| 1. |
- Ambrazaitis, Gilbert, 1979-, et al.
(författare)
-
Accentual falls and rises vary as a function of accompanying head and eyebrow movements
- 2018
-
Ingår i: Proceedings FONETIK 2018. - Gothenburg : University of Gothenburg. ; , s. 5-7
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- In this study we examine prosodic prominence from a multimodal perspective. Our research question is whether the phonetic realization of accentual falls and rises in Swedish complex pitch accents varies as a function of accompanying head and eyebrow movements. The study is based on audio and video data from 60 brief news readings from Swedish Television (SVT Rapport), comprising 1936 words in total, or about 12 minutes of speech from five news anchors (two female, three male). The results suggest a tendency for a cumulative relation of verbal and visual prominence cues: the more visual cues accompanying, the higher the pitch peaks and the larger the rises and falls.
|
|
| 2. |
- Ambrazaitis, Gilbert, 1979-, et al.
(författare)
-
Auditory vs. audiovisual prominence ratings of speech involving spontaneously produced head movements
- 2022
-
Ingår i: Proceedings of the 11th International Conference on Speech Prosody : Speech Prosody 2022 - Speech Prosody 2022. - : International Speech Communication Association. - 2333-2042. ; , s. 352-356, s. 352-356
-
Konferensbidrag (refereegranskat)abstract
- Visual information can be integrated in prominence perception, but most available evidence stems from controlled experimental settings, often involving synthetic stimuli. The present study provides evidence from spontaneously produced head gestures that occurred in Swedish television news readings. Sixteen short clips (containing 218 words in total) were rated for word prominence by 85 adult volunteers in a between-subjects design (44 in an audio-visual vs. 41 in an audio-only condition) using a web-based rating task. As an initial test of overall rating behavior, average prominence across all 218 words was compared between the two conditions, revealing no significant difference. In a second step, we compared normalized prominence ratings between the two conditions for all 218 words individually. These results displayed significant (or near significant, p
|
|
| 3. |
- Ambrazaitis, Gilbert, 1979-, et al.
(författare)
-
Compounds in interaction : patterns of synchronization between manual gestures and lexically stressed syllables in spontaneous Swedish
- 2020
-
Ingår i: Proceedings of Gesture and Speech in Interaction (GESPIN2020). - : KTH Royal Institute of Technology.
-
Konferensbidrag (refereegranskat)abstract
- Prosody and gesture share common functions in communication, one of which is the signaling of prominence. However, we still know relatively little about how these two domains are coordinated with one another. The current study explores timing relationships between hand gestures and stressed syllables in Swedish compounds. Compounds in Swedish usually have two stressed syllables, enabling us to investigate possible differences in alignment between gestures and stressed syllables as a function of their phonological status (primary, secondary stress). We find a tendency for stressed syllables to be accompanied by movement phases of gestures, with the primary stressed syllable somewhat more likely to arise with a movement phase than the secondary stressed syllable. The stressed syllables also show tendencies for close temporal alignment with transitions between one gesture phase to another, though not all gesture phase types may start in alignment with the second stressed syllable. Our data thus provide additional support for the common function of prosody and gesture in signaling prominence in spoken communication and indicate that transitions between gesture phases as well as the gestures themselves may contribute to prominence marking.
|
|
| 4. |
|
|
| 5. |
- Ambrazaitis, Gilbert, 1979-, et al.
(författare)
-
Do dialect-specific prosodic properties shape the path to contrastive focus? - Production and comprehension data from 3-5 year-old children acquiring Stockholm or Scania Swedish
- 2023
-
Konferensbidrag (refereegranskat)abstract
- Cuing information structure (IS) is a fundamental function of prosody in many languages. For instance, in English or Dutch, a contrastive focus on the color adjective in a phrase like ‘the green boat’ is marked by a pitch accent on GREEN and, crucially, a lack of post-focal accent on BOAT: ‘the GREEN boat’. Listeners infer from this prosodic structure that ‘boat’ represents already activated (given) information and that the color that is specially mentioned represents one out of a set of alternatives.According to research findings, children achieve proficiency in the production of prosodic IS encoding within the age range of 4 to 8 years, displaying considerable variability. This variability is suggested to result from structural prosodic differences between languages. For instance, Stockholm Swedish speaking children mark (non-contrastive) focus using the Swedish prominence cuing H(igh) tone in an adult-like manner already at 4-5 years, while Dutch speaking children handle Dutch intonational pitch accents only after the age of 7-8 years. One hypothesis is that this relates to the presence of lexical pitch accents in Swedish, which could make Swedish speaking children more sensitive to prosodic contrasts; in addition, the combination of lexical accent + prominence H results in a complex contour which is particularly salient. However, studies investigating this have usually had a strict focus on speech production.The