Listener Ratings Research Articles

Articulatory-acoustic vowel space: Associations between acoustic and perceptual measures of clear speech

Purpose: The current investigation examined the relationship between perceptual ratings of speech clarity and acoustic measures of speech production. Included among the acoustic measures was the Articulatory-Acoustic Vowel Space (AAVS), which provides a measure of working formant space derived from continuously sampled formant trajectories in connected speech.Method: Acoustic measures of articulation and listener ratings of speech clarity were obtained from habitual and clear speech samples produced by 10 neurologically healthy adults. Perceptual ratings of speech clarity were obtained from visual-analogue scale ratings and acoustic measures included the AAVS measure, articulation rate and percentage pause.Result: Clear speech was characterised by a higher perceptual clarity rating, slower articulation rate, greater percentage pause and larger AAVS compared to habitual speech. Additionally, correlation analysis revealed a significant relationship between the perceptual clear speech effect and the relative clarity-related change in the AAVS and articulation rate measures.Conclusion: The current findings suggest that, along with speech rate measures, the recently introduced AAVS is sensitive to changes in speech clarity.

Listener ratings of singer expressivity in musical performance

The objective of this study was to analyze whether listeners’ ratings of a singer’s expressivity in musical examples in two contemporary commercial styles (pop and country) were affected by the presence of vocal fry and to see if there was a difference in listener ratings according to the singer’s gender. Subjects listened to musical samples with and without vocal fry that were recorded by one male and one female singer, and were asked to rate the singer’s expressivity on a numerical scale. Certain musical examples were presented twice, with and without vocal fry, in order to directly compare a performance of the same song in both modes. Some samples were repeated to test the reliability of the raters. Listener ratings indicated a preference for performances with vocal fry made by the female singer; for the male singer, performances without vocal fry were rated higher.

Genre-Adaptive Semantic Computing and Audio-Based Modelling for Music Mood Annotation

This study investigates whether taking genre into account is beneficial for automatic music mood annotation in terms of core affects valence, arousal, and tension, as well as several other mood scales. Novel techniques employing genre-adaptive semantic computing and audio-based modelling are proposed. A technique called the ACTwg employs genre-adaptive semantic computing of mood-related social tags, whereas ACTwg-SLPwg combines semantic computing and audio-based modelling, both in a genre-adaptive manner. The proposed techniques are experimentally evaluated at predicting listener ratings related to a set of 600 popular music tracks spanning multiple genres. The results show that ACTwg outperforms a semantic computing technique that does not exploit genre information, and ACTwg-SLPwg outperforms conventional techniques and other genre-adaptive alternatives. In particular, improvements in the prediction rates are obtained for the valence dimension which is typically the most challenging core affect dimension for audio-based annotation. The specificity of genre categories is not crucial for the performance of ACTwg-SLPwg. The study also presents analytical insights into inferring a concise tag-based genre representation for genre-adaptive music mood analysis.

Perceived gender in clear and conversational speech

Much of our understanding of gender perception in voice and speech is based on sustained vowels or single words, which eliminates temporal, prosodic, and articulatory cues available in more natural, connected speech. Further, while many studies have examined acoustic and sociolinguistic differences between male and female voices, the relationship between talker speaking style and perception of gender has not yet been explored. Clear speech, adopted by talkers who perceive some barrier to effective communication, is one such speaking style change. The present study examines the relationship between clear speech and talker gender perception. Clear and conversational neutral sentences produced by all 41 talkers from the Ferguson Clear Speech Database (Ferguson, 2004) were presented to young listeners with normal hearing. They rated the gender of the talker using a visual analog scale with endpoints labeled masculine and feminine, chosen to capture small within-category changes in perceived gender. Acoustic analyses of these sentences, including fundamental frequency, formant frequencies, speaking rate, fundamental frequency range, and cepstral peak prominence will be undertaken to determine the relationship between acoustic correlates of clear speech and listener ratings of femininity or masculinity. Applications to transgender voice therapy will be discussed.

Visuospatial Working Memory Capacity Predicts Physiological Arousal in a Narrative Task.

Physiological arousal that occurs during narrative production is thought to reflect emotional processing and cognitive effort (Bar-Haim et al. in Dev Psychobiol 44:238-249, 2004). The purpose of this study was to determine whether individual differences in visuospatial working memory and/or verbal working memory capacity predict physiological arousal in a narrative task. Visuospatial working memory was a significant predictor of skin conductance level (SCL); verbal working memory was not. When visuospatial working memory interference was imposed, visuospatial working memory was no longer a significant predictor of SCL. Visuospatial interference also resulted in a significant reduction in SCL. Furthermore, listener ratings of narrative quality were contingent upon the visuospatial working memory resources of the narrator. Potential implications for educators and clinical practitioners are discussed.

