of listeners. However, the ONH group, experienced additional effort when completing the tasks.

Effects of signal bandwidth and noise on individual speaker identification.

Two experiments were conducted to evaluate the effects of increasing spectral bandwidth from 3 to 10 kHz on individual speaker recognition in noisy conditions (+5, 0, and -5 dB signal-to-noise ratio). Experiment 1 utilized h(Vowel)d (hVd) signals, while experiment 2 utilized sentences from the Rainbow Passage. Both experiments showed significant improvements in individual speaker identification in the 10 kHz bandwidth condition (6% for hVds; 10% for sentences). These results coincide with the extant machine recognition literature demonstrating significant amounts of individual speaker information present in the speech signal above approximately 3-4 kHz. Cues from the high-frequency region for speaker identity warrant further study.

Hearing and orally mimicking different acoustic-semantic categories of natural sound engage distinct left hemisphere cortical regions.

Oral mimicry is thought to represent an essential process for the neurodevelopment of spoken language systems in infants, the evolution of language in hominins, and a process that could possibly aid recovery in stroke patients. Using functional magnetic resonance imaging (fMRI), we previously reported a divergence of auditory cortical pathways mediating perception of specific categories of natural sounds. However, it remained unclear if or how this fundamental sensory organization by the brain might relate to motor output, such as sound mimicry. Here, using fMRI, we revealed a dissociation of activated brain regions preferential for hearing with the intent to imitate and the oral mimicry of animal action sounds versus animal vocalizations as distinct acoustic-semantic categories. This functional dissociation may reflect components of a rudimentary cortical architecture that links systems for processing acoustic-semantic universals of natural sound with motor-related systems mediating oral mimicry at a category level. The observation of different brain regions involved in different aspects of oral mimicry may inform targeted therapies for rehabilitation of functional abilities after stroke.

Gender Identification Using High-Frequency Speech Energy: Effects of Increasing the Low-Frequency Limit.

OBJECTIVE: The purpose of this study was to investigate the ability of normal-hearing listeners to use high-frequency energy for gender identification from naturally produced speech signals. DESIGN: Two experiments were conducted using a repeated-measures design. Experiment 1 investigated the effects of increasing high-pass filter cutoff (i.e., increasing the low-frequency spectral limit) on gender identification from naturally produced vowel segments. Experiment 2 studied the effects of increasing high-pass filter cutoff on gender identification from naturally produced sentences. Confidence ratings for the gender identification task were also obtained for both experiments. RESULTS: Listeners in experiment 1 were capable of extracting talker gender information at levels significantly above chance from vowel segments high-pass filtered up to 8.5 kHz. Listeners in experiment 2 also performed above chance on the gender identification task from sentences high-pass filtered up to 12 kHz. CONCLUSIONS: Cumulatively, the results of both experiments provide evidence that normal-hearing listeners can utilize information from the very high-frequency region (above 4 to 5 kHz) of the speech signal for talker gender identification. These findings are at variance with current assumptions regarding the perceptual information regarding talker gender within this frequency region. The current results also corroborate and extend previous studies of the use of high-frequency speech energy for perceptual tasks. These findings have potential implications for the study of information contained within the high-frequency region of the speech spectrum and the role this region may play in navigating the auditory scene, particularly when the low-frequency portion of the spectrum is masked by environmental noise sources or for listeners with substantial hearing loss in the low-frequency region and better hearing sensitivity in the high-frequency region (i.e., reverse slope hearing loss).

Automated Classification of Vowel Category and Speaker Type in the High-Frequency Spectrum.

The high-frequency region of vowel signals (above the third formant or F3) has received little research attention. Recent evidence, however, has documented the perceptual utility of high-frequency information in the speech signal above the traditional frequency bandwidth known to contain important cues for speech and speaker recognition. The purpose of this study was to determine if high-pass filtered vowels could be separated by vowel category and speaker type in a supervised learning framework. Mel frequency cepstral coefficients (MFCCs) were extracted from productions of six vowel categories produced by two male, two female, and two child speakers. Results revealed that the filtered vowels were well separated by vowel category and speaker type using MFCCs from the high-frequency spectrum. This demonstrates the presence of useful information for automated classification from the high-frequency region and is the first study to report findings of this nature in a supervised learning framework.

Gaps-in-noise detection and gender identification from noise-vocoded vowel segments: Comparing performance of active musicians to non-musicians.

This study evaluated performance on a gender identification and temporal resolution task among active musicians and age-matched non-musicians. Brief duration (i.e., 50 and 100 ms) vowel segments produced by four adult male and four adult female speakers were spectro-temporally degraded using various parameters and presented to both groups for gender identification. Gap detection thresholds were measured using the gaps-in-noise (GIN) test. Contrary to the stated hypothesis, a significant difference in gender identification was not observed between the musician and non-musician listeners. A significant difference, however, was observed on the temporal resolution task, with the musician group achieving approximately 2 ms shorter gap detection thresholds on the GIN test compared to the non-musician counterparts. These results provide evidence supporting the potential benefits of musical training on temporal processing abilities, which have implications for the processing of speech in degraded listening environments and the enhanced processing of the fine-grained temporal aspects of the speech signal. The results also support the GIN test as an instrument sensitive to temporal processing differences among active musicians and non-musicians.

