Search

A speech production experiment with electroglottography investigated how voicing is affected by consonants of differing degrees of constriction. Measures of glottal contact [closed quotient (CQ)] and strength of voicing [strength of excitation (SoE)] were used in conditional inference tree analyses. Broadly, the results show that as the degree of constriction increases, both CQ and SoE values decrease, indicating breathier and weaker voicing. Similar changes in voicing quality are observed throughout the course of the production of a given segment. Implications of these results for a greater understanding of source-tract interactions and for the phonological notion of sonority are discussed.

“Creaky voice” is a term that covers multiple kinds of voicing, and there is no single defining acoustic property shared by all subtypes of creaky voice. Here we explore the distinct characteristics of each subtype. We identify three main properties of creaky voice: low f0, irregular f0, and constricted glottis (as shown by electroglottography). Prototypical creaky voice has all of these properties; other subtypes are characterized by different subsets of properties. We will describe, with reference to previous literature, multiply-pulsed creak (a special case of irregular f0, along with low f0), open-glottis creak (low and often irregular f0, but unconstricted glottis), vocal fry (low but regular f0, constricted glottis), and creak with such irregular pulsing that no f0 can be recovered. Building on our previous work [Keating et al. 2015 Proc. ICPhS], we show how various acoustic measures pattern for each subtype. Results from parametric speech synthesis provide support for our acoustic observations.

The IPA currently does not specify how to represent prenasalization, preglottalization or preaspiration. We first review some current transcription practices, and phonetic and phonological literature bearing on the unitary status of prenasalized, preglottalized and preaspirated segments. We then propose that the IPA adopt superscript diacritics placed before a base symbol for these three phenomena. We also suggest how the current IPA Diacritics chart can be modified to allow these diacritics to be fit within the chart.

Little is known about human and machine speaker discrimination ability when utterances are very short and the speaking style is variable. This study compares text-independent speaker discrimination ability of humans and machines based on utterances shorter than 2 s in two different speaking styles (read sentences and speech directed towards pets, characterized by exaggerated prosody). Recordings of 50 female speakers drawn from the UCLA Speaker Variability Database were used as stimuli. Performance of 65 human listeners was compared to i-vector-based automatic speaker verification systems using mel-frequency cepstral coefficients, voice quality features, which were inspired by a psychoacoustic model of voice perception, or their combination by score-level fusion. Humans always outperformed machines, except in the case of style-mismatched pairs from perceptually-marked speakers. Speaker representations by humans and machines were compared using multi-dimensional scaling (MDS). Canonical correlation analysis showed a weak correlation between machine and human MDS spaces. Multiple regression showed that means of voice quality features could represent the most important human MDS dimension well, but not the dimensions from machines. These results suggest that speaker representations by humans and machines are different, and machine performance might be improved by better understanding how different acoustic features relate to perceived speaker identity.Little is known about human and machine speaker discrimination ability when utterances are very short and the speaking style is variable. This study compares text-independent speaker discrimination ability of humans and machines based on utterances shorter than 2 s in two different speaking styles (read sentences and speech directed towards pets, characterized by exaggerated prosody). Recordings of 50 female speakers drawn from the UCLA Speaker Variability Database were used as stimuli. Performance of 65 human listeners was compared to i-vector-based automatic speaker verification systems using mel-frequency cepstral coefficients, voice quality features, which were inspired by a psychoacoustic model of voice perception, or their combination by score-level fusion. Humans always outperformed machines, except in the case of style-mismatched pairs from perceptually-marked speakers. Speaker representations by humans and machines were compared using multi-dimensional scaling (MDS). Canonical correlation analysi...

