Auditory and acoustic descriptions are presented for the tones and tone sandhi of two Wenzhou spe... more Auditory and acoustic descriptions are presented for the tones and tone sandhi of two Wenzhou speakers in disyllabic words with tones from the historical ping + qu categories. Differences between the speakers both in isolation tones and tone sandhi are demonstrated. It is suggested that the opaque morphotonemic relationships between sandhi and isolation tones are better accounted for historically, rather than as synchronic operations on unique underliers. The effects of stress / metrical strength are an important factor.
Annual Conference of the International Speech Communication Association, 1998
Acoustic and auditory data are presented from an ongoing large-scale investigation into the tones... more Acoustic and auditory data are presented from an ongoing large-scale investigation into the tones and tone sandhi of the Wu dialects of Zhejiang province in East Central China. The citation tones from 4 sites (3 hitherto undescribed) in the little known Central Zhejiang area are described: Pujiang, Tonglu, Shengxian and Tiantai. Mean F0 and duration data are presented for the tones of these dialects. The data demonstrate a high degree of complexity, having no less than 25 Linguistic-tonetically different tones, including 3 different falling tones, and 4 different falling-level tones. The nature of the complexity of these forms is discussed.
L'auteur donne une description perceptuelle, acoustique et physiologique des modes phonat... more L'auteur donne une description perceptuelle, acoustique et physiologique des modes phonatoires - chuchottement, voix chuchottée, et "grondement", qui caractérisent les syllabes aux tons yang du dialecte Wu de Zhenhai. Il en discute également les implications phonologiques.
Important aspects of Technical Forensic Speaker Recognition, particularly those associated with e... more Important aspects of Technical Forensic Speaker Recognition, particularly those associated with evidence, are exemplified and critically discussed, and comparisons drawn with generic Speaker Recognition. The centrality of the Likelihood Ratio of BayesÕ theorem in correctly evaluating strength of forensic speech evidence is emphasised, as well as the many problems involved in its accurate estimation. It is pointed out that many different types of evidence are of use, both experimentally and forensically, in discriminating same-speaker from different-speaker speech samples, and some examples are given from real forensic case-work to illustrate the Likelihood Ratio-based approach. The extent to which Technical Forensic Speaker Recognition meets the Daubert requirement of testability is also discussed.
This paper describes a discrimination experiment in forensic speaker recognition using the Austra... more This paper describes a discrimination experiment in forensic speaker recognition using the Australian English diphthong /a�/. A two-level kernel density multivariate likelihood ratio is used as a discriminant function to investigate how well noncontemporaneous same-speaker speech samples of /a� / can be forensically discriminated from different-speaker speech samples using just this diphthong’s Fpattern at its two targets. Natural speech elicited from 25 Australian-English speaking males is extrinsically evaluated against a reference population of 166 male speakers from Bernard’s database. Comparing samples with 12 diphthong tokens each, a respectable well-calibrated EER of between ca. 8 % and 10 % is obtained. Forensically important aspects of the results are discussed, including an assessment of the suitability of the reference population. 1.
Forensic Phonetics is an important application of Linguistics that has emerged as a discipline ov... more Forensic Phonetics is an important application of Linguistics that has emerged as a discipline over the last decade. This paper describes a Forensic Phonetic experiment which investigates the nature of within- and between-speaker variation in the acoustic characteristics of the wordhelloin demonstrably similar-sounding voices. The nature of within-segment variation is determined in repeats of the same word said under different prosodic conditions in order to exclude as much of the linguistically determined variation as is consistent with the realities of the forensic situation, thus providing a good estimate of variation associated with speakers. Intonationally varying tokens of the naturally produced single word utterancehellofrom six adult Australian males are compared with respect to fundamental frequency, and to centre frequencies and bandwidths of the F-pattern below 5 KHz. This comprises the first five formants and extra resonances, including a possible singer’s formant and tr...
Acoustic descriptions of citation tones are presented for speakers from seven sites in the Oujian... more Acoustic descriptions of citation tones are presented for speakers from seven sites in the Oujiang sub-group of the Wu dialects of east central China. The homogeneity of two tones within the Oujiang sub-group is demonstrated by quantified comparison with Shanghai tones, and it is pointed out that the normalised tonal acoustics can be interpreted historically as reconstructions of proto-Oujiang tones. 1.
