Dr. Eva Lasarcyk

Department of Computational Linguistics and Phonetics
Building C7.2
Saarland University
66041 Saarbrücken

I joined Speech at Amazon in December 2014. However, I still receive, and intermittently read, my coli emails in case you would like to contact me.

Recent Activities

Post-doctoral researcher in the Project C1: Information Density and the Predictability of Phonetic Structure of the Collaborative Research Center SFB 1102: Information Density and Linguistic Encoding.

Evaluation of different speech synthesis systems with regard to their possibilities to carry out voice manipulations for behavioral and EEG perception experiments, in collaboration with Heiner Drenhaus.

Research Areas

I am interested in state-of-the-art speech technology and its use in everyday life as well as in research, currently using synthesis techniques to study human speech production and perception.

How does the quality of synthetic speech influence perception?

How can speech synthesis be used to model articulatory mechanisms and characteristics of speech, such as regional pronunciation or smiling?

Furthermore, I am interested in techniques for conducting web-based listening tests, and especially also interested in applications of speech technology to everyday settings in the home, mobile, e-health, and assisted living environments.

PhD Thesis using VocalTractLab

I completed my Ph.D. thesis in phonetics under the supervision of Prof. emeritus William J. Barry and Prof. Dr. Bernd Möbius, at the department of Computational Linguistics and Phonetics, Saarland University, and the International Research Training Group Language Technology and Cognitive Systems.

The focus of my thesis is twofold: I have been addressing questions of speech production mechanisms regarding (mainly) paralinguistic phenomena, and assessing the capabilities of a new "discovery tool" for phonetic research.

I have used the articulatory speech synthesis system VocalTractLab, developed by Peter Birkholz, to explore production and perception aspects of articulatory synthetic speech. The speech phenomena investigated so far include larynx height and voice quality, laughter in conversational speech, speech laughs and smiled speech, means to convey (un-)certainty in speech, synthesis of vocal age, and the articulatory simulation of a German regional accent (Saxon).

The studies are mostly explorative in nature to find out how well the articulatory speech synthesis system is suited as a "discovery tool" for phonetic research, where the goal is to investigate articulatory processes in speech production. The experiments cover sub-glottal, glottal and supra-glottal aspects of speech production, and they are evaluated by acoustic, visual-articulatory, and perceptual means.

At our institute, Ingmar Steiner also worked with this articulatory synthesis software.

Past Activities

Teaching introductory courses (Introductions to the speech sciences, transcription and phonology) at the Institute for German language and literature studies (Germanistik), Abteilung Neuere Deutsche Sprachwissenschaft (NDS), Saarland University. SS 2013, WS 2012/13, SS 2012. Besides teaching basics of phonology and transcription of Standard German pronunciation, I focused on the transcription of regional pronunciations and foreign-accented pronunciations.

Project work at Lehrstuhl für Sprach- und Signalverarbeitung der Universität des Saarlandes in the BMBF-Projekt "Softwarecluster EMERGENT" (2011-2012).

Mitarbeit am Lehrstuhl für Experimentelle Neuropsychologie der Universität des Saarlandes / International Research Training Group "Adaptive Minds" (2010-2011).

Visiting PhD student at the Centre for Speech Technology Research (CSTR) at the University of Edinburgh during autumn 2009.

Organization of a research project during the 2008 IRTG Annual Research Meeting in Edinburgh: Search Engine for Phonetic Corpora.

Participation in the Johns Hopkins Summer School and Summer Research Workshop in Baltimore in 2008, working in the team of Vocal Aging Explained by Vocal Tract Modeling.

Organization of a research project during the IGK Annual Research Meeting 2007: aXenTTS - Synthesis of Foreign-Accented Speech.

Organization of the social activities during the annual research meeting.

Perl-Programmierkurs SS 2007 (with Ingmar Steiner).

Documents

Eva Lasarcyk, Heiner Drenhaus, Bern Möbius (accepted). Experimente zur Wahrnehmung gezielt degradierter synthetischer Sprache. 26. Konferenz Elektronische Sprachsignalverarbeitung (ESSV 2015). Eichstätt, March 2015.

William J. Barry, Bernd Möbius and Jürgen Trouvain (eds.) (2014). PHONUS 17: Eva Lasarcyk: Empirical evaluation of the articulatory synthesizer VocalTractLab as a discovery tool for phonetic research: Articulatory-acoustic investigations of paralinguistic speech phenomena. Saarbrücken: Institute of Phonetics, Saarland University. (PhD thesis, Saarland University.) [pdf] [PHONUS main page]

Eva Lasarcyk, Peter Birkholz and William J. Barry (accepted). Imitating a bi-dialectal speaker using acoustic-to-articulatory inversion: Articulatory basis of vowels in Saxon and Standard High German. International Workshop on "Multilinguality in Speech Research: Data, Methods and Models". Schloss Dagstuhl, April 2014.

