Muddy, muddled, or muffled? Understanding the perception of audio quality in music by hearing aid users.

Introduction: Previous work on audio quality evaluation has demonstrated a developing convergence of the key perceptual attributes underlying judgments of quality, such as timbral, spatial and technical attributes. However, across existing research there remains a limited understanding of the crucial perceptual attributes that inform audio quality evaluation for people with hearing loss, and those who use hearing aids. This is especially the case with music, given the unique problems it presents in contrast to human speech. Method: This paper presents a sensory evaluation study utilising descriptive analysis methods, in which a panel of hearing aid users collaborated, through consensus, to identify the most important perceptual attributes of music audio quality and developed a series of rating scales for future listening tests. Participants (N = 12), with a hearing loss ranging from mild to severe, first completed an online elicitation task, providing single-word terms to describe the audio quality of original and processed music samples; this was completed twice by each participant, once with hearing aids, and once without. Participants were then guided in discussing these raw terms across three focus groups, in which they reduced the term space, identified important perceptual groupings of terms, and developed perceptual attributes from these groups (including rating scales and definitions for each). Results: Findings show that there were seven key perceptual dimensions underlying music audio quality (clarity, harshness, distortion, spaciousness, treble strength, middle strength, and bass strength), alongside a music audio quality attribute and possible alternative frequency balance attributes. Discussion: We outline how these perceptual attributes align with extant literature, how attribute rating instruments might be used in future work, and the importance of better understanding the music listening difficulties of people with varied profiles of hearing loss.

Comparison of ideal mask-based speech enhancement algorithms for speech mixed with white noise at low mixture signal-to-noise ratios.

The literature shows that the intelligibility of noisy speech can be improved by applying an ideal binary or soft gain mask in the time-frequency domain for signal-to-noise ratios (SNRs) between -10 and +10 dB. In this study, two mask-based algorithms are compared when applied to speech mixed with white Gaussian noise (WGN) at lower SNRs, that is, SNRs from -29 to -5 dB. These comprise an Ideal Binary Mask (IBM) with a Local Criterion (LC) set to 0 dB and an Ideal Ratio Mask (IRM). The performance of three intrusive Short-Time Objective Intelligibility (STOI) variants-STOI, STOI+, and Extended Short-Time Objective Intelligibility (ESTOI)-is compared with that of other monaural intelligibility metrics that can be used before and after mask-based processing. The results show that IRMs can be used to obtain near maximal speech intelligibility (>90% for sentence material) even at very low mixture SNRs, while IBMs with LC = 0 provide limited intelligibility gains for SNR < -14 dB. It is also shown that, unlike STOI, STOI+ and ESTOI are suitable metrics for speech mixed with WGN at low SNRs and processed by IBMs with LC = 0 even when speech is high-pass filtered to flatten the spectral tilt before masking.

Dataset of British English speech recordings for psychoacoustics and speech processing research: The clarity speech corpus.

This paper presents the Clarity Speech Corpus, a publicly available, forty speaker British English speech dataset. The corpus was created for the purpose of running listening tests to gauge speech intelligibility and quality in the Clarity Project, which has the goal of advancing speech signal processing by hearing aids through a series of challenges. The dataset is suitable for machine learning and other uses in speech and hearing technology, acoustics and psychoacoustics. The data comprises recordings of approximately 10,000 sentences drawn from the British National Corpus (BNC) with suitable length, words and grammatical construction for speech intelligibility testing. The collection process involved the selection of a subset of BNC sentences, the recording of these produced by 40 British English speakers, and the processing of these recordings to create individual sentence recordings with associated transcripts and metadata.

Intelligibility prediction for speech mixed with white Gaussian noise at low signal-to-noise ratios.

The effect of additive white Gaussian noise and high-pass filtering on speech intelligibility at signal-to-noise ratios (SNRs) from -26 to 0 dB was evaluated using British English talkers and normal hearing listeners. SNRs below -10 dB were considered as they are relevant to speech security applications. Eight objective metrics were assessed: short-time objective intelligibility (STOI), a proposed variant termed STOI+, extended short-time objective intelligibility (ESTOI), normalised covariance metric (NCM), normalised subband envelope correlation metric (NSEC), two metrics derived from the coherence speech intelligibility index (CSII), and an envelope-based regression method speech transmission index (STI). For speech and noise mixtures associated with intelligibility scores ranging from 0% to 98%, STOI+ performed at least as well as other metrics and, under some conditions, better than STOI, ESTOI, STI, NSEC, CSIIMid, and CSIIHigh. Both STOI+ and NCM were associated with relatively low prediction error and bias for intelligibility prediction at SNRs from -26 to 0 dB. STI performed least well in terms of correlation with intelligibility scores, prediction error, bias, and reliability. Logistic regression modeling demonstrated that high-pass filtering, which increases the proportion of high to low frequency energy, was detrimental to intelligibility for SNRs between -5 and -17 dB inclusive.