Acoustic and auditory data are presented for the citation tones of single speakers from nine site... more Acoustic and auditory data are presented for the citation tones of single speakers from nine sites (eight hitherto undescribed in English) from the little-studied Chuqu 处衢 subgroup of Wu 吴 in East Central China: Lìshuĭ 丽水, Lóngquán 龙泉, Qìngyuán 庆元, Lóngyóu 龙游, Jìnyún 缙云, Qīngtián 青田, Yúnhé 云和, Jĭngníng 景宁, and Táishùn 泰顺. The data demonstrate a high degree of complexity, having no less than 22 linguistic-tonetically different tones. The nature of the complexity of these forms is discussed, especially with respect to whether the variation is continuous or categorical, and inferences are drawn on their historical development.
Mean fundamental frequency and duration data are presented for citation tones of Hong Kong Canton... more Mean fundamental frequency and duration data are presented for citation tones of Hong Kong Cantonese on short stopped syllables for five male and five female young native speakers. The relationship between the acoustics of the speakers’ stopped and unstopped tones is examined and it is shown that the low-stopped tone F0 is for most speakers acoustically more similar to the low falling unstopped-tone than to the low level unstopped-tone. It is furthermore shown with likelihood ratios that, contrary to current assumptions concerning the phonology of Cantonese, and the notion of phonological categoricality in general, it is not possible to state for Cantonese that, on the basis of its F0, the low stopped-tone is related to any single unstopped tone.
Acoustic and auditory data are presented for the citation tones of single speakers from nine site... more Acoustic and auditory data are presented for the citation tones of single speakers from nine sites (eight hitherto undescribed in English) from the little-studied Chuqu 处衢 subgroup of Wu 吴 in East Central China: Lishuĭ丽水, Longquan龙泉, Qingyuan庆元, Longyou龙游, Jinyun缙云, Qīngtian青田, Yunhe 云和, Jĭngning景宁, and Taishun 泰顺. The data demonstrate a high degree of complexity, having no less than 22 linguistic-tonetically different tones. The nature of the complexity of these forms is discussed, especially with respect to whether the variation is continuous or categorical, and inferences are drawn on their historical development. Index terms: tone, linguistic-tonetics, Chinese, Wu, Chuqu, areal variation.
This paper describes a likelihood ratio-based forensic voice comparison experiment using acoustic... more This paper describes a likelihood ratio-based forensic voice comparison experiment using acoustics from the formant pattern of the Cantonese diphthong /ei/ in the word /sei/ “4”. Non-contemporaneous recordings from 18 young male native speakers are used, each containing nine replicates. Multivariate likelihood ratios derived from the cubic polynomial coefficients of the /ei/ F-pattern are obtained with a two-level kernel density multivariate likelihood ratio. A calibrated cross-validated log-likelihood ratio cost of 0.46 is obtained, with an equal error rate of about 14%. It is concluded that the /ei/ diphthongal F-pattern can be useful in forensic speaker identification in Cantonese.
An acoustic and auditory-phonetic analysis of the data in Bain's emergency call shows strong evid... more An acoustic and auditory-phonetic analysis of the data in Bain's emergency call shows strong evidence in support of the hypothesis that the disputed utterance begins with Bain saying I can't, and against the hypothesis that it begins with I shot. Neither is it convincing to argue that the disputed utterance was not speech.
The tonal acoustics are described of a Nyo speaker fluent in 3 tonally distinct varieties of Tai:... more The tonal acoustics are described of a Nyo speaker fluent in 3 tonally distinct varieties of Tai: Standard Thai with 5 tonemes, Lao with 7, and Nyo with 4 tonemes. Mean fundamental frequency and duration values are presented for the allotones of these tonemes on syllables with unstopped Rhymes. A perceptual experiment is described to determine how many of the 16 allotones are identifiable by Nyo tridialectal listeners. The acoustic and perceptual data are used to demonstrate the existence of 10 linguistic-tonetically distinct tones.
Auditory and acoustic data are presented to document some tonologically challenging aspects of th... more Auditory and acoustic data are presented to document some tonologically challenging aspects of the seven tones of Wénchéng, a south-west Oujiang Wu dialect of Chinese. Tones are described both in isolation form and in selected lexical tone sandhi combinations. It is shown how the complexity of the tonal system results in clashes between tonological representation and phonetic reality.
This paper describes a likelihood ratio-based forensic voice comparison using the Cantonese triph... more This paper describes a likelihood ratio-based forensic voice comparison using the Cantonese triphthong /iau/. The features for comparison are both the formant pattern and tonal F0 of /iau/ contained in 15 native speakers’ non-contemporaneous natural speech. Kernel density multivariate likelihood ratios are calculated based on polynomial coefficients converted from raw frequency values of F1, F2, F3 and F0. A loglikelihood-ratio cost (Cllr) value for the F-pattern of 0.6 is obtained, showing that comparisons with /iau/ F-pattern have reasonable validity, and also reasonable strength of evidence can be achieved. Likelihood ratios calculated based on tonal F0 do not show good performance (Cllr=0.95), and are therefore not fused with the F-pattern LRs.