Little is known about the nature or extent of everyday variability in voice quality. This paper describes a series of principal component analyses to explore within- and between-talker acoustic variation and the extent to which they conform to expectations derived from current models of voice perception. Based on studies of faces and cognitive models of speaker recognition, the authors hypothesized that a few measures would be important across speakers, but that much of within-speaker variability would be idiosyncratic. Analyses used multiple sentence productions from 50 female and 50 male speakers of English, recorded over three days. Twenty-six acoustic variables from a psychoacoustic model of voice quality were measured every 5 ms on vowels and approximants. Across speakers the balance between higher harmonic amplitudes and inharmonic energy in the voice accounted for the most variance (females = 20%, males = 22%). Formant frequencies and their variability accounted for an additional 12% of variance across speakers. Remaining variance appeared largely idiosyncratic, suggesting that the speaker-specific voice space is different for different people. Results further showed that voice spaces for individuals and for the population of talkers have very similar acoustic structures. Implications for prototype models of voice perception and recognition are discussed.

Speakers of North American English are known to use a variety of tap/flap articulations depending on phonetic context (Derrick and Gick, 2011); it is also known that NAE taps/flaps are sometimes associated with a greatly lowered F4 frequency (Warner and Tucker, 2017). It has been less clear whether only certain articulatory variants show this acoustic effect. Since retroflex stops are also associated with lowered F4 (Blumstein and Stevens, 1975), we predict that flap retroflexion is associated with lowered F4. To test this prediction, synchronized ultrasound and audio recordings were made of words containing /t, d/ in a variety of contexts known to give rise to tap/flap variants. Based on visual inspection of ultrasound videos, these were coded as one of four articulatory variants (low tap, high tap, up flap, down flap: Derrick and Gick, 2011); formant frequencies were extracted from the audio at several timepoints relative to the tap/flap. Preliminary results from one speaker support the hypothesis: high taps and down flaps (variants with initial retroflexion) show an F4 drop into the consonant, while high taps and up flaps (variants with final retroflexion) show an F4 rise out of the consonant.

This study investigates the importance of source spectrum slopes in the perception of phonation by White Hmong listeners. In White Hmong, nonmodal phonation (breathy or creaky voice) accompanies certain lexical tones, but its importance in tonal contrasts is unclear. In this study, native listeners participated in two perceptual tasks, in which they were asked to identify the word they heard. In the first task, participants heard natural stimuli with manipulated F0 and duration (phonation unchanged). Results indicate that phonation is important in identifying the breathy tone, but not the creaky tone. Thus, breathiness can be viewed as contrastive in White Hmong. Next, to understand which parts of the source spectrum listeners use to perceive contrastive breathy phonation, source spectrum slopes were manipulated in the second task to create stimuli ranging from modal to breathy sounding, with F0 held constant. Results indicate that changes in H1-H2 (difference in amplitude between the first and second harmonics) and H2-H4 (difference in amplitude between the second and fourth harmonics) are independently important for distinguishing breathy from modal phonation, consistent with the view that the percept of breathiness is influenced by a steep drop in harmonic energy in the lower frequencies.

How are listeners able to identify whether the pitch of a brief isolated sample of an unknown voice is high or low in the overall pitch range of that speaker? Does the speaker's voice quality convey crucial information about pitch level? Results and statistical models of two experiments that provide answers to these questions are presented. First, listeners rated the pitch levels of vowels taken over the full pitch ranges of male and female speakers. The absolute f0 of the samples was by far the most important determinant of listeners' ratings, but with some effect of the sex of the speaker. Acoustic measures of voice quality had only a very small effect on these ratings. This result suggests that listeners have expectations about f0s for average speakers of each sex, and judge voice samples against such expectations. Second, listeners judged speaker sex for the same speech samples. Again, absolute f0 was the most important determinant of listeners' judgments, but now voice quality measures also played a role. Thus it seems that pitch level judgments depend on voice quality mostly indirectly, through its information about sex. Absolute f0 is the most important information for deciding both pitch level and speaker sex.

The Council of the IPA is establishing committees that will take on some of the tasks of running the Association, and will allow us to expand the kinds of projects that we can undertake.