Eva Lasarcyk and William J. Barry (2013). Articulatory basis of vowels in Saxon vs. Standard High German - As suggested by acoustic-to-articulatory inversion using articulatory speech synthesis. Poster at Phonetik und Phonologie 9 (P&P 9), Zürich, Switzerland, October 2013. [abstract]

Eva Lasarcyk, Charlotte Wollermann, Bernhard Schröder, and Ulrich Schade (2013). On the Modelling of Prosodic Cues in Synthetic Speech - What are the Effects on Perceived Uncertainty and Naturalness? In Proceedings 10th International Workshop on Natural Language Processing and Cognitive Science (NLPCS 2013), Marseille, France, October 2013. [pdf] (PR)

Charlotte Wollermann, Eva Lasarcyk, Ulrich Schade, and Bernhard Schröder (2013). Disfluencies and uncertainty perception - Evidence from a human-machine scenario. In Proceedings 6th Workshop on Disfluency in Spontaneous Speech (DiSS 2013), TMH-QPSR 54(1), Stockholm, Sweden. [pdf] (PR)

Michael Wiegand, Benjamin Roth, Eva Lasarcyk, Stephanie Köser, and Dietrich Klakow (2012): A Gold Standard for Relation Extraction in the Food Domain. In: Proceedings of the international conference on Language Resources and Evaluation (LREC), Istanbul, May 23-25. 507-514. [pdf] (PR)
Website Relation Extraction in the Food Domain.

Eva Lasarcyk and Charlotte Wollermann (2010): Do Prosodic Cues Influence Uncertainty Perception in Articulatory Speech Synthesis? In: Proceedings 7th ISCA Worskhop on Speech Synthesis (SSW7), Kyoto, September 22-24. 230-235. [pdf] (PR)

Eva Lasarcyk (2010): Acoustics vs. Articulation in Articulatory Speech Synthesis: One Vocal Tract Target Configuration Has More Than One Sound. In: Konferenz Elektronische Sprachsignalverarbeitung (ESSV'10), Berlin, September 8-10. [pdf] [abstract]

Werner Spiegl, Georg Stemmer, Eva Lasarcyk, Varada Kolhatkar, Andrew Cassidy, Blaise Potard, Stephen Shum, Young Chol Song, Puyang Xu, Peter Beyerlein, James Harnsberger, Elmar Nöth (2009). Analyzing Features for Automatic Age Estimation on Cross-Sectional Data. In: Proceedings 10th Interspeech, Brighton, September 6-10. 2923-2926. [pdf] (PR)

Eva Lasarcyk (2009): VocalTractLab als adäquates Entdeckungswerkzeug für Phonetik und Phonologie. 5. Jahrestreffen Phonetik und Phonologie, Köln, 3. März 2009. [abstract]

Jürgen Trouvain, Sascha Fagel and Eva Lasarcyk (2009). Überlegungen zur Messung von Larynxhöhe. 5. Jahrestreffen Phonetik und Phonologie, Köln, 3. März 2009. [pdf] [abstract]

Sascha Fagel, Jürgen Trouvain and Eva Lasarcyk (2009). Observing lip and vertical larynx movements during smiled speech (and laughter). Interdisciplinary Workshop on Laughter and other Interactional Vocalisations in Speech, Berlin, February 27-28, 2009. [pdf]

Eva Lasarcyk and Jürgen Trouvain (2008): Spread Lips + Raised Larynx + Higher F0 = Smiled Speech? - An Articulatory Synthesis Approach. Proceedings 8th International Speech Production Seminar (ISSP), Strasbourg, December 8-12. 345-348. [pdf] [Supplementary files] (PR)

Eva Lasarcyk and Varada Kolhatkar (2008). Error Analysis of Formant Tracking Algorithms. In: Peter Beyerlein, Andrew Cassidy, Varada Kholhatkar, Eva Lasarcyk, Elmar Nöth, Blaise Potard, Stephen Shum, Young Chol Song, Werner Spiegl, Georg Stemmer, Puyang Xu: Vocal Aging Explained by Vocal Tract Modelling: 2008 JHU Summer Workshop Final Report. Center for Language and Speech Processing, Johns Hopkins University, Baltimore. [pdf]

Charlotte Wollermann and Eva Lasarcyk (2007): Modeling and perceiving of different degrees of certainty in articulatory speech synthesis. In: Proceedings 6th ISCA Workshop on Speech Synthesis, Bonn, August 22-24. 40-45. [pdf] (PR)

Eva Lasarcyk (2007): Investigating Larynx Height With An Articulatory Speech Synthesizer. In: Proceeding 16th International Congress of the Phonetic Sciences (ICPhS), Saarbrücken, August 6-10. [pdf and supplementary files] (PR)

Eva Lasarcyk and Jürgen Trouvain (2007): Imitating Conversational Laughter With An Articulatory Speech Synthesizer. In: Proceedings Interdisciplinary Workshop on the Phonetics of Laughter, Saarbrücken, August 4-5. 43-48. [pdf][Homepage] (PR)

Eva Lasarcyk (2006): "Phonetik und Phonologie". In: Irene Cramer and Sabine Schulte im Walde (Editors). Computerlinguistik und Sprachtechnologie: Eine Studienbibliographie. Institut für Deutsche Sprache, Stauffenberg Verlag, Mannheim, 2006. [Online references]

Eva Lasarcyk (2005): Wie deutsch klingt Englisches im Deutschen - und warum? Magisterarbeit, Institut für Kommunikationsforschung und Phonetik, Universität Bonn. (How German do English words sound in German - and why? Master's thesis, University of Bonn.) [pdf]

(PR) indicates peer-reviewed publications.