An acoustically-based description is given of the isolation tones and right-dominant tone sandhi ... more An acoustically-based description is given of the isolation tones and right-dominant tone sandhi in disyllabic words of a male speaker of the Chinese Oūjiāng 甌江 Wu吳 dialect of Wencheng 文成. His seven isolation tones show typical Wu complexity, comprising two mid-level, two rising, two falling-rising and one depressed level pitch shapes. Typical too is his three-way voicing contrast in syllable-Onset stops. However, the typical Wu relationship between tonal register and phonemic Onset voicing is shown to be disrupted, Onset voicing no longer correlating with tonal pitch height. The word-final tones in sandhi are shown to be straightforwardly related, phonologically and phonetically, to the isolation tones, with biuniqueness preserved. The realization of the word-initial tones in sandhi, on the other hand, involves complex mergers conditioned by largely non-phonetic factors related to historical tone categories, resulting in five extra sandhi tones that do not occur in isolation. It is...
A large-scale forensic discrimination experiment is described that investigates how well same-spe... more A large-scale forensic discrimination experiment is described that investigates how well same-speaker speech samples can be discriminated from different-speaker speech samples using acoustic parameters from Australian English vowels. A multivariate likelihood ratio is used as a discriminant function on the five tense and six lax vowel phonemes of 171 male speakers. In 171 target trials and 58,140 non-target trials, comparing samples with just one token per vowel each gives EERs of between 17% and 40%, which drop to between 10% and 14% when fused. Kernel density modeling of the reference population is shown to outperform normal, and performance degrades under quasi-realistic conditions.
Auditory and acoustic descriptions are presented for the tones and tone sandhi of two Wenzhou spe... more Auditory and acoustic descriptions are presented for the tones and tone sandhi of two Wenzhou speakers in disyllabic words with tones from the historical ping + qu categories. Differences between the speakers both in isolation tones and tone sandhi are demonstrated. It is suggested that the opaque morphotonemic relationships between sandhi and isolation tones are better accounted for historically, rather than as synchronic operations on unique underliers. The effects of stress / metrical strength are an important factor.
Annual Conference of the International Speech Communication Association, 1998
Acoustic and auditory data are presented from an ongoing large-scale investigation into the tones... more Acoustic and auditory data are presented from an ongoing large-scale investigation into the tones and tone sandhi of the Wu dialects of Zhejiang province in East Central China. The citation tones from 4 sites (3 hitherto undescribed) in the little known Central Zhejiang area are described: Pujiang, Tonglu, Shengxian and Tiantai. Mean F0 and duration data are presented for the tones of these dialects. The data demonstrate a high degree of complexity, having no less than 25 Linguistic-tonetically different tones, including 3 different falling tones, and 4 different falling-level tones. The nature of the complexity of these forms is discussed.
L'auteur donne une description perceptuelle, acoustique et physiologique des modes phonat... more L'auteur donne une description perceptuelle, acoustique et physiologique des modes phonatoires - chuchottement, voix chuchottée, et "grondement", qui caractérisent les syllabes aux tons yang du dialecte Wu de Zhenhai. Il en discute également les implications phonologiques.
Important aspects of Technical Forensic Speaker Recognition, particularly those associated with e... more Important aspects of Technical Forensic Speaker Recognition, particularly those associated with evidence, are exemplified and critically discussed, and comparisons drawn with generic Speaker Recognition. The centrality of the Likelihood Ratio of BayesÕ theorem in correctly evaluating strength of forensic speech evidence is emphasised, as well as the many problems involved in its accurate estimation. It is pointed out that many different types of evidence are of use, both experimentally and forensically, in discriminating same-speaker from different-speaker speech samples, and some examples are given from real forensic case-work to illustrate the Likelihood Ratio-based approach. The extent to which Technical Forensic Speaker Recognition meets the Daubert requirement of testability is also discussed.