A set of equations for numerically calculating the interaural level difference in the horizontal plane.

The variation of interaural level difference (ILD) with direction and frequency is particularly complex and convoluted. The purpose of this work was to determine a set of parametric equations that can be used to calculate ILDs continuously at any value of frequency and azimuth in the horizontal plane. They were derived by fitting equations to ILDs derived from the azimuthal-dependence data tabulated by Shaw and Vaillancourt [(1985). J. Acoust. Soc Am. 78, 1120-1123] and assuming left-right symmetry. The equations are shown to fit those data to an overall RMS error less than 0.5 dB.

The Effect of Pulmonary Function on the Incidence of Vocal Fatigue Among Teachers.

INTRODUCTION: Females face a significantly higher risk of presenting with voice problems than males. This discrepancy has been associated with a number of differences in respiratory behavior and the physiology of the laryngeal and endocrine systems. METHODS: In conjunction with established spirometry measures, the Vocal Fatigue Index (VFI) was used to determine (1) if there is a relationship between base pulmonary function and vocal fatigue among teachers; and (2) if that relationship is different in females from males. One hundred and twenty-two elementary and middle school teachers (96 females and 26 males) from the Jordan School District in Northern Utah participated in the study. RESULTS: VFI factors were predictors of the outcomes of several raw spirometry measures for female participants, but the same predictive relationship was not found for male participants. Additionally, there appeared to be no relationship between VFI and spirometry measures in females when using normalized, rather than raw, spirometry metrics. CONCLUSIONS: The results suggest that the pulmonary physiology that would result in reduced raw pulmonary function, in combination with other differences associated with gender, may lead to a greater incidence of vocal fatigue among female teachers than their male counterparts.

Speech produced in noise: Relationship between listening difficulty and acoustic and durational parameters.

Conversational speech produced in noise can be characterised by increases in intelligibility relative to such speech produced in quiet. Listening difficulty (LD) is a metric that can be used to evaluate speech transmission performance more sensitively than intelligibility scores in situations in which performance is likely to be high. The objectives of the present study were to evaluate the LD of speech produced in different noise and style conditions, to evaluate the spectral and durational speech modifications associated with these conditions, and to determine whether any of the spectral and durational parameters predicted LD. Nineteen subjects were instructed to speak at normal and loud volumes in the presence of background noise at 40.5 dB(A) and babble noise at 61 dB(A). The speech signals were amplitude-normalised, combined with pink noise to obtain a signal-to-noise ratio of -6 dB, and presented to twenty raters who judged their LD. Vowel duration, fundamental frequency and the proportion of the spectral energy in high vs low frequencies increased with the noise level within both styles. LD was lowest when the speech was produced in the presence of high level noise and at a loud volume, indicating improved intelligibility. Spectrum balance was observed to predict LD.

Evaluation of the starting point of the Lombard Effect.

Speakers increase their vocal effort when their communication is disturbed by noise. This adaptation is termed the Lombard effect. The aim of the present study was to determine whether this effect has a starting point. Hence, the effects of noise at levels between 20 and 65 dB(A) on vocal effort (quantified by sound pressure level) and on both perceived noise disturbance and perceived vocal discomfort were evaluated. Results indicate that there is a Lombard effect change-point at a background noise level (Ln) of 43.3 dB(A). This change-point is anticipated by noise disturbance, and is followed by a high magnitude of vocal discomfort.

Silence and Voicing Accumulations in Italian Primary School Teachers With and Without Voice Disorders.

