Robert Espesser

About the relationship between eyebrow movements and Fo variations

Speech production is always accompanied by facial and gestural activity. The study is part of a broader research project on how head movements and facial expressions are related to voice variations in different speech situations. Ten normal subjects were recorded while reading aloud answering yes/no questions, and dialoguing with an interviewer. Rapid rising-falling eyebrow movements produced by the subjects as they spoke were associated with Fo rises in only 71% of the cases. This suggests that eyebrow movements and fundamental frequency changes are not automatically linked (i.e., they are not the result of muscular synergy), but are more a consequence of linguistic and communicational choices. Note also that 38% of the eyebrow movements were produced while the subject was not speaking. Thus, eyebrow movements may also serve as back-channel signals or play a role in turn-taking during conversation.

Levels of Representation and Levels of Analysis for the Description of Intonation Systems

The linguistic description of the intonation systems of different languages, like that of any other aspect of language, can be thought of as a rather indirect process of extracting linguistic information from measurable physical data. As has long been known, there is no automatic technique for performing this operation. As Chomsky (1964) has pointed out, there is no general ‘discovery procedure’ we can appeal to.

Intensive training of phonological skills in progressive aphasia: A model of brain plasticity in neurodegenerative disease

Three patients with a typical syndrome of nonfluent primary progressive aphasia (Mesulams syndrome) were trained daily with a remediation protocol including auditory exercises specifically designed to involve several aspects of phonological processing, a domain known to be specifically affected in the condition. The speech content of the exercises was based on the temporal theory of phonological processes according to which increasing the duration of formant transition should facilitate phoneme discrimination and phoneomic awareness. Significantly improved performance on the trained tasks was demonstrated in the three patients. Improvement further generalized to other tasks such as nonword repetition and reading. We conclude that such results (1) argue for using intensive focused therapy of language impairment in neurodegenerative disorders, (2) may constitute a good model of brain plasticity in neurodegenerative disorders in general, and (3) support theories of phonological processing emphasizing temporal features of the auditory signal.

Use of the magnitude estimation technique for assessing the performance of text‐to‐speech synthesis systems

As text-to-speech systems develop, it becomes necessary to compare various solutions and to evaluate whether a change in the synthesis procedure has an effect on the listeners attitude to the system. The possibility of directly scaling intelligibility, naturalness, and users satisfaction (i.e., acceptability) with the magnitude estimation technique is investigated. A magnitude estimation protocol suitable for this purpose is described. In general, within the limits of the methodological constraints discussed in this paper, the procedure appears to be reliable and valid for quantifying the perceived attributes of synthesized speech.

Vowel reduction in conversational speech in French: the role of lexical factors

Abstract In this study we investigate vowel reduction and the role of some lexical factors in the production of vowels extracted from a corpus of French conversations. Vowel durations and spectral quality are examined with respect to (1) their interaction in the corpus, (2) the position of vowels in words, and (3) word frequency and word category. The analyses are conducted on vowels produced by 16 speakers. Our study provides strong evidence that vowel reduction (decrease in durations and more centralized spectral values) affects most of the vowels in conversational speech. The results show that vowels in final syllables of words were less often reduced while the preceding ones show reduced durations and centralized formant values. Moreover vowels are more reduced in monosyllabic function words than in monosyllabic content words. Nevertheless, we did not find a clear effect of word frequency on vowel durations. Finally, our study shows that vowel reduction depends on several factors related to lexical properties (word category) and to prosodic properties (stress and final lengthening).

Statistical distributions of speech for various languages

Statistical distributions of speech spectral energy were obtained for Danish, German, Italian, English, and French. The variable whose distribution was analyzed is termed “perceived spectral energy.” It represents the energy of speech contained within a band 1 Hz wide at the output of an exponential time window. Various window time constants were used in the analysis. Speech material was obtained from the EUROM. O database. The results indicate that there are no significant differences between the languages. Therefore, the normative values can be specified across the languages. However, significant sex differences are found and norms are given for each sex. Finally, these distributions will be compared to those of synthetic speech. [Work supported by a grant from the EEC Esprit SAM project.] Also at the Speech and Hearing Ctr., Univ. of Iowa, Iowa City, IA 52242.

