Special Session: Sources of Prosodic Variation across Recording Settings
Sources of Prosodic Variation across Recording Settings:
- Oliver Niebuhr (Institute of Design and Communication, Innovation Research Cluster Alsion, University of Southern Denmark, Sønderborg, Denmark)
- Petra Wagner (Bielefeld University, Germany)
The task of describing and understanding speech communication places certain requirements on the acoustic quality and experimental control of speech signals and moreover involves a number of instrumental approaches. Thus, even though phonetic research actually targets speech communication outside the laboratory, “lab speech is indispensable in our quest to understand the underlying mechanisms of human language” (Xu 2010:329). Despite this inevitable fact, we know surprisingly little about the social, environmental, and task-specific factors that shape speech production inside the laboratory, apart from the fact that read speech differs in many respects from spontaneous speech (e.g,., Mixdorff & Pfitzinger, 2005; Dellwo et al., 2015). In the recent words of Wagner et al. (2015:1): “we need a much better understanding of the extent to which our methods and our ways of collecting speech data influence our results”.
With reference to the INSPECT initiative (i.e. “Innovating Speech Elicitation Techniques”) that was started by Niebuhr and colleagues in 2013 (see Niebuhr & Michaud 2015 for a recent overview), our special session aims at bringing together papers that can sensitize researchers to the challenges, sources of variation, and possible pitfalls of gathering production data inside the laboratory. Thus, possible topics include (but are to restricted to) in-depth analyses or comparisons of elicitation tasks (including types of reading material), phonetic/prosodic convergence, the issue of generalization of lab-speech findings, reliability of measurements as well as how different recordings conditions, signal qualities, and signal-compression procedures affect this reliability, individual speaker differences (e.g., with respect to expressive speech, fluency, laughter, breathing), effects and artifacts of recording conditions, such as type of microphone, eye contact, task duration (fatigue or order/repetition effects), or environmental noise. Papers that address these or related issues from a perceptual point of view are also welcome.
Our special session is only a small step towards a better understanding of how speech elicitation techniques affect lab speech. Yet, every small step will help us gain control over the factors that shape speech production and, ultimately, also enables us to exploit these factors for triggering or supporting certain phonetic/prosodic features or speaking styles.
Dellwo, V., A. Leemann, M.-J. Kolly (2015). The recognition of read and spontaneous speech in local vernacular: The case of Zurich German. J. of Phonetics 48, 13-28
Mixdorff, H., H.R. Pfitzinger (2005). Analysing Fundamental Frequency Contours and Speech Rate in Map Task Dialogs. Speech Communication 46, 310-325.
Niebuhr, O., A. Michaud (2015). Speech data acquisition – The underestimated challenge. Kalipho 3, 1-42.
Wagner, P., J. Trouvain, F. Zimmerer (2015). In defense of stylistic diversity in speech research. J. of Phonetics 48, 1-12. Xu, Y. (2010). In defense of lab speech. Journal of Phonetics 38, 329-336.