San Felipe Jalapa de Díaz (Jalapa) Mazatec is unusual in possessing a three-way phonation contrast and three-way level tone contrast independent of phonation. This study investigates the acoustics of how phonation and tone interact in this language, and how such interactions are maintained across variables like speaker sex, vowel timecourse, and presence of aspiration in the onset. Using a large number of words from the recordings of Mazatec made by Paul Kirk and Peter Ladefoged in the 1980s and 1990s, the results of our acoustic and statistical analysis support the claim that spectral measures like H1-H2 and mid-range spectral measures like H1-A2 best distinguish each phonation type, though other measures like Cepstral Peak Prominence are important as well. This is true regardless of tone and speaker sex. The phonation type contrasts are strongest in the first third of the vowel and then weaken towards the end. Although the tone categories remain distinct from one another in terms of F0 throughout the vowel, for laryngealized phonation the tone contrast in F0 is partially lost in the initial third. Consistent with phonological work on languages that cross-classify tone and phonation type (i.e. ‘laryngeally complex’ languages, Silverman 1997), this study shows that the complex orthogonal three-way phonation and tone contrasts do remain acoustically distinct according to the measures studied, despite partial neutralizations in any given measure.

This study investigates effects of three prosodic factors—prosodic boundary (Utterance-initial vs. Utterance-medial), lexical stress (primary vs. secondary) and phrasal accent (accented vs. unaccented)—on articulatory and acoustic realizations of word-initial CVs (/ne/, /te/) in trisyllabic English words. The consonantal measures were linguopalatal Peak contact and Release contacts (by electropalatography), Seal duration, Nasal duration and Nasal energy for /n/, VOT, RMS burst energy and spectral Center of Gravity at the release for /t/; and the vocalic measures were linguopalatal Vowel contact, Vowel F1, Vowel duration and Vowel amplitude. Several specific points emerge. Firstly, domain-initial articulation is differentiated from stress- or accent-induced articulations along several measures. Secondly, the vowel is effectively louder domain-initially, suggesting that the boundary effect is not strictly local to the initial consonant. Thirdly, some accentual effects can be seen in secondary-stressed syllables, suggesting that accentual influences spread beyond the primary-stressed syllable. Finally, unlike domain-initial effects, prominence effects are not cumulative. Thus we conclude that, at least for the kind of word-initial syllables tested here, different aspects of prosodic structure (domain boundary vs. prominence) are differentially encoded.

In a study of optical cues to the visual perception of stress, three American English talkers spoke words that differed in lexical stress and sentences that differed in phrasal stress, while video and movements of the face were recorded. The production of stressed and unstressed syllables from these utterances was analyzed along many measures of facial movement, which were generally larger and faster in the stressed condition. In a visual perception experiment, 16 perceivers identified the location of stress in forced-choice judgments of video clips of these utterances (without audio). Phrasal stress was better perceived than lexical stress. The relation of the visual intelligibility of the prosody of these utterances to the optical characteristics of their production was analyzed to determine which cues are associated with successful visual perception. While most optical measures were correlated with perception performance, chin measures, especially Chin Opening Displacement, contributed the most to correct perception independently of the other measures. Thus, our results indicate that the information for visual stress perception is mainly associated with mouth opening movements.

A complete understanding of visual phonetic perception (lipreading) requires linking perceptual effects to physical stimulus properties. However, the talking face is a highly complex stimulus, affording innumerable possible physical measurements. In the search for isomorphism between stimulus properties and phonetic effects, second-order isomorphism was examined between the perceptual similarities of video-recorded perceptually identified speech syllables and the physical similarities among the stimuli. Four talkers produced the stimulus syllables comprising 23 initial consonants followed by one of three vowels. Six normal-hearing participants identified the syllables in a visual-only condition. Perceptual stimulus dissimilarity was quantified using the Euclidean distances between stimuli in perceptual spaces obtained via multidimensional scaling. Physical stimulus dissimilarity was quantified using face points recorded in three dimensions by an optical motion capture system. The variance accounted for in the relationship between the perceptual and the physical dissimilarities was evaluated using both the raw dissimilarities and the weighted dissimilarities. With weighting and the full set of 3-D optical data, the variance accounted for ranged between 46% and 66% across talkers and between 49% and 64% across vowels. The robust second-order relationship between the sparse 3-D point representation of visible speech and the perceptual effects suggests that the 3-D point representation is a viable basis for controlled studies of first-order relationships between visual phonetic perception and physical stimulus attributes.