Effet de l'accélération artificielle du signal de parole sur la perception des dysarthries cérébelleuses: à propos de deux cas

Objectif: Dans les dysarthries cérébelleuses, la perte de la précision des mouvements articulatoires de la parole induit une dysrégulation temporelle avec un allongement irrégulier des durées contribuant globalement au ralentissement du débit de la parole. Ainsi, nous supposons que les distorsions du signal de parole sont largement influencées par le ralentissement et que la correction de ce ralentissement, par accélération du signal de parole, peut améliorer l’intelligibilité. Patientes et méthode: Deux patientes avec une dysarthrie ataxique ont été étudiées. L’intelligibilité de la parole a été évaluée subjectivement avec une échelle visuelle analogique et objectivement par la méthode des stimuli constants associée à une tâche d’identification d’un phonème cible en temps réel. L’accélération temporelle des stimuli est effectuée par l’algorithme SOLA. Un total de 144 stimuli a été créé à 3 vitesses différentes: 48 phrases à «vitesse naturelle» (sans accélération, VN), 48 phrases à «vitesse intermédiaire» (50% de l’accélération maximale, VI) et 48 phrases à «vitesse maximale» (vitesse moyenne des locuteurs témoins, VM). Les 144 phrases, dans les 3 présentations expérimentales, ont été réparties en 3 listes, chaque auditeur entendant une seule fois chaque phrase. Chaque liste est écoutée par 8 auditeurs. Les phrases sont présentées dans un ordre randomisé. Pour estimer l’effet de l’accélération, nous avons utilisé un modèle logit mixte pour le score de l’épreuve subjectif et les temps de réaction, un test de χ2 pour le nombre des erreurs. Résultats: Pour les 2 locutrices, la probabilité d’une meilleure intelligibilité à l’épreuve subjective est significativement plus grande pour la condition VI ou VM que pour la condition VN. Le passage de la condition VN à la condition VI augmente le taux de satisfaction des auditeurs quant à l’intelligibilité de 78 à 87% pour Mme B et de 46 à 63% pour Mme Z. Les erreurs sont moins fréquentes avec la compression pour Mme Z. L’effet de l’accélération n’est jamais significatif sur les temps de réaction (F2, 1024 = 2,14, p = 0,12). Conclusions: Cette étude est la première à analyser l’effet d’une accélération temporelle uniforme sur de la parole pathologique. Elle montre que chez 2 patientes présentant une dysarthrie ataxique, le niveau perçu de l’intelligibilité est amélioré.

Reducing context effects in the subjective evaluation of speech quality

This study investigates whether the contextual invariance and subject invariance of categorical and magnitude estimates of speech quality could be improved by introducing a reference system and by normalizing the results with respect to it. One hundred and forty‐four subjects were equally divided into 12 experimental groups and tested in various stimulus contexts using either a magnitude estimation (ME) or categorical estimation (CE) procedures. Various normalization procedures were compared to the nonnormalized results with respect to their immunity to the context/group effects. Different listening conditions were created by using various text‐to‐speech synthesis algorithms and external noise. The reference condition was natural speech. The results indicate that the introduction of a reference system and appropriate normalization algorithms is beneficial. [Work supported by a grant from the EEC Esprit SAM project.] Also at the Speech and Hearing Center, Univ. of Iowa, Iowa City, IA 52242.

How does vowel duration affect loudness in a CV syllable

Abstract The objective of our current experiments is to examine the effect of duration on loudness in CV syllables. A natural [ta] was manipulated with the aid of a computer so as to generate a reference stimulus of 250 ms at 70 dB SPL and a series of test stimuli with duration of 75, 100, 150, 200, and 250 ms. The stimulus interval was 200 ms in experiment I, and 600 ms in experiment II. A computer-assisted method of adjustment was set up. The measured level difference in dB between the reference stimulus and the adjusted test stimulus is less than 2 dB for the shortest duration, and is equal to 0.2 dB for the longest one. In experiment III, the Up-Down-Transformed-Response procedure was used with the same stimuli as in experiment II. Resulting level differences values, when plotted, had the same loudness contour as in experiment I and II, but were slightly higher.