few previous studies that have conducted parallel production and comprehension experiments have typically used offline methods to assess comprehension. More recent studies using online methods such as eye tracking have usually not included children younger than 6 years of age and have not been complemented by production data. In this study we combine production and comprehension experiments, using eye tracking, to study contrastive focus prosody in 3- to 5-year-old children speaking either Scanian or Stockholm Swedish. In Scanian, instead of adding the prominence H-tone for focus, phrase-level prominence is encoded through phonetic adjustments of the (lexical) HL accent patterns. By comparing these two Swedish varieties we can thus control for phonological features (incl. lexical tone), as well as grammar and lexicon, when exploring effects of prosodic-typological differences.In our production experiment we elicit adjective-noun phrases in three different focus conditions (broad, contrast on adjective, contrast on noun), using an interactive video/card game. Production data are analyzed acoustically and auditorily.As for comprehension, our visual-word eye-tracking experiment makes use of the same pictures of colored objects to investigate whether and how children rely on prosody for reference resolution (e.g., Where is the yellow boat? And where is the GREEN boat?). The time course of eye movements will be analyzed using growth curves. Both production and eye-tracking data will be analyzed as a function of dialect, age and standardized measures of language production and comprehension (The New Reynell Developmental Language Scales), as well as compared to data from adult controls. Data are currently being collected. A preliminary analysis of eye-tracking data from 24 Scanian children (ages 3-5 years collapsed) and a subset of adults from both dialects suggests similar comprehension of focus prosody as in adults (as a mismatched focus prosody in the adjective successfully elicits looks at the foil item in all groups), although processing appears to be slower, and anticipatory strategies differ slightly from those of adults (as the color of the first-mentioned adjective in a trial elicits looks at the color-matched distractor in adults, but not in children). An analysis contrasting comprehension data for both dialects, as well as a preliminary analysis of production data will be presented at the conference.
|
|
| 6. |
- Ambrazaitis, Gilbert, 1979-, et al.
(författare)
-
Focal F0 peak shape and sentence mode in Swedish
- 2015
-
Ingår i: Proceedings of the 18th International Congress of Phonetic Sciences. - Glasgow : University of Glasgow. - 9780852619414 - 9780852619421
-
Konferensbidrag (refereegranskat)abstract
- Shape characteristics of rising-falling accentual F0 peaks of Stockholm Swedish Accent I words in narrow focus are studied in a corpus of 287 read sentences. The corpus includes statements and three types of polar questions. Results reveal a clear effect of sentence mode on the shape of the accentual rises: Statements are predominantly characterized by convex rises, questions by concave rises.
|
|
| 7. |
- Ambrazaitis, Gilbert, 1979-, et al.
(författare)
-
Is duration a (true) correlate of the Swedish word accents?
- 2021
-
Ingår i: Fonologi i Norden. - : University of Helsinki. ; , s. 2-2
-
Konferensbidrag (refereegranskat)abstract
- Swedish exhibits a binary tonal word-accent distinction (Accent 1, Accent 2) which is acoustically manifested in the fundamental frequency (fo) contour. Further acoustic correlates beyond fo have hardly been acknowledged in the past, although minor differences in segmental durations between A1 and A2 have been observed in some previous studies (e.g., Elert, 1964; Lundmark Svensson et al., 2017). In this study, we aim to scrutinize these durational patterns for two Swedish dialects (Scania and Stockholm Swedish) by means of testing whether duration as a correlate of Swedish word accents can be observed independently of effects such as focal lengthening and final lengthening. The few previously reported duration data for A1 and A2 in Swedish are, we will argue, not well in line with the numerous available reports on tone languages, which suggest that more complex tonal patterns are reflected by longer durations (e.g., K hnlein, 2015, and references therein). For Swedish, however, previous research has not consequently revealed longer durations where it would be predicted based on tonal complexity. For instance, for Stockholm Swedish, Elert (1964) observed a longer stressed vowel in A1 than in A2, although longer durations could be predicted for the more complex H*LH-pattern in A2 than for the (H)L*H in A1. One possible explanation of this inconsistency could be that the Swedish word accents per se might not significantly differ in duration at all, and the small durational differences between A1 and A2 are, instead, a bi-product of sentence-level prosody, which might, for reasons to be discussed at the conference, affect one of the word accents stronger than the other. In this study, we therefore compare segmental durations for A1 and A2 in focal and non-focal conditions in different positions in the utterance. Furthermore, we present two parallel analyses for two dialects of Swedish – Stockholm and Scania – which differ critically in the tonal composition (or rather timing) of the word accents, as well as in how focus is marked tonally.The analyses are based on recordings from 24 speakers in total (6 women and 6 men per dialect), and 36 utterances per speaker (12 conditions, 3 repetitions). Speakers