Gender identification from high-pass filtered vowel segments: the use of high-frequency energy.

The purpose of this study was to examine the use of high-frequency information for making gender identity judgments from high-pass filtered vowel segments produced by adult speakers. Specifically, the effect of removing lower-frequency spectral detail (i.e., F3 and below) from vowel segments via high-pass filtering was evaluated. Thirty listeners (ages 18-35) with normal hearing participated in the experiment. A within-subjects design was used to measure gender identification for six 250-ms vowel segments (/æ/, /ɪ /, /ɝ/, /ʌ/, /ɔ/, and /u/), produced by ten male and ten female speakers. The results of this experiment demonstrated that despite the removal of low-frequency spectral detail, the listeners were accurate in identifying speaker gender from the vowel segments, and did so with performance significantly above chance. The removal of low-frequency spectral detail reduced gender identification by approximately 16 % relative to unfiltered vowel segments. Classification results using linear discriminant function analyses followed the perceptual data, using spectral and temporal representations derived from the high-pass filtered segments. Cumulatively, these findings indicate that normal-hearing listeners are able to make accurate perceptual judgments regarding speaker gender from vowel segments with low-frequency spectral detail removed via high-pass filtering. Therefore, it is reasonable to suggest the presence of perceptual cues related to gender identity in the high-frequency region of naturally produced vowel signals. Implications of these findings and possible mechanisms for performing the gender identification task from high-pass filtered stimuli are discussed.

Identification of high-pass filtered male, female, and child vowels: The use of high-frequency cues.

Vowels are characteristically described according to low-frequency resonance characteristics, which are presumed to provide the requisite information for identification. Classically, the study of vowel perception has focused on the lowest formant frequencies, typically F1, F2, and F3. Lehiste and Peterson [Phonetica 4, 161-177 (1959)] investigated identification accuracy of naturally produced male vowels composed of various amounts of low- and high-frequency content. Results showed near-chance identification performance for vowel segments containing only spectral information above 3.5 kHz. The authors concluded that high-frequency information was of minor importance for vowel identification. The current experiments report identification accuracy for high-pass filtered vowels produced by two male, two female, and two child talkers using both between- and within-subject designs. Identification performance was found to be significantly above chance for the majority of vowels even after high-pass filtering to remove spectral content below 3.0-3.5 kHz. Additionally, the filtered vowels having the highest fundamental frequency (child talkers) often had the highest identification accuracy scores. Linear discriminant function analysis mirrored perceptual performance when using spectral peak information between 3 and 12 kHz.

The awareness of doctoral-level professions among entering college students.

PURPOSE: The purpose of this study was to survey entering college students' awareness of the profession of audiology, as well as to compare the students' awareness of audiology to their awareness of roles in 2 other doctoral-level professions. METHOD: A survey examining students' awareness and perception of the profession of audiology was administered in 2009 at the California University of Pennsylvania (Emanuel, Donai, & Araj, 2012). A modified survey, which included questions about awareness of the profession of audiology as well as podiatry and optometry, was distributed in 2012. RESULTS: Survey data suggest that, overall, students were more accurate in describing the profession of optometry compared to audiology, but no difference existed between their awareness of podiatry and audiology. However, students with self-reported awareness of the three professions were more accurate in describing the professions of podiatry and optometry as compared to audiology. CONCLUSIONS: Results of this distribution suggest an upward trend in audiology awareness among students. However, the awareness of audiology as a potential career path is low relative to that of the professions of optometry and podiatry for those with self-reported knowledge. Future surveys should be distributed to entering college students at other universities, varying in size, location, and demographics.

The awareness of the profession of audiology among entering college students.

PURPOSE: The purpose of this study was to develop and pilot test a survey of entering college students' awareness of the profession of audiology and to report the preliminary results of the survey. METHOD: A survey was developed and administered in 2003 at the California University of Pennsylvania. A modified survey was administered in 2009. Survey questions asked students to identify what audiologists do and the education requirements to become an audiologist. RESULTS: Seventeen percent of the students self-reported that they knew what an audiologist did and were able to accurately describe the profession. Approximately 30% of the students learned about audiology from family/friends. Students reported selecting their major based on interest in a specific field and not on market-driven forces such as job opportunities and salary. CONCLUSION: Future surveys should be conducted to confirm the extent of the lack of visibility of audiology as a profession and to serve as a metric for the efficacy of future marketing efforts in the profession.