The present pilot study revives an old approach to intonation and reintroduces it as a new experi... more The present pilot study revives an old approach to intonation and reintroduces it as a new experimental method: the successive drawing of perceived intonation contours. It has been shown that intonation drawings made by untrained native German listeners for sets of controlled stimulus utterances can yield valid and reliable results patterns. Additionally, intonation drawings are more straightforward than other reproduction methods and allow more detailed insights into the perception of intonation than other meaning-based 2AFC tasks. Based on the results received for two classes of meaningful intonational units -(nuclear) pitch accents and phrase-final intonation movements -it is argued that the relationship between the production and perception of intonation is characterized by a multiparametric coding that goes far beyond F0 and additionally crosses the traditional segment-prosody divide. Since the acoustic complexity can be translated into simpler perceptual patterns, phonological models of intonation are formulated more successfully and consistently at the level of perception than at the level of acoustic F0.
German knows two plateau-based phrase-final intonation con-tours: the high level plateau of the c... more German knows two plateau-based phrase-final intonation con-tours: the high level plateau of the continuation rise and the descending plateau sequence of the calling contour. They oc-cur within a narrow scaling range of only a few semitones. The paper presents production and perception evidence for a third plateau-based phrase-final intonation contour inside this narrow scaling range. The new plateau contour shows a F0 de-crease of between 1-3 st (in the form of a slightly declining plateau or a descending plateau sequence), involves additional lengthening of the vowels underneath the plateau, and occurs when resistance is futile, i.e. when speakers signal that they finally, but reluctantly, give in to a demand of the dialogue partner. Phonological implications are briefly outlined.
Taking up anecdotal evidence, our general research aim is to investigate and to quantify the phon... more Taking up anecdotal evidence, our general research aim is to investigate and to quantify the phonetic characteristics of Low German in different Northern German regions on the basis of detailed acoustic and auditory analyses. In the initial pilot study presented here, we focus on phonologically long vowels. The analyses are based on a sample of long-vowel tokens, which were produced by 18 Northern German speakers in spontaneous translations of the 'Wenker' sentences. The speakers had comparable dialectal competences, but came from different regions of Northern Germany, i. e. Schleswig, Holstein, Dithmarschen, Ostfriedland, Mecklenburg-Vorpommern, or Nordbrandenburg. The selected vowel tokens occurred in contexts that were phonetically controlled in terms of both consonantal coarticulation and prosodic structure. The acoustic analysis included measuring durations, formant frequencies (F1-F3) and intonation characteristics (pitch-accent F0 patterns). The auditory analysis was made by narrow phonetic transcriptions. The results of the two analyses agree in showing clear regional differences. They concern the distributions of the long vowels within the vowel space (i.e. the vowel qualities, their stabilities and phonetic distances to each other) as well as the pitch-accent intonation patterns that co-occur with the long vowels. Differences in vowel duration were not found. Nordbrandenburg and Schleswig are characterized by a wide spectrum of monophthongal long-vowel qualities. In contrast, the long vowels in Holstein, Mecklenburg-Vorpommern and Ostfriesland show less diverse quality differences in the vowel centre. However, overall the phonetic diversity is still there in terms of diphthongizations that start or end at very different qualities. The long vowels of Dithmarschen combine the two characteristics, i.e. diphthongal qualities with relatively large differences in the vowel centre. As regards the intonation patterns across the vowels, we found that Dithmarschen and Holstein are both characterized by rising-falling pitch-accent peaks, but with different alignments relative to the vowel boundaries. The pitch-accent intonations of our speakers from Ostfriesland also rose and fell across the vowel. However, unlike in all other regions they additionally showed a striking F0 shape with a long high plateau in between the rising and falling movements. Our results are discussed with regard to the current claims about the dialectal organization of Northern Germany.
The line of (ongoing) research presented in this talk was inspired by the notion of truncation of... more The line of (ongoing) research presented in this talk was inspired by the notion of truncation of falling intonations by voiceless consonants at the end of utterances. The term truncation (in contrast to compression) was originally introduced for Swedish by Erikson and Altermark (1972) and refined by Bannert and Bredvad-Jensen (1975) in order to account for variation in word-accent realizations due to changes of vowel duration. Later, Grabe (1998) applied the concept of truncation to German, claiming that utterance-final falling intonations are truncated by voiceless segments, whereas utterance-final rising intonations are compressed and hence realized entirely before the voiceless segments. In the course of his prosodic labelling of the Kiel Corpus of Spontaneous Speech, it was observed by the author that, even though utterance-final F0 falls were indeed truncated to different degrees, the spectral energy distributions of aspiration noises of plosives following the truncated F0 fal...
The paper presents two perception experiments that start from the well-established phenomenon of ... more The paper presents two perception experiments that start from the well-established phenomenon of auditory restoration and address the occurrence and nature of this phenomenon in German reduced speech. The experimental stimuli based on sound patterns that can be decoded as either disyllabic nouns or highly reduced strings of function words whose canonical reference forms differ in the number of syllables and/or phon-emes. Comparisons of these stimuli with []-like hum sounds of different durations in AX tests suggest that reduced speech is perceptually restored to fuller (though maybe not canon-ical) forms, however, rather on a syllable than on a phoneme basis. In this way, the results raise further doubts about the central role of the phoneme in spoken communication.