Little is known about how to characterize normal variability in voice quality within and across utterances from normal speakers. Our previous study of female voices suggested that only a few acoustic parameters consistently distinguish speakers, with most of the work being done by idiosyncratic subsets of parameters. The present study extends this research to samples of 50 men’s voices. The men read 5 sentences twice on 3 days—30 sentences per speaker. The VoiceSauce analysis program estimated means and standard deviations for many acoustic parameters (including F0, harmonic amplitude differences, harmonic-to-noise ratios, and formant frequencies) for the vowels and approximant consonants in each sentence. Multidimensional scaling and linear discriminant analysis were used to examine the acoustic characteristics of the overall voice space, and to measure how well each speaker’s set of 30 sentences could be acoustically distinguished from all other speakers’ sentences. Additional analyses of small subsets of voices compared the importance of global versus local details in discriminating voices acoustically. Results will be compared to those for female speakers, and implications for recognition by listening will be discussed. [Work supported by NSF and NIH.]Little is known about how to characterize normal variability in voice quality within and across utterances from normal speakers. Our previous study of female voices suggested that only a few acoustic parameters consistently distinguish speakers, with most of the work being done by idiosyncratic subsets of parameters. The present study extends this research to samples of 50 men’s voices. The men read 5 sentences twice on 3 days—30 sentences per speaker. The VoiceSauce analysis program estimated means and standard deviations for many acoustic parameters (including F0, harmonic amplitude differences, harmonic-to-noise ratios, and formant frequencies) for the vowels and approximant consonants in each sentence. Multidimensional scaling and linear discriminant analysis were used to examine the acoustic characteristics of the overall voice space, and to measure how well each speaker’s set of 30 sentences could be acoustically distinguished from all other speakers’ sentences. Additional analyses of small subsets ...

© 2014 Acoustical Society of America. This study explores the vocal fold contact patterns of one type of phonation contrast - the tense vs lax phonation contrasts of three Yi (Loloish) languages. These contrasts are interesting because neither phonation category is very different from modal voice, and because both phonations are largely independent of the languages' tonal contrasts. Electroglottographic (EGG) recordings were made in the field, and traditional EGG measures were derived. These showed many small but significant differences between the phonations, with tense phonation having greater contact quotients and briefer but slower changes in contact. Functional data analysis was then applied to entire EGG pulse shapes. The resulting first principal component was found to be mostly strongly related to the phonation contrasts, and correlated with almost all the traditional EGG measures. Unlike the traditional measures, however, this component also seems to capture differences in abruptness of contact. Furthermore, previously collected perceptual responses from native speakers of one of the languages correlated better with this component than with any other EGG measure or any acoustic measure. The differences between these tense and lax phonations are not large, but apparently they are consistent enough, and perceptually robust enough, to support this linguistic contrast.

The study investigates the importance of phonation cues in White Hmong tone identification.

There is a consensus that dyslexia is on a continuum with normal reading skill and that dyslexics fall at the low end of the normal range in phonological skills. However, there is still substantial variability in phonological skill among dyslexic children. Recent studies have focused on the high end of the continuum of phonological skills in dyslexics, identifying a “surface” dyslexic, or “delayed” profile in which phonological skills are not out of line with other aspects of word recognition. The present study extended this work to a longitudinal context, and explored differences among subgroups of dyslexics on a battery of component reading skills. Third grade dyslexics (n=72) were classified into two subgroups, phonological dyslexics and delayed dyslexics, based on comparisons to younger normal readers at the same reading level (RL group). The children were tested at two points (in third and fourth grade). The results revealed that the classification of dyslexics produced reliable, stable, and valid groups. About 82 percent of the children remained in the same subgroup category when retested a year later. Phonological dyslexics were lower in phoneme awareness and expressive language. Delayed dyslexics tended to be slower at processing printed letters and words but not at rapid automatic naming of letters, and relied more heavily on phonological recoding in reading for meaning than did phonological dyslexics. A subset of the delayed dyslexics with the traditional “surface dyslexic” pattern (relatively high pseudoword and low exception word reading) was also identified. The surface subgroup resembled the RL group on most measures and was not very stable over one year. The results are discussed in light of current models of dyslexia and recent subgrouping schemes, including the Double-Deficit Hypothesis.