The Actual vs. Predicted Effects of Intonation Accuracy on Vocal Performance Quality

The belief that intonation accuracy is a key determinant of musical performance quality is ubiquitous in music pedagogy; nonetheless, empirical validation of this belief is lacking. We investigated the effects of intonation accuracy on perceived performance quality and assessed whether music professionals’ beliefs about the importance of intonation are consistent with these effects. In Experiments 1 and 2, participants listened to vocal performances that were in tune, moderately out of tune, or severely out of tune. Only severe mistunings caused decreases in performance quality ratings for intonation insensitive listeners (those who performed poorly on a mistuning detection prescreening). However, both moderate and severe mistunings were detrimental to the ratings of intonation sensitive listeners. These results indicate that moderate mistunings exert a negative influence on the perceived quality of a performance only if the listener can explicitly detect the mistunings. If a listener cannot explicitly detect the mistunings, those mistunings do not implicitly exert a negative influence on the perception of the performance. In Experiment 3, music professionals heard samples of performances from Experiments 1 and 2 in each intonation condition and were asked to estimate the impact of the mistunings on listeners’ ratings of performance quality. Their predictions were compared to the actual performance quality ratings obtained in Experiment 2. Music professionals overestimated the impact of moderate mistunings for both intonation sensitive and insensitive listeners, suggesting that music professionals may hold inaccurate beliefs about the importance of intonation accuracy in vocal performance.

Using Listener Judgments to Investigate Linguistic Influences on L2 Comprehensibility and Accentedness: A Validation and Generalization Study

The current study investigated linguistic influences on comprehensibility (ease of understanding) and accentedness (linguistic nativelikeness) in second language (L2) learners’ extemporaneous speech. Target materials included picture narratives from 40 native French speakers of English from different proficiency levels. The narratives were subsequently rated by 20 native speakers with or without linguistic and pedagogical experience for comprehensibility, accentedness, and 11 linguistic variables spanning the domains of phonology, lexis, grammar, and discourse structure. Results showed that comprehensibility was associated with several linguistic variables (vowel/consonant errors, word stress, fluency, lexis, grammar), whereas accentedness was chiefly linked to pronunciation (vowel/consonant errors, word stress). Native-speaking listeners thus appear to pay particular attention to pronunciation, rather than lexis and grammar, to evaluate nativelikeness but tend to consider various sources of linguistic information in L2 speech in judging comprehensibility. The use of listener ratings (perceptual measures) in evaluating linguistic aspects of learner speech and their implications for language assessment and pedagogy are discussed.

Foreign accentedness of English sentences spoken by Japanese EFL learners and Japanese teachers of English: A first report

This study reports the results of a foreign-accented-rating experiment that investigate the foreign accentedness of spoken English sentences by two Japanese groups, Japanese EFL learners and Japanese teachers of English. This study aims first to investigate whether spoken English sentences by Japanese teachers of English are judged less foreign-accented than those by Japanese EFL learners, and second to investigate whether American-English listeners rate spoken English sentences by Japanese speakers more severely than Japanese listeners do. The stimuli were five sentences adopted from Flege, Munro, and McKay (1995) spoken by 33 Japanese EFL learners and 33 Japanese teachers of English. Ten American-English speakers and ten Japanese speakers were asked to rate the stimuli presented visually and auditorily by clicking along a line on the computer screen for their ratings. The participants’ original responses were converted to 10 scales and were submitted to the analyses. The results showed that spoken English sentences by Japanese teachers of English were rated significantly less Japanese-accented than those by Japanese EFL learners. The results further revealed that as in the previous studies, the American-English speakers rated spoken English sentences by two Japanese groups significantly more severely than Japanese speakers. [Work supported by JSPS.]