We show on the basis of German that prosodic patterns change in the course of a traditional sente... more We show on the basis of German that prosodic patterns change in the course of a traditional sentence-list elicitation. Two fre-quent methods are analyzed: sentence-frame and syntax-frame elicitations. While only the sentences of the sentence-frame elicitation show an increase in speaking rate, both elicitation methods cause a drastic reduction in the alignment variability of nuclear pitch-accent rises. So, the starting point for the idea of segmental anchoring, i.e. the characteristic stable alignment of L and H targets, could primarily be due to a training effect based on the continuous production of analogously constructed or identical carrier sentences. Detailed pitch-accent analyses also offer alternative interpretations for anchoring patterns. Methodologically, in order to avoid training effects in pitch-accent production, our findings suggest using the syntax-frame method and short sentence lists of 40 items or less.
The study scrutinizes the role of alignment of F0 move-ments in identifying two different pitch a... more The study scrutinizes the role of alignment of F0 move-ments in identifying two different pitch accents. Although this general issue was addressed for German, the pitch-accent contrast that was studied occurs cross-linguistically and is known as 'early' vs. 'medial' or H+L* vs. L+H*. The early pitch accent reaches the F0-peak maximum befo-re the accented-vowel onset and hence falls into the vowel, while the medial pitch accent peaks after the vowel onset. This alignment-based identification model was recently un-dermined by studies that varied the slopes and ranges of the F0 movements or the extension of the F0-peak maximum. The latter parameter is taken up in the present perception experiments. Starting from a pointed rising-falling peak ali-gned at the accented-vowel onset, a peak and a plateau series were resynthesized by shifting either the entire peak or just the rising or falling movement into and away from the ac-cented vowel. The peak and plateau stimuli were...
The results of a production experiment with 8 native speakers of German show that the acoustic pr... more The results of a production experiment with 8 native speakers of German show that the acoustic profiles of contrastive focus differ between read monologues, read dialogues and spontaneous dialogues. The differences suggest that contrastive focus is only a matter of F0, and that two F0-related subtypes, comparative and corrective focus, may be distinguished. The additional changes in duration and intensity that were found for corrective focus are not due to establishing an informationstructure contrast, but to signaling emphatic reinforcement.
Based on German production data from the ‘Kiel Corpus of Spontaneous Speech’, we conducted two pe... more Based on German production data from the ‘Kiel Corpus of Spontaneous Speech’, we conducted two perception experiments, using an innovative interactive task in which participants gave real oral responses to resynthesized question stimuli. Differences in the time interval between stimulus question and response show that segmental reduction, intensity level, and the shape of the phrase-final rise all function as cues to turn-taking in conversation. Thus, the phonetics of turn-taking goes beyond the traditional triad of duration, voice quality, and F0 level.
The paper deals with perceived speech rhythm, starting from the observation that two nouns with a... more The paper deals with perceived speech rhythm, starting from the observation that two nouns with a conjunction in between ('X and/or Y', cf. title) sound more rhythmical in a particular noun order. A perception experiment on German with real and pseudo nouns provides evidence that speech rhythm is not just created prosodically by means of high and low or long and short syllables, but that the phonetic properties of the vowel nuclei and of the consonantal onsets and offsets of the stressed syllables are separate segmental constituents of speech rhythm.
Our paper presents an initial investigation into the acoustic form and function of emphatic redup... more Our paper presents an initial investigation into the acoustic form and function of emphatic reduplication in German. We found that emphatic reduplications are used to attract attention in order to prepare listeners for important information. Readspeech data elicited on this basis showed that the emphatic reduplications of grade particles and verbal imperatives both start with a prosodic 'ritardando', but differ subsequently in the overall intonation patterns and their association with either decreasing or increasing stress and accent levels.
The presented study concerns the influence of the syllabic structure on perceived prominence. We ... more The presented study concerns the influence of the syllabic structure on perceived prominence. We examined how gaps in the F0 contour due to unvoiced consonants affect prominence perception, given that such gaps can either be filled or blinded out by listeners. For this purpose we created a stimulus set of real disyllabic words which differed in the quantity of the vo-wel of the accented syllable nucleus and the types of sub-sequent intervocalic consonant(s). Results include, inter alia, that stimuli with unvoiced gaps in the F0 contour are indeed perceived as less prominent. The prominence reduction is smaller for monotonous stimuli than for stimuli with F0 excur-sions across the accented syllable. Moreover, in combination with F0 excursions, it also mattered whether F0 had to be interpolated or extrapolated, and whether or not the gap in-cluded a fricative sound. The results support both the filling-in and blinding-out of F0 gaps, which fits in well with earlier experiments on the ...
This paper continues our work on the perception of prominence as a function of F0 continuity. In ... more This paper continues our work on the perception of prominence as a function of F0 continuity. In an earlier study the first author had shown that F0 intervals occurring at lexically stressed syllables – and measured using the amplitude of Fujisaki model accent commands – strongly contribute to the perceived prominence of that syllable. More recent work explored how F0 continuity influenced prominence ratings of single word utterances. The outcome indicated that listeners made use of the physically available F0 information and there-fore words containing gaps in the contour were perceived as less prominent. It was also shown that subjects were able to interpolate missing parts as long as the F0 peak was still pre-sent. The current study explores whether subjects compensate the lack of prominence in words containing F0 gaps by asking them to produce a word with the same accent strength as that of a spoken word stimulus, the spoken word being either the same or different from the one t...
Uploads
Papers by Oliver Niebuhr