In March–April 2015, ten additional members of the new Council of the International Phonetic Association were elected. They are: Plinio Barbosa, Martine Grice, Asher Laufer, Wai-Sum Lee, Kofi Adu Manyah, Alexis Michaud, Jane Stuart-Smith, Marija Tabain, Masaki Taniguchi, and Jacqueline Vaissière. John Ohala, elected in the previous round, has resigned from the Council.

In this paper it is shown that at the edges of prosodic domains, initial consonant and final vowels have more extreme (less reduced) lingual articulations, which are called articulatory strengthening. Linguopalatal contact for consonants and vowels in different prosodic positions was compared, using reiterant-speech versions of sentences with a variety of phrasings read by three speakers of American English. Four prosodic domains were considered: the phonological word, the phonological (or intermediate) phrase, the intonational phrase, and the utterance. Domain-initial consonants show more linguopalatal contact than domain-medial or domain-final consonants, at three prosodic levels. Most vowels, on the other hand, show less linguopalatal contact in domain-final syllables compared to domain-initial and domain-medial. As a result, the articulatory difference between segments is greater around a prosodic boundary, increasing the articulatory contrast between consonant and vowels, and prosodic domains are marked at both edges. Furthermore, the consonant initial strengthening is generally cumulative, i.e., the higher the prosodic domain, the more linguopalatal contact the consonant has. However, speakers differed in how many and which levels were distinguished in this way. It is suggested that this initial strengthening could provide an alternative account for previously observed supralaryngeal declination of consonants. Acoustic duration of the consonants is also affected by prosodic position, and this lengthening is cumulative like linguopalatal contact, but the two measures are only weakly correlated.

Little is known about how to characterize normal variability in voice quality within and across utterances from normal speakers. Given a standard set of acoustic measures of voice, how similar are samples of 50 women’s voices? Fifty women, all native speakers of English, read 5 sentences twice on 3 days—30 sentences per speaker. The VoiceSauce analysis program estimated many acoustic parameters for the vowels and approximant consonants in each sentence, including F0, harmonic amplitude differences, harmonic-to-noise ratios, formant frequencies. Each sentence was then characterized by the mean and standard deviation of each measure. Linear discriminant analysis tested how well each speaker’s set of 30 sentences could be acoustically distinguished from all other speakers’ sentences. Initial work testing just 3 speakers from this sample found that the speakers could be completely discriminated (classified) by these measures, and largely discriminated by just 2 of them. Such a simple result is not expected for the larger sample of speakers. We will present results concerning how successfully speakers can be discriminated, how well different numbers of discriminant functions do, and which acoustic measures do the most work. Implications for recognition by listening will be discussed. [Work supported by NSF and NIH.]

To determine if the speaking fundamental frequency (F0) profiles of English and Mandarin differ, a variety of voice samples from male and female speakers were compared. The two languages’ F0 profiles were sometimes found to differ, but these differences depended on the particular speech samples being compared. Most notably, the physiological F0 ranges of the speakers, determined from tone sweeps, hardly differed between the two languages, indicating that the English and Mandarin speakers’ voices are comparable. Their use of F0 in single-word utterances was, however, quite different, with the Mandarin speakers having higher maximums and means, and larger ranges, even when only the Mandarin high falling tone was compared with English. In contrast, for a prose passage, the two languages were more similar, differing only in the mean F0, Mandarin again being higher. The study thus contributes to the growing literature showing that languages can differ in their F0 profile, but highlights the fact that the choice of speech materials to compare can be critical.