Verbal communication with the Blom low profile and Passy-Muir one-way tracheotomy tube speaking valves

ObjectiveTo investigate physiologic parameters, voice production abilities, and functional verbal communication ratings of the Blom low profile voice inner cannula and Passy-Muir one-way tracheotomy tube speaking valves. Study designCase series with planned data collection. SettingLarge, urban, tertiary care teaching hospital. Subjects and methodsReferred sample of 30 consecutively enrolled adults requiring a tracheotomy tube and tested with Blom and Passy-Muir valves. Physiologic parameters recorded were oxygen saturation, respiration rate, and heart rate. Voice production abilities included maximum voice intensity in relation to ambient room noise and maximum phonation duration of the vowel/a/. Functional verbal communication was determined from randomized and blinded listener ratings of counting 1–10, saying the days of the week, and reading aloud the sentence, “There is according to legend a boiling pot of gold at one end.” ResultsThere were no significant differences (p>0.05) between the Blom and Passy-Muir valves for the physiologic parameters of oxygen saturation, respiration rate, and heart rate; voice production abilities of both maximum intensity and duration of/a/; and functional verbal communication ratings. Both valves allowed for significantly greater maximum voice intensity over ambient room noise (p<0.001). ConclusionsThe Blom low profile voice inner cannula and Passy-Muir one-way speaking valves exhibited equipoise regarding patient physiologic parameters, voice production abilities, and functional verbal communication ratings.Learning outcomes: Readers will understand the importance of verbal communication for patients who require a tracheotomy tube; will be able to determine the differences between the Blom low profile voice inner cannula and Passy-Muir one-way tracheotomy tube speaking valves; and will be confident in knowing that both the Blom and Passy-Muir one-way tracheotomy tube speaking valves are equivalent regarding physiological functioning and speech production abilities.

The effect of playback method on listeners' judgments of concert hall auralizations

Previous studies of the perception of concert hall acoustics have generally employed one of two methods for soliciting listeners’ judgments. The first is to have listeners rate the sound of a hall while physically present in that hall. The second is to record the acoustics of a hall and later simulate those acoustics in a laboratory setting. While the first method offers a completely authentic presentation of the concert experience, the second allows for more direct comparisons between different spaces and affords the researcher greater control over experimental variables. Higher-order spherical microphone arrays offer a way to capture the spatial components of a concert hall’s acoustics, which can then be reproduced using a loudspeaker array. In this study, eight different concert and recital halls were measured using both a spherical microphone array and binaural dummy head as receivers. Listeners were then presented with auralizations of the halls using an ambisonic loudspeaker array and headphones, and asked to rate the halls based on subjective preference and on similarity to one another. The responses were analyzed using multidimensional scaling methods in order to examine the effect of the auralization system on the listeners’ judgments.

Online crowdsourcing for efficient rating of speech: A validation study

Blinded listener ratings are essential for valid assessment of interventions for speech disorders, but collecting these ratings can be time-intensive and costly. This study evaluated the validity of speech ratings obtained through online crowdsourcing, a potentially more efficient approach. 100 words from children with /r/ misarticulation were electronically presented for binary rating by 35 phonetically trained listeners and 205 naïve listeners recruited through the Amazon Mechanical Turk (AMT) crowdsourcing platform. Bootstrapping was used to compare different-sized samples of AMT listeners against a “gold standard” (mode across all trained listeners) and an “industry standard” (mode across bootstrapped samples of three trained listeners). There was strong overall agreement between trained and AMT listeners. The “industry standard” level of performance was matched by bootstrapped samples with n=9 AMT listeners. These results support the hypothesis that valid ratings of speech data can be obtained in an efficient manner through AMT. Researchers in communication disorders could benefit from increased awareness of this method.Learning outcomes: Readers will be able to (a) discuss advantages and disadvantages of data collection through the crowdsourcing platform Amazon Mechanical Turk (AMT), (b) describe the results of a validity study comparing samples of AMT listeners versus phonetically trained listeners in a speech-rating task.

Semantic Computing of Moods Based on Tags in Social Media of Music

Social tags inherent in online music services such as Last.fm provide a rich source of information on musical moods. The abundance of social tags makes this data highly beneficial for developing techniques to manage and retrieve mood information, and enables study of the relationships between music content and mood representations with data substantially larger than that available for conventional emotion research. However, no systematic assessment has been done on the accuracy of social tags and derived semantic models at capturing mood information in music. We propose a novel technique called Affective Circumplex Transformation (ACT) for representing the moods of music tracks in an interpretable and robust fashion based on semantic computing of social tags and research in emotion modeling. We validate the technique by predicting listener ratings of moods in music tracks, and compare the results to prediction with the Vector Space Model (VSM), Singular Value Decomposition (SVD), Nonnegative Matrix Factorization (NMF), and Probabilistic Latent Semantic Analysis (PLSA). The results show that ACT consistently outperforms the baseline techniques, and its performance is robust against a low number of track-level mood tags. The results give validity and analytical insights for harnessing millions of music tracks and associated mood data available through social tags in application development.

