Robert Mannell

Does a rater's familiarity with a candidate's pronunciation affect the rating in oral proficiency interviews?

This study investigated factors that could affect inter-examiner reliability in the pronunciation assessment component of speaking tests. We hypothesized that the rating of pronunciation is susceptible to variation in assessment due to the amount of exposure examiners have to nonnative English accents. An inter-rater variability analysis was conducted on the English pronunciation ratings of three test candidate interlanguages: Chinese, Korean, and Indian English. Pronunciation was rated by 99 International English Language Testing System (IELTS) examiners across five geographically dispersed test centres. The examiners had either prolonged exposure, or no, or little exposure to the interlanguage of the candidates. A significant proportion of examiners rated pronunciation higher when they had prolonged exposure, and lower when they had no, or little, exposure to the candidates’ interlanguage. The location of the test centre also had a significant effect on the pronunciation rating, independent of the familiarity variable, with a significant proportion of NNS raters scoring candidates from their home country higher than those who were not. It is recommended that interlanguage phonology familiarity should be considered in the design of speaking tests and rater training and that research is required into test centre bias and the phonological judgements and awareness of OPI raters.

Development and evaluation of Mandarin disyllabic materials for speech audiometry in China

The purpose of this study is to develop and evaluate disyllabic Mandarin speech test materials (MSTMs) in order to facilitate wider use of speech audiometry in Chinese audiology clinics. Phonologically balanced Mandarin disyllabic materials with high familiarity were designed based on the basic rules for developing speech materials and the particular characteristics of Mandarin, and recorded digitally. In order to establish the validity and reliability of these Mandarin disyllabic materials, equivalence of difficulty between the word lists was evaluated for a group of 60 subjects (age-range 18–25 years) with normal hearing. Subsequently, performance-intensity (PI) functions were measured in a group of 30 subjects with normal hearing (age-range 18–25 years), and a group of 35 subjects with sensineural hearing loss. The nine lists of Mandarin disyllabic materials were found to have sufficient reliability and validity to be used in clinical situations

Relative Contributions of Temporal Envelope and Fine Structure Cues to Lexical Tone Recognition in Hearing-Impaired Listeners

It has been reported that normal-hearing Chinese speakers base their lexical tone recognition on fine structure regardless of temporal envelope cues. However, a few psychoacoustic and perceptual studies have demonstrated that listeners with sensorineural hearing impairment may have an impaired ability to use fine structure information, whereas their ability to use temporal envelope information is close to normal. The purpose of this study is to investigate the relative contributions of temporal envelope and fine structure cues to lexical tone recognition in normal-hearing and hearing-impaired native Mandarin Chinese speakers. Twenty-two normal-hearing subjects and 31 subjects with various degrees of sensorineural hearing loss participated in the study. Sixteen sets of Mandarin monosyllables with four tone patterns for each were processed through a “chimeric synthesizer” in which temporal envelope from a monosyllabic word of one tone was paired with fine structure from the same monosyllable of other tones. The chimeric tokens were generated in the three channel conditions (4, 8, and 16 channels). Results showed that differences in tone responses among the three channel conditions were minor. On average, 90.9%, 70.9%, 57.5%, and 38.2% of tone responses were consistent with fine structure for normal-hearing, moderate, moderate to severe, and severely hearing-impaired groups respectively, whereas 6.8%, 21.1%, 31.4%, and 44.7% of tone responses were consistent with temporal envelope cues for the above-mentioned groups. Tone responses that were consistent neither with temporal envelope nor fine structure had averages of 2.3%, 8.0%, 11.1%, and 17.1% for the above-mentioned groups of subjects. Pure-tone average thresholds were negatively correlated with tone responses that were consistent with fine structure, but were positively correlated with tone responses that were based on the temporal envelope cues. Consistent with the idea that the spectral resolvability is responsible for fine structure coding, these results demonstrated that, as hearing loss becomes more severe, lexical tone recognition relies increasingly on temporal envelope rather than fine structure cues due to the widened auditory filters.