This paper describes a discrimination experiment in forensic speaker recognition using the Austra... more This paper describes a discrimination experiment in forensic speaker recognition using the Australian English diphthong /a�/. A two-level kernel density multivariate likelihood ratio is used as a discriminant function to investigate how well noncontemporaneous same-speaker speech samples of /a� / can be forensically discriminated from different-speaker speech samples using just this diphthong’s Fpattern at its two targets. Natural speech elicited from 25 Australian-English speaking males is extrinsically evaluated against a reference population of 166 male speakers from Bernard’s database. Comparing samples with 12 diphthong tokens each, a respectable well-calibrated EER of between ca. 8 % and 10 % is obtained. Forensically important aspects of the results are discussed, including an assessment of the suitability of the reference population. 1.
Forensic Phonetics is an important application of Linguistics that has emerged as a discipline ov... more Forensic Phonetics is an important application of Linguistics that has emerged as a discipline over the last decade. This paper describes a Forensic Phonetic experiment which investigates the nature of within- and between-speaker variation in the acoustic characteristics of the wordhelloin demonstrably similar-sounding voices. The nature of within-segment variation is determined in repeats of the same word said under different prosodic conditions in order to exclude as much of the linguistically determined variation as is consistent with the realities of the forensic situation, thus providing a good estimate of variation associated with speakers. Intonationally varying tokens of the naturally produced single word utterancehellofrom six adult Australian males are compared with respect to fundamental frequency, and to centre frequencies and bandwidths of the F-pattern below 5 KHz. This comprises the first five formants and extra resonances, including a possible singer’s formant and tr...
Acoustic descriptions of citation tones are presented for speakers from seven sites in the Oujian... more Acoustic descriptions of citation tones are presented for speakers from seven sites in the Oujiang sub-group of the Wu dialects of east central China. The homogeneity of two tones within the Oujiang sub-group is demonstrated by quantified comparison with Shanghai tones, and it is pointed out that the normalised tonal acoustics can be interpreted historically as reconstructions of proto-Oujiang tones. 1.
Acoustic and auditory data are presented for the citation tones of single speakers from nine site... more Acoustic and auditory data are presented for the citation tones of single speakers from nine sites (eight hitherto undescribed in English) from the little-studied Chuqu 处衢 subgroup of Wu 吴 in East Central China: Lìshuĭ 丽水, Lóngquán 龙泉, Qìngyuán 庆元, Lóngyóu 龙游, Jìnyún 缙云, Qīngtián 青田, Yúnhé 云和, Jĭngníng 景宁, and Táishùn 泰顺. The data demonstrate a high degree of complexity, having no less than 22 linguistic-tonetically different tones. The nature of the complexity of these forms is discussed, especially with respect to whether the variation is continuous or categorical, and inferences are drawn on their historical development.
Mean fundamental frequency and duration data are presented for citation tones of Hong Kong Canton... more Mean fundamental frequency and duration data are presented for citation tones of Hong Kong Cantonese on short stopped syllables for five male and five female young native speakers. The relationship between the acoustics of the speakers’ stopped and unstopped tones is examined and it is shown that the low-stopped tone F0 is for most speakers acoustically more similar to the low falling unstopped-tone than to the low level unstopped-tone. It is furthermore shown with likelihood ratios that, contrary to current assumptions concerning the phonology of Cantonese, and the notion of phonological categoricality in general, it is not possible to state for Cantonese that, on the basis of its F0, the low stopped-tone is related to any single unstopped tone.
Acoustic and auditory data are presented for the citation tones of single speakers from nine site... more Acoustic and auditory data are presented for the citation tones of single speakers from nine sites (eight hitherto undescribed in English) from the little-studied Chuqu 处衢 subgroup of Wu 吴 in East Central China: Lishuĭ丽水, Longquan龙泉, Qingyuan庆元, Longyou龙游, Jinyun缙云, Qīngtian青田, Yunhe 云和, Jĭngning景宁, and Taishun 泰顺. The data demonstrate a high degree of complexity, having no less than 22 linguistic-tonetically different tones. The nature of the complexity of these forms is discussed, especially with respect to whether the variation is continuous or categorical, and inferences are drawn on their historical development. Index terms: tone, linguistic-tonetics, Chinese, Wu, Chuqu, areal variation.
This paper describes a likelihood ratio-based forensic voice comparison experiment using acoustic... more This paper describes a likelihood ratio-based forensic voice comparison experiment using acoustics from the formant pattern of the Cantonese diphthong /ei/ in the word /sei/ “4”. Non-contemporaneous recordings from 18 young male native speakers are used, each containing nine replicates. Multivariate likelihood ratios derived from the cubic polynomial coefficients of the /ei/ F-pattern are obtained with a two-level kernel density multivariate likelihood ratio. A calibrated cross-validated log-likelihood ratio cost of 0.46 is obtained, with an equal error rate of about 14%. It is concluded that the /ei/ diphthongal F-pattern can be useful in forensic speaker identification in Cantonese.