This paper reports a set of studies of some phonetic characteristics of the American English represented in the TIMIT speech database. First we describe some relevant characteristics of TIMIT, and how we use the non-speech files on the TIMIT CD with a commercial database program. Two studies are then described: one using only the non-audio parts of TIMIT (segmental transcriptions and durations, and speaker information), and one using the audio signal for acoustic analysis. Results of such studies should be useful not only to linguistic phoneticians but also for speech recognition lexicons and text-to-speech systems.

Within the context of a grammar containing both phonological and phonetic rules, three kinds of representation all qualify as “ phonetic”. One is the categorical output of the phonology. The others are two forms of output of the quantitative phonetics: articulatory and acoustic. These representations are discussed and exemplified primarily with respect to laryngeal features for consonants.

The integrity of phonological representation/processing in dyslexic children was explored with a gating task in which children listened to successively longer segments (gates) of a word. At each gate, the task was to decide what the entire word was. Responses were scored for overall accuracy as well as the children's sensitivity to coarticulation from the final consonant. As a group, dyslexic children were less able than normally achieving readers to detect coarticulation present in the vowel portion of the word, particularly on the most difficult items, namely those ending in a nasal sound. Hierarchical regression and path analyses indicated that phonological awareness mediated the relation of gating and general language ability to word and pseudoword reading ability.

Increasing evidence suggests that voices are best thought of as complex auditory patterns, and that listeners perceive and remember voices with reference to a “prototype” or “average” for that talker. Little is known about how, and how much, individual talkers vary their voice quality across situations that arise in every-day speaking, so the nature and extent of variability underlying these abstract averages, and thus the nature of the averages themselves, is unclear. The theoretical relationship between acoustic similarity and confusability in the context of a prototype model also remains unclear. In this preliminary study, 9 tokens of the vowel /a/ were recorded from 5 females on three dates. Measures of F0, spectral slope, HNR, and formant frequencies and their variability were gathered for all voice samples and acoustic distances between talkers were calculated under the assumption that all acoustic variables were equally important perceptually. Perceptual confusability was assessed in a same/different task, and predictions under the equal perceptual weight assumption were tested. Discussion will focus on how much variability is required before a voice sample no longer sounds like the originating talker, and on how the perceptual importance of each acoustical variable varies across talkers and acoustic contexts. [Work supported by NSF and NIH.]

While almost all of Ken Stevens’s research has been influential in linguistic phonetics—especially his work on the quantal nature of speech and on enhancement theory—this presentation will focus more on aspects not covered by others in this session. A noteworthy aspect of Ken’s career is that he frequently collaborated with linguists, notably in research on phonetic features and their structure. In work with Blumstein and with Halle, he provided acoustic correlates of place of articulation, laryngeal, and nasalization features. Much of this work is summarized in his 1980 paper in JASA, “Acoustic correlates of some phonetic categories.” In work with Keyser, he suggested an overall organization of features to define major classes of sounds. This work was published in linguistics journals, e.g., their 1994 paper in Phonology, “Feature geometry and the vocal tract.” Ken’s importance to linguistics, as an engineer interested in linguistic sound systems and eager to work with phoneticians and phonologists, cannot be overestimated, and is a legacy continued by many of his students.

During 2014, the Association granted three IPA Student Awards to student members participating in phonetics conferences. The recipients of the Awards were Martin Kohlberger of Leiden University for participation in Sound Change in Interacting Human Systems (University of California, Berkeley, USA), Caroline Sigouin of Université Laval (Québec, Canada) for participation in 30es Journées d’études sur la parole (Le Mans, France), and Valerie Freeman of the University of Washington for participation in NWAV 43 (New Ways of Analyzing Variation) in Chicago. Congratulations to these student members.