Austrian Academy of Sciences
Acoustics Research Institute
Secondary Stress: A Speaker-Specific
Characteristic?
Sylvia Moosmüller
http://www.kfs.oeaw.ac.at
[email protected]
Institut für Schallforschung, Österreichische Akademie der Wissenschaften
Wohllebengasse 12-14, A-1040 Wien, Austria
Tel: +43 1/4277-29503 Fax: +43 1/4277-9295 E-mail: [email protected]
Vowels in Standard Austrian German
Introducation: Theoretical Framework
•Socio-pragmatic foundation of phonology and
phonetics
•The phonological system
•Standard Austrian German
Methods and Data
Interpretation of Formant Frequencies
Vowel Inventory and Features
Coarticulation
Vowel Variability
Speaker-specific traits
Data
Six speakers of Standard Austrian German
Two speaking tasks (reading, spontaneous)
~ 11.000 vowels were analysed
F1, F2, F3, F0, duration
Method
Sampling frequency: 22 050 Hz, 16 Bit
LPC
26 coefficients, pre-emphasis 0.95
46 ms gliding Hanning window
Overlap: 95%
20 to 150 measurement points per vowel,
depending on the duration
Method
Spectrogram of the item „Kies“ (gravel). Speaker sp012, reading sentences. Bottom
panel: fundamental frequency, Next panel from bottom: waveform window, Third panel
from bottom: spectrogram window, Left upper panel: waveform zoom window, rigth upper
panel: amplitude spectrum window.
Vowels in Standard Austrian German
Vowel system consists of 13 vowels on 5 constriction
locations:
•Pre-palatal:
/i, ç, y, Y/
•Palatal:
/e, E, ë, ê/
•Velar:
/u, ï/
•Upper-pharyngeal: /o, O/
•Lower-pharyngeal: /A/
These vowels are additionally distinguished by
constriction degree (or lip aperture) and lip protrusion.
Acoustic Correlates of Primary Stress in
German
Duration (Goldbeck & Sendlmeier 1988, Jessen et al.
1995, Mengel 1997, Dogil & Williams 1999)
Spectral tilt (Claßen et al. 1998)
Vowel quality (Jessen et al. 1995)
Difference between primary stressed and
unstressed vowels
Difference stressed - unstressed (in %)
35,000
30,000
25,000
20,000
15,000
10,000
5,000
0,000
/i/ c
/i/ uc
/e/ c
/e/ uc
/u/ c
/u/ uc
/o/ c
/o/ uc
/a/
F1
7,229
5,265
8,558
10,412
11,193
7,324
9,189
10,058
21,264
F2
11,192
6,211
15,346
10,718
27,561
24,602
28,686
16,073
5,825
F3
10,455
4,654
9,867
3,352
2,759
4,144
4,520
Mean values of the statistically significant differences (in %) between stressed and
unstressed vowels, broken for F1, F2, and F3, pooled over all speakers and speaking
tasks. Legend: c = [+constricted], uc = [-constricted]
Secondary Stress
No acoustic correlates found sofar for secondary stress
(Mengel 2000, Kleber & Klipphahn 2006)
Secondary Stress is a perceptual phenomenon (Mengel
2000, Schreuder 2006)
Secondary Stress
Secondary stress in compounds e.g.:
Fi'nanzmi]nister
'Bundes]kanzler
Be'suchs]recht
Be'suchs]tage
'Jugend]amt
(finance minister)
(chancellor)
(visiting rights)
(visiting days)
(youth welfare office)
In the spontaneous speech of the data at issue,
approximately 25% of all nouns were compounds.
The vowel /A/
Schematized change in vowel quality:
F1 primary stress > F1 secondary stress > F1 unstressed
F2 primary stress < F2 secondary stress < F2 unstressed
F3 primary stress < F3 secondary stress < F3 unstressed
The vowel /A/
F3
F2
F1
sp180
sp129
sp082
sp012
sp126
sp127
Statistically significant changes of F1, F2, and F3 (p < 0.05) of the vowel /A/ dependent on stress,
sentence reading task. Within each column denoting the speakers, the left-most crossbar denotes
the relative formant frequency position of primary stressed vowels, the middle crossbar the one of
secondary stressed vowels, and the rightmost crossbar the one of unstressed vowels.
