Katrin Schweitzer

Katrin Schweitzer
Telefon 0049 711 685-84589
Universität Stuttgart
Institut für