OBJECTIVES: The relationship between the silence and voicing accumulations of primary school teachers and the teachers' clinical status was examined to determine whether more voicing accumulations and fewer silence accumulations were measured for the vocally unhealthy subjects than for the healthy subjects, which would imply more vocal loading and fewer short-term recovery moments. METHODS: Twenty-six Italian primary school teachers were allocated by clinicians to three groups: (1) with organic voice disorders, (2) with subjectively mild organic alteration or functional voice symptoms, and (3) normal voice quality and physiology. Continuous silence and voicing periods were measured with the APM3200 during the teachers' 4-hour workdays. The accumulations were grouped into seven time intervals, ranging from 0.03-0.9 to 3.16-10 seconds, according to Italian prosody. The effects of group on silence and voicing accumulations were evaluated. RESULTS: Regarding silence accumulations, Group 1 accumulated higher values in intervals between 0.1 and 3.15 seconds than other groups, whereas Groups 2 and 3 did not differ from each other. Voicing accumulations between 0.17 and 3.15 seconds were higher for subjects with a structural disorder. A higher time dose was accumulated by these subjects (40.6%) than other subjects (Group 2, 31.9%; Group 3, 32.3%). CONCLUSIONS: Although previous research has suggested that a rest period of a few seconds may produce some vocal fatigue recovery, these results indicate that periods shorter than 3.16 seconds may not have an observable effect on recovery. The results provide insight into how vocal fatigue and vocal recovery may relate to voice disorders in occupational voice users.

Effect of Training and Level of External Auditory Feedback on the Singing Voice: Pitch Inaccuracy.

BACKGROUND: One of the most important aspects of singing is the control of fundamental frequency. OBJECTIVES: The effects on pitch inaccuracy, defined as the distance in cents in equally tempered tuning between the reference note and the sung note, of the following conditions were evaluated: (1) level of external feedback, (2) tempo (slow or fast), (3) articulation (legato or staccato), (4) tessitura (low, medium, or high), and (5) semi-phrase direction (ascending or descending). METHODS: The subjects were 10 nonprofessional singers and 10 classically trained professional or semi-professional singers (10 men and 10 women). Subjects sang one octave and a fifth arpeggi with three different levels of external auditory feedback, two tempi, and two articulations (legato or staccato). RESULTS: It was observed that inaccuracy was greatest in the descending semi-phrase arpeggi produced at a fast tempo and with a staccato articulation, especially for nonprofessional singers. The magnitude of inaccuracy was also relatively large in the high tessitura relative to the low and the medium tessitura for such singers. Contrary to predictions, when external auditory feedback was strongly attenuated by the hearing protectors, nonprofessional singers showed greater pitch accuracy than in the other external feedback conditions. This finding indicates the importance of internal auditory feedback in pitch control. CONCLUSIONS: With an increase in training, the singer's pitch inaccuracy decreases.

Effects of speech style, room acoustics, and vocal fatigue on vocal effort.

Vocal effort is a physiological measure that accounts for changes in voice production as vocal loading increases. It has been quantified in terms of sound pressure level (SPL). This study investigates how vocal effort is affected by speaking style, room acoustics, and short-term vocal fatigue. Twenty subjects were recorded while reading a text at normal and loud volumes in anechoic, semi-reverberant, and reverberant rooms in the presence of classroom babble noise. The acoustics in each environment were modified by creating a strong first reflection in the talker position. After each task, the subjects answered questions addressing their perception of the vocal effort, comfort, control, and clarity of their own voice. Variation in SPL for each subject was measured per task. It was found that SPL and self-reported effort increased in the loud style and decreased when the reflective panels were present and when reverberation time increased. Self-reported comfort and control decreased in the loud style, while self-reported clarity increased when panels were present. The lowest magnitude of vocal fatigue was experienced in the semi-reverberant room. The results indicate that early reflections may be used to reduce vocal effort without modifying reverberation time.

Effect of Training and Level of External Auditory Feedback on the Singing Voice: Volume and Quality.

BACKGROUND: Previous research suggests that classically trained professional singers rely not only on external auditory feedback but also on proprioceptive feedback associated with internal voice sensitivities. OBJECTIVES: The Lombard effect and the relationship between sound pressure level (SPL) and external auditory feedback were evaluated for professional and nonprofessional singers. Additionally, the relationship between voice quality, evaluated in terms of singing power ratio (SPR), and external auditory feedback, level of accompaniment, voice register, and singer gender was analyzed. METHODS: The subjects were 10 amateur or beginner singers and 10 classically trained professional or semiprofessional singers (10 men and 10 women). Subjects sang an excerpt from the Star-Spangled Banner with three different levels of the accompaniment, 70, 80, and 90 dBA and with three different levels of external auditory feedback. SPL and SPR were analyzed. RESULTS: The Lombard effect was stronger for nonprofessional singers than professional singers. Higher levels of external auditory feedback were associated with a reduction in SPL. As predicted, the mean SPR was higher for professional singers than nonprofessional singers. Better voice quality was detected in the presence of higher levels of external auditory feedback. CONCLUSIONS: With an increase in training, the singer's reliance on external auditory feedback decreases.