An acoustic and auditory-phonetic analysis of the data in Bain's emergency call shows strong evid... more An acoustic and auditory-phonetic analysis of the data in Bain's emergency call shows strong evidence in support of the hypothesis that the disputed utterance begins with Bain saying I can't, and against the hypothesis that it begins with I shot. Neither is it convincing to argue that the disputed utterance was not speech.
The tonal acoustics are described of a Nyo speaker fluent in 3 tonally distinct varieties of Tai:... more The tonal acoustics are described of a Nyo speaker fluent in 3 tonally distinct varieties of Tai: Standard Thai with 5 tonemes, Lao with 7, and Nyo with 4 tonemes. Mean fundamental frequency and duration values are presented for the allotones of these tonemes on syllables with unstopped Rhymes. A perceptual experiment is described to determine how many of the 16 allotones are identifiable by Nyo tridialectal listeners. The acoustic and perceptual data are used to demonstrate the existence of 10 linguistic-tonetically distinct tones.
Auditory and acoustic data are presented to document some tonologically challenging aspects of th... more Auditory and acoustic data are presented to document some tonologically challenging aspects of the seven tones of Wénchéng, a south-west Oujiang Wu dialect of Chinese. Tones are described both in isolation form and in selected lexical tone sandhi combinations. It is shown how the complexity of the tonal system results in clashes between tonological representation and phonetic reality.
This paper describes a likelihood ratio-based forensic voice comparison using the Cantonese triph... more This paper describes a likelihood ratio-based forensic voice comparison using the Cantonese triphthong /iau/. The features for comparison are both the formant pattern and tonal F0 of /iau/ contained in 15 native speakers’ non-contemporaneous natural speech. Kernel density multivariate likelihood ratios are calculated based on polynomial coefficients converted from raw frequency values of F1, F2, F3 and F0. A loglikelihood-ratio cost (Cllr) value for the F-pattern of 0.6 is obtained, showing that comparisons with /iau/ F-pattern have reasonable validity, and also reasonable strength of evidence can be achieved. Likelihood ratios calculated based on tonal F0 do not show good performance (Cllr=0.95), and are therefore not fused with the F-pattern LRs.
An acoustically-based description is given of the isolation tones and right-dominant tone sandhi ... more An acoustically-based description is given of the isolation tones and right-dominant tone sandhi in disyllabic words of a male speaker of the Chinese Oūjiāng 甌江 Wu吳 dialect of Wencheng 文成. His seven isolation tones show typical Wu complexity, comprising two mid-level, two rising, two falling-rising and one depressed level pitch shapes. Typical too is his three-way voicing contrast in syllable-Onset stops. However, the typical Wu relationship between tonal register and phonemic Onset voicing is shown to be disrupted, Onset voicing no longer correlating with tonal pitch height. The word-final tones in sandhi are shown to be straightforwardly related, phonologically and phonetically, to the isolation tones, with biuniqueness preserved. The realization of the word-initial tones in sandhi, on the other hand, involves complex mergers conditioned by largely non-phonetic factors related to historical tone categories, resulting in five extra sandhi tones that do not occur in isolation. It is...
A large-scale forensic discrimination experiment is described that investigates how well same-spe... more A large-scale forensic discrimination experiment is described that investigates how well same-speaker speech samples can be discriminated from different-speaker speech samples using acoustic parameters from Australian English vowels. A multivariate likelihood ratio is used as a discriminant function on the five tense and six lax vowel phonemes of 171 male speakers. In 171 target trials and 58,140 non-target trials, comparing samples with just one token per vowel each gives EERs of between 17% and 40%, which drop to between 10% and 14% when fused. Kernel density modeling of the reference population is shown to outperform normal, and performance degrades under quasi-realistic conditions.
Impressionistic and acoustic data are presented for the nine citation tones, and a small part of ... more Impressionistic and acoustic data are presented for the nine citation tones, and a small part of the disyllabic tone sandhi, of a speaker of the previously undescribed Chinese dialect of Maodian 毛店 from the Wuzhou 婺州 subgroup of Wu 吳. The data are used to refine the typology of the apparent right-dominant tone sandhi characteristic of the southern Wu and Min area. It is shown that not all word-final tones are the same as citation tones; and that therefore preservation of word-final tones cannot be criteria for right-dominance.
Uploads
Papers by Phil Rose