Development of Mandarin monosyllabic speech test materials in China

In this study, monosyllabic Mandarin speech test materials (MSTMs) were developed for use in word recognition tests for speech audiometry in Chinese audiology clinics. Mandarin monosyllabic materials with high familiarity were designed with regard to phonological balance and recorded digitally with a male voice. Inter-list equivalence of difficulty was evaluated for a group of 60 subjects (aged 18–25 years) with normal hearing. Seven lists with 50 words each were found to be equivalent. These seven equivalent lists were used to measure performance-intensity (PI) functions for a group of 32 subjects with normal hearing and a group of 40 subjects with mild to moderate sensorineural hearing loss. The mean slope of PI function was found to be 4.1%/dB and 2.7%/dB, respectively. The seven lists of Mandarin monosyllabic materials were found to have sufficient reliability and validity to be used in clinical situations.

Contribution of spectral cues to mandarin lexical tone recognition in normal-hearing and hearing-impaired mandarin Chinese speakers

Objective: The purpose of this study was to investigate the contribution of spectral fine structure and spectral envelope cues to recognition of Mandarin lexical tones in normal-hearing and sensorineural hearing-impaired Mandarin-speaking listeners. Design: Four groups of subjects participated in the study, including 20 normal-hearing, 20 moderately, 20 moderately to severely, and 8 severely hearing-impaired listeners. The original speech materials consisted of 16 sets of Mandarin monosyllables spoken by a male and a female. Each monosyllable had four tonal patterns, resulting in a total of 64 combinations of consonants, vowels, and tones. A Linear Predictive Coding (LPC) algorithm was used to create two sets of synthesized materials, including 128 tokens with the original spectral fine structure mixed with the spectral envelope from a different tone, as well as 128 tokens with noise fine structure and the original spectral envelope. All subjects participated in tone recognition tests using the two sets of chimeric tone tokens. Oral responses to tones were recorded and scored as percent correct. Results: Hearing-impaired listeners could take advantage of spectral fine structure in the recognition of lexical tones, but with increasing hearing loss, the ability of hearing-impaired listeners to recognize tones became worse, especially for severely hearing-impaired listeners. Hearing-impaired listeners showed significant differences in tone recognition between the male and female voices. Tone 3 was the easiest tone to perceive, followed by tone 2, whereas tones 1 and 4 were hard for all subjects, particularly when only the spectral envelope cue was available. Hearing-impaired listeners showed a significantly lower level of lexical tone recognition than normal-hearing listeners when using spectral envelope cues in comparison with normal-hearing listeners. Conclusions: These results demonstrate that the spectral fine structure cue dominates lexical tone recognition for all subjects. Listeners with sensorineural hearing impairment showed reduced ability in the recognition of lexical tones using both spectral fine structure and spectral envelope cues, which may result from their impaired auditory spectral resolution.

Mandarin lexical tone recognition in sensorineural hearing-impaired listeners and cochlear implant users

Abstract Conclusions: As the hearing loss becomes more severe, the tone recognition performance of hearing-impaired listeners gradually but slowly reduces. The tone recognition performance of cochlear implant listeners is below or close to the performance of severely hearing-impaired listeners. Objectives: The present study aimed to investigate the Mandarin lexical tone recognition performance of sensorineural hearing-impaired listeners and post-lingually deafened cochlear implant users. Methods: Tone recognition performance was measured for 30 normal-hearing subjects, 41sensorineural hearing-impaired listeners, and 12 cochlear implant users using 128 monosyllables recorded by a male and a female adult native Mandarin speaker. Results: The results indicated that the accuracy of tone recognition was 99.3%, 96.4%, 93.7%, 83.9%, and 81.0% for the normal-hearing, moderate, moderate to severe, severely hearing-impaired, and cochlear implant subjects, respectively. For the hearing-impaired subjects, a significantly negative correlation was observed between tone recognition performance and the audiometric hearing thresholds. For cochlear implant subjects, Tone 3 was the easiest one to perceive and Tone 2 was the hardest one to perceive. They tended to misperceive Tone 1 as Tone 2, and misperceive Tone 2 as Tones 1 and 3.