Locus equations and coarticulation in three Australian languages.

Locus equations were applied to F2 data for bilabial, alveolar, retroflex, palatal, and velar plosives in three Australian languages. In addition, F2 variance at the vowel-consonant boundary, and, by extension, consonantal coarticulatory sensitivity, was measured. The locus equation slopes revealed that there were place-dependent differences in the magnitude of vowel-to-consonant coarticulation. As in previous studies, the non-coronal (bilabial and velar) consonants tended to be associated with the highest slopes, palatal consonants tended to be associated with the lowest slopes, and alveolar and retroflex slopes tended to be low to intermediate. Similarly, F2 variance measurements indicated that non-coronals displayed greater coarticulatory sensitivity to adjacent vowels than did coronals. Thus, both the magnitude of vowel-to-consonant coarticulation and the magnitude of consonantal coarticulatory sensitivity were seen to vary inversely with the magnitude of consonantal articulatory constraint. The findings indicated that, unlike results reported previously for European languages such as English, anticipatory vowel-to-consonant coarticulation tends to exceed carryover coarticulation in these Australian languages. Accordingly, on the F2 variance measure, consonants tended to be more sensitive to the coarticulatory effects of the following vowel. Prosodic prominence of vowels was a less significant factor in general, although certain language-specific patterns were observed.

Effects of voice style, noise level, and acoustic feedback on objective and subjective voice evaluations.

Speakers adjust their vocal effort when communicating in different room acoustic and noise conditions and when instructed to speak at different volumes. The present paper reports on the effects of voice style, noise level, and acoustic feedback on vocal effort, evaluated as sound pressure level, and self-reported vocal fatigue, comfort, and control. Speakers increased their level in the presence of babble and when instructed to talk in a loud style, and lowered it when acoustic feedback was increased and when talking in a soft style. Self-reported responses indicated a preference for the normal style without babble noise.

Teachers and Teaching: Speech Production Accommodations Due to Changes in the Acoustic Environment.

School teachers have an elevated risk of voice problems due to the vocal demands in the workplace. This manuscript presents the results of three studies investigating teachers' voice use at work. In the first study, 57 teachers were observed for 2 weeks (waking hours) to compare how they used their voice in the school environment and in non-school environments. In a second study, 45 participants performed a short vocal task in two different rooms: a variable acoustic room and an anechoic chamber. Subjects were taken back and forth between the two rooms. Each time they entered the variable acoustics room, the reverberation time and/or the background noise condition had been modified. In this latter study, subjects responded to questions about their vocal comfort and their perception of changes in the acoustic environment. In a third study, 20 untrained vocalists performed a simple vocal task in the following conditions: with and without background babble and with and without transparent plexiglass shields to increase the first reflection. Relationships were examined between [1] the results for the room acoustic parameters; [2] the subjects' perception of the room; and [3] the recorded speech acoustic. Several differences between male and female subjects were found; some of those differences held for each room condition (at school vs. not at school, reverberation level, noise level, and early reflection).

Videos Influence Behavior Change Measures for Voice and Speech in Individuals with Parkinson's Disease.

The majority of individuals with Parkinson's disease (PD) experience voice and speech difficulties at some point over the course of the disease. Voice therapy has been found to help improve voice and speech in individuals with PD, but the majority of these individuals do not enroll in voice therapy. The purpose of this study was to determine whether watching short videos about voice symptoms and treatment in Parkinson's disease influences readiness to change, stages of change, and self-efficacy in individuals with PD. Eight individuals with PD participated in the study. Fifteen videos were chosen, three representing each of the five stages of change. We chose videos from YouTube that represented variety in speakers, content, and genre. We found that readiness to change significantly increased after watching videos, suggesting that watching videos helped these individuals move closer to actively improving their voice and speech. In addition, five of the eight participants showed forward movement in stages of change. Finally, self-efficacy demonstrated a positive trend following video watching. Overall, our results demonstrate that watching videos available on the internet can influence individuals with Parkinson's disease in changing vocal behavior. Implications for future wireless health applications are described.