Association between objective measurement of the speech intelligibility of young people with dysarthria and listener ratings of ease of understanding

Purpose: This study aimed to investigate the association between listeners’ ratings of how much effort it took to understand the speech of young people with cerebral palsy and the percentage of words listeners actually understood.Method: Thirty-one young people with dysarthria and cerebral palsy (16 males, 15 females; mean age = 11 years, SD = 3) were audio recorded repeating single words and producing speech. Objective measures of intelligibility were calculated for multiple familiar and unfamiliar listeners using a forced choice paradigm for single words and verbatim orthographic transcriptions for connected speech. Listeners rated how much effort it took to understand speech in each condition using a 5-point ordinal ease of listening (EOL) scale.Results: Agreement on EOL within rater groups was high (ICC > 0.71). An effect of listener was observed for familiar listeners, but not for unfamiliar listeners. EOL agreement between familiar and unfamiliar listeners was weak–moderate (ICC = 0.46). EOL predicted the percentage of speech actually understood by familiar and unfamiliar listeners (r > 0.56, p < 0.001 for all predictions). Strongest associations between EOL and intelligibility were observed for speakers with mild and profound impairments.Conclusions: The findings of this study demonstrate that listeners can judge how well they have understood dysarthric speech. EOL is associated with listener familiarity, speech task and speech impairment severity. EOL is appropriate for use in clinical practice as a measure of communication activity.

Comparing social meanings across listener and speaker groups: The indexical field of Spanish /s/

AbstractRecent research has suggested that the contribution of individual sociolinguistic variables to the social perception of a speaker is influenced by other available information about the speaker (Campbell-Kibler, 2007; Pharao, Maegaard, Spindler Møller, &amp; Kristiansen, 2014). Here we investigate the impact of listener awareness of regional sociolinguistic variation on sociolinguistic perception. Specifically, we compare how the social meanings attributed to word-internal, preconsonantal /s/ differ based on whether the listeners and speakers use predominantly /s/-weakening Puerto Rican Spanish or predominantly non–/s/-weakening Mexican Spanish. We find that for measures of status, Puerto Rican and Mexican listeners both show a smaller effect of /s/ when rating Puerto Rican as opposed to Mexican speakers. However, we see no effect of speaker nationality on heteronormativity, and Puerto Rican listeners and male Mexican listeners rate strong /s/ as less heteronormative across the board. Mexican female listeners, however, rate strong /s/ as more heteronormative. These results suggest that listeners integrate their own local ideologies with their understanding of regional differences when socially evaluating language variation.

The Validity of Likeability as a New Construct of ESL Pronunciation

The most significant constructs of ESL pronunciation models are accentedness , intelligibility , and comprehensibility (Derwing &amp; Munro 2005; Jenkins, 2000; Pickering, 2006). It is clear though that the assessment of these characteristics shows no consideration for the potential affection of the assessor towards the speaker. Thus, the study presented here proposes a notion named “likeability” (L) to examine whether the ratings of listeners are determined in part by the esteem in which they hold the speaker. A questionnaire compiled by Murphy (2014) was given to 11 US and 11 ELLs. Participants were provided with three YouTube links of interviews of Nelson Mandela, a former President of South Africa, Ban Ki-Moon, the current Secretary-General of the United Nations, and Bandar Bin Sultan, the former Saudi ambassador to the United States. The first hypothesis of the study is that the American raters would be harsher in their judgments of Ban Ki-moon and Bin Sultan due to the political conflict and the dramatic events that have been associated with them in the past but this would not be the case with Mandela who has many ardent admirers in the U.S.A. The second hypothesis is that the participants among the international group who share the language background of the speaker would rate that speaker more highly than the others. Unexpectedly, however, both US and ELL listeners rated Mandela the lowest in terms of comprehensibility and Bin Sultan the most comprehensible. All the participants in the international group provided similar ratings regardless of whether they share the language background of the speaker or not. The paper concludes with a suggestion for more thorough investigation of the concept of likeability in selecting ESL speech models, particularly with assigning speakers with different professional fields, ethnicity, and language background. This investigation is crucial in finding the most suitable models for pronunciation teaching.