Text-to-speech rule and dictionary development

Abstract This paper describes the development and evaluation of the grapheme-to-phoneme sub-system of a complete real-time synthesis system under development at Macquarie University. It has been developed around a lexicon knowledge base which contains the 4000–5000 most common English words and which has been augmented by a suffix stripper and a set of grapheme to phoneme rules. Evaluation and development of this system has been facilitated by using weighted statistics which reflect the frequency of occurence of each word in the LOB and Brown corpora of English. These statistics are derived from a test word database which includes all acceptable Australian pronunciations (as defined by the Macquarie Dictionary) of each word, as well as their LOB and Brown frequency counts. The pronunciation derived by the system is compared to the Macquarie Dictionary pronunciations and given a score proportional to its frequency in the two corpora. These scores facilitate decisions to be made about which alterations to the rules or lexicon will have the greatest effect on total system accuracy in ordinary running text (as reflected by the corpora frequencies).

Cochlear Implant Rate Pitch and Melody Perception as a Function of Place and Number of Electrodes

Six Nucleus cochlear implant recipients participated in a study investigating the effect of place of stimulation on melody perception using rate-pitch cues. Each stimulus was a pulse train delivered on either a single electrode or multiple electrodes sequentially. Four spatial stimulation patterns were used: a single apical electrode, a single mid electrode, a pair of electrodes (apical and mid), and 11 electrodes (from apical to mid). Within one block of trials, all stimuli had the same spatial stimulation pattern, with pulse rate varying from 131 to 262 pps. An additional pulse rate range of 262 to 523 pps was tested with the single-electrode stimuli. Two experimental procedures were used: note ranking; and a modified melodies test with backwards and warp modification. In each trial of the modified melodies test, a familiar melody and a version with modified pitch were presented (in random order), and the subject’s task was to select the unmodified melody. There were no significant differences in performance for stimulation on 1, 2, or 11 electrodes, implying that recipients were unable to combine temporal information from different places in the cochlea to give a stronger pitch cue. No advantage of apical electrodes was found: at the lower pulse rates, there were no significant differences between electrodes; and at the higher pulse rates, scores on the apical electrode dropped more than those on the mid electrode.

The Combined effect of speech codec quality and transmission delay on human performance during complex spoken interactions

This paper examines the effect of interaction between speech codec output quality and simulated satellite or VoIP transmission delay time on talker performance in a complex interaction. A hardware test codec (both single and tandem) was compared against a number of processed speech reference conditions to determine the relative subjective quality of the test codecs against conditions with known Mean Opinion Scores (MOS). The two codec conditions plus an additional higher quality condition were then used in an experiment that examined the effect of the interaction of transmitted speech quality and simulated transmission delay on a speech shadowing task and an accompanying error repair task involving two speakers. One person (the “reader”) read a passage. The second person (the “shadower”) shadowed the read passage by repeating immediately the words spoken by the reader. The reader, whilst reading, also listened for errors spoken by the shadower and repaired those errors by verbally reporting them to the shadower. A significant interaction between codec quality and transmission delay was found for the error repair task, but only for cases where the shadower made a significant number of errors. These results suggest that, for highly complex interactions which involve significant cognitive load, human performance will degrade more rapidly with increases in delay for transmission systems using speech codecs with lower quality output. This is assumed to be due to the additional demands upon working memory imposed by the transmission delay.

Acoustic cues for lexical tone perception in hearing-impaired listeners

This series of studies was aimed to investigate how listeners with sensorineural hearing loss (SNHL) and with auditory neuropathy syndrome disorder (ANSD) achieved lexical tone recognition using either the temporal envelope (E) or the fine structure (FS) cues. Five groups of Mandarin-speaking subjects, including (1) 22 normal-hearing subjects, (2) 8 moderate, (3) 13 moderate to severe, (4) 10 severe SNHL patients with various degrees of SNHL, and (5) 10 patients with ANSD, participated in the study. Monosyllabic words were processed through a 16-channel “auditory chimera” in which E from a monosyllabic word of one tone was paired with FS from the same monosyllable of other tones. On average, 92.0%, 67.4%, 58.1%, 37.5%, and 17.1% of the tone responses were consistent with FS cues, while 5.8%, 23.7%, 31.1%, 45.2%, 42.7% of the tone responses were consistent with E cues for the 5 groups of subjects mentioned above. Therefore, as the hearing loss becomes more severe, the ability of SNHL patients to use FS for...