The vowel /i/
Schematized change in vowel quality:
F1 primary stress < F1 secondary stress < F1 unstressed
F2 primary stress > F2 secondary stress > F2 unstressed
F3 primary stress > F3 secondary stress > F3 unstressed
The vowel /i/
F3
F2
F1
sp180
sp129
sp082
sp012
sp126
sp127
Statistically significant changes of F1, F2, and F3 (p < 0.05) of the vowel /A/ dependent on stress,
sentence reading task. Within each column denoting the speakers, the left-most crossbar denotes
the relative formant frequency position of primary stressed vowels, the middle crossbar the one of
secondary stressed vowels, and the rightmost crossbar the one of unstressed vowels.
The vowel /ç/
Schematized change in vowel quality:
F1 primary stress < F1 secondary stress < F1 unstressed
F2 primary stress > F2 secondary stress > F2 unstressed
F3 primary stress > F3 secondary stress > F3 unstressed
The vowel /ç/
F3
F2
F1
sp180
sp129
sp082
sp012
sp126
sp127
Statistically significant changes of F1, F2, and F3 (p < 0.05) of the vowel /A/ dependent on stress,
sentence reading task. Within each column denoting the speakers, the left-most crossbar denotes
the relative formant frequency position of primary stressed vowels, the middle crossbar the one of
secondary stressed vowels, and the rightmost crossbar the one of unstressed vowels.
The vowel /E/
Schematized change in vowel quality:
F1 primary stress > F1 secondary stress > F1 unstressed
F2 primary stress > F2 secondary stress > F2 unstressed
F3 primary stress > F3 secondary stress > F3 unstressed
The vowel /E/
F3
F2
F1
sp180
sp129
sp082
sp012
sp126
sp127
Statistically significant changes of F1, F2, and F3 (p < 0.05) of the vowel /A/ dependent on stress,
sentence reading task. Within each column denoting the speakers, the left-most crossbar denotes
the relative formant frequency position of primary stressed vowels, the middle crossbar the one of
secondary stressed vowels, and the rightmost crossbar the one of unstressed vowels.
Conclusion
High variability among the speakers (see also Kleber
& Klipphahn 2006)
Difference between primary stressed position and
unstressed position codified
Secondary stress can be expressed in different ways:
in the same way as the primary stressed vowel
in the same way as unstressed vowel
not at all
Depends on speaker and speaking task.
References
Claßen, Kathrin, Grzegorz Dogil, Michael Jessen, Krzysztof Marasek & Wolfgang Wokurek. 1998.
„Stimmqualität und Wortbetonung im Deutschen.“ Linguistische Berichte 174, 202-245.
Dogil Grzegorz & Briony Williams. 1999. "The phonetic manifestation of word stress." In: Harry van
der Hulst (ed.), Word Prosodic Systems in the Languages of Europe. Berlin, New York: Mouton de
Gruyter, 273-334.
Goldbeck, Thomas. P. & Walter F. Sendlmeier. 1988. "Wechselbeziehung zwischen Satzmodalität
und Akzentuierung in satzfinaler Position bei der Realisierung von Intonationskonturen." In: Hans
Altmann (ed.), Intonationsforschungen. Tübingen: Niemeyer, 305-321.
Jessen, Michael, Krzystof Marasek, Katrin Schneider & Kathrin Claßen. 1995. „Acoustic Correlates
of Word Stress and the tense/lax Opposition in the Vowel System of German.“ Proceedings of the
ICPhS ’95. Stockholm. Vol 4, 428-431.
Kleber, Felicitas & Nadine Klipphahn. 2006. "An acoustic investigation of secondary stress in
German." Arbeitsberichte des Instituts für Phonetik und digitale Sprachverarbeitung der Universität
Kiel, AIPUK 37, 1-18.
Mengel, Andreas. 1997. "Das akustische Korrelat des deutschen Wortakzents." Vortrag gehalten
anläßlich der Konferenz für Elektronische Sprachsignalverarbeitung (ESSV 1997) in Cottbus.
Mengel, Andreas. 2000. Deutscher Wortakzent. Symbole, Signale. Norderstedt: Books on Demand
GmbH.
Schreuder, Maartje. 2006. Prosodic Processes in Language and Music. PhD Dissertation,
Groningen Dissertations in Linguistics (GRODIL) 60, University of Groningen
THANK YOU FOR YOUR ATTENTION

Pres