Perception of Speech Produced by Native and Nonnative Talkers by Listeners with Normal Hearing and Listeners with Cochlear Implants

PURPOSE The aim of this study was to evaluate the understanding of English sentences produced by native (English) and nonnative (Spanish) talkers by listeners with normal hearing (NH) and listeners with cochlear implants (CIs). METHOD Sentence recognition in noise was measured in adult subjects with CIs and subjects with NH, all of whom were native talkers of American English. Test sentences were from the Hearing in Noise Test (HINT) database and were produced in English by four native and eight nonnative talkers. Subjects also rated the intelligibility and accent for each talker. RESULTS The speech recognition thresholds in noise of subjects with CIs and subjects with NH were 4.23 dB and 1.32 dB poorer with nonnative talkers than with native talkers, respectively. Performance was significantly correlated with talker intelligibility and accent ratings for subjects with CIs but only correlated with talker intelligibility ratings for subjects with NH. For all subjects, performance with individual nonnative talkers was significantly correlated with talkers' number of years of residence in the United States. CONCLUSION CI users exhibited a larger deficit in speech understanding with nonnative talkers than did subjects with NH, relative to native talkers. Nonnative talkers' experience with native culture contributed strongly to speech understanding in noise, intelligibility ratings, and accent ratings of both listeners with NH and listeners with CIs.

Perceptual tonal spaces in whispered speech

Whispered speech is made when no vocal fold vibration occurs. However, pitch from whispers can still be perceived for some listeners because it corresponds to the change in F1 and F2. Listeners rely on the laryngeal movements, which not only associate with pitch change but also result in the change of the tongue position as well as the shape of the oral cavity. This study examines the acoustical-perceptual relationships in the identification of pitch during whispered tone production. Listeners (English and Taiwanese) rate the similarity of fifty AX tonal pairs (Taiwanese [ka℧] and [ɕ] in modal voice vs. whispered voice). Multidimensional scaling (MDS) analysis is used to map English listeners’ and Taiwanese listeners’ perceptual tonal spaces regarding modal speech and whispered speech stimuli. The correlations between the similarity ratings, the reciprocal of their reaction time, and the selected acoustic measures are examined.

The effect of background noise on intelligibility of adults and children with dysphonia

A majority of patients with dysphonia report reduced intelligibility in their daily communication environments. Laryngeal pathology often causes abnormal vibration and incomplete closure of the vocal folds, resulting in increased noise and decreased harmonic power in the speech signal. These acoustic consequences likely make dysphonic speech more difficult to understand, particularly in the presence of background noise. The study tested two hypotheses: (1) intelligibility of dysphonic speech is more negatively affected by background noise than that of normal speech, and (2) listener ratings of intelligibility will correlate with clinical measures of dysphonia. One hundred twenty speech samples were collected from 6 adults and 4 children with normal voice and 6 adults and 4 children with varying degrees of dysphonia. Each sample consisted of a short phrase or sentence and was characterized by two acoustic measures commonly associated with degree of dysphonia: cepstral peak prominence (CPP) and harmonic to noise ratio (HNR). Samples were combined with three levels of “cafeteria” noise ( + 0 dB SNR, + 5 dB SNR, and no noise) and then subjected to a speech perception experiment with 60 normal listeners. This project is ongoing. Preliminary results support hypothesis 1; additional findings related to hypothesis 2 will also be discussed.

The perception of high frequency sibilants in Hungarian male speech

The aim of this paper is to report on an experiment designed to evaluate the perception of high frequency sibilant articulations in Hungarian male speech and to theorise on the results. The main findings of the experiment are that the Hungarian listeners rate high frequency sibilants with femininity. These findings suggest that there is at least some social awareness of sibilant frequency in Hungarian. What follows from this is, in turn, that the sociolinguistic salience of sibilants as a variable is not confined to dialects of English, where the phenomenon has been most thoroughly described and discussed.

Hearing tongue loops: Perceptual sensitivity to acoustic signatures of articulatory dynamics

Previous work has shown that velar stops are produced with a forward movement during closure, forming a forward (anterior) loop for a VCV sequence, when the preceding vowels are back or mid. Are listeners aware of this aspect of articulatory dynamics? The current study used articulatory synthesis to examine how such kinematic patterns are reflected in the acoustics, and whether those acoustic patterns elicit different goodness ratings. In Experiment I, the size and direction of loops was modulated in articulatory synthesis. The resulting stimuli were presented to listeners for a naturalness judgment. Results show that listeners rate forward loops as more natural than backward loops, in agreement with typical productions. Acoustic analysis of the synthetic stimuli shows that forward loops exhibit shorter and shallower VC transitions than CV transitions. In Experiment II, three acoustic parameters were employed incorporating F3-F2 distance, transition slope, and transition length to systematically modulate the magnitude of VC and CV transitions. Listeners rated the naturalness in accord with those of Experiment I. This study reveals that there is sufficient information in the acoustic signature of "velar loops" to affect perceptual preference. Similarity to typical productions seemed to determine preferences, not acoustic distinctiveness.