were asked to read variants of the sentence Boven/ar hade vinet/er i bilen/ar ‘(The) villain(s) had wine(s) in (the) car(s)’, where the three test words (villain, wine, car) are associated either with A1 (= def. sg. form) or A2 (= indef. pl.); a few relevant combinations of A1 and A2 forms were selected for this study. Furthermore, a narrow focus was elicited (using context questions) on either the first, second or third test word. Data are analyzed using linear mixed regression models. The results reveal an overall stable and uniform effect of word accent on the duration of the vowel in the stressed syllable (/i:/ in vin and bil) and the subsequent consonant (either /n/ or /l/), irrespective of focus condition, position in utterance (vin vs. bil) and dialect: both segments tend to be slightly longer in A2 than in A1 (except in the /l/ in post-focal bilen in Stockholm Swedish).To conclude, this study suggests duration as a systematic correlate of the Swedish word accents, which, however, does not obviously seem to relate to tonal composition or complexity. We aim to collect fruitful comments and explanatory accounts from the conference delegates.ReferencesElert, C.-C. (1964). Phonologic Studies of Quantity in Swedish. Based on Material from Stockholm Speakers. Stockholm: Almqvist & Wiksell.Köhnlein, B. (2015). The complex durational relationship of contour tones and level tones: Evidence from diachrony. Diachronica, 32(2), 231-267. doi:10.1075/dia.32.2.03kohLundmark Svensson, M., Ambrazaitis, G., & Ewald, O. (2017). Exploring multidimensionality: Acoustic and articulatory correlates of Swedish word accents. Proc. INTERSPEECH 2017, Stockholm, Sweden. http://dx.doi.org/10.21437/Interspeech.2017-1502
|
|
| 8. |
- Ambrazaitis, Gilbert, 1979-, et al.
(författare)
-
Is phonetic prominence underlyingly multimodal?
- 2019
-
Ingår i: Phonetics and Phonology in Europe 2019 satellite workshop. - Köln : University of Cologne.
-
Konferensbidrag (refereegranskat)
|
|
| 9. |
|
|
| 10. |
- Ambrazaitis, Gilbert, 1979-, et al.
(författare)
-
Multimodal prominence ratings : Effects of screen size and audio device
- 2019
-
Ingår i: Book of Abstracts MMSYM 2019. - : University of Leuven. ; , s. 2-3
-
Konferensbidrag (refereegranskat)abstract
- Prosodic prominence is a multimodal phenomenon involving both acoustic and kinematic dimensions. In order to study the multimodal nature of prominence, we need to collect prominence ratings based on audio-visual speech material from large groups of speakers. This is feasible by means of a web-based crowdsourcing set-up, allowing volunteers to participate using a private computer or mobile phone. However, this freedom also implies a certain reduction of experimental control due to variation in hardware used by the raters. In this pilot study we explore potential effects of two hardware features – screen size and audio device (headphones vs. loudspeakers) – on multimodal prominence ratings. To this end, 16 brief clips from Swedish television news (218 words in total) were rated by 31 native Swedish volunteers using a web-based set-up. In our GUI, orthographic representations of the text were displayed below the video player. Each word was to be rated as either non-prominent, moderately prominent, or strongly prominent, by means of clicking on the word in question until the desired prominence level was encoded through a specific color (yellow: moderate; red: strong). Participants were free to use a mobile phone, a tablet, or a computer, and headphones or loudspeakers, and we collected information about their hardware using a questionnaire. In addition, we automatically logged the screen size of the participant’s computer/phone. We applied two different approaches to analyze the participant’s rating behavior as a function of the hardware features under discussion. First, we calculated a selection of five variables from the raw prominence ratings: (i) the sum of all ratings (over all 218 words), (ii) the percentage of words rated as (moderately or strongly) prominent, (iii) among prominent words, the proportion of words rated as strongly prominent, and (iv-v) the relative prominence rating of two selected words. Effects of screen size and audio device on these variables were analyzed using linear regression models. Second, we calculated inter-rater reliability for multiple raters using Fleiss’ kappa, both for all raters as a reference and for subgroups concerning audio device and screen size. The results reveal a significant model fit for variable (iii) defined above (proportion of strong ratings; F(5;21) = 5.332; p=.0022**), suggesting a significantly higher proportion of strong prominent ratings obtained with loudspeakers (34.0% of words rated as prominent on average) compared to with headphones (18.3%; t=2.944; p=.0073**), as well as with medium size screens (34.2%) compared to with small screens (24.4%; t=2.433; p=.0232*); however, the proportion of strong prominent ratings tended to be lowest with large screens (14.2% on average). Effects of screen size were also reflected in inter-rater reliability, revealing the highest kappa for users with medium-sized screens (kappa=.566, when ratings are recoded to a binary decision) compared to large (kappa=.485) and small screens (mobile phones, kappa=.437). However, inter-rater reliability was less affected by the listening condition (headphones vs. loudspeakers). 3 To conclude, the choice of hardware might have effects on multimodal prominence ratings, which has to be considered in crowdsourcing approaches. More detailed results will be presented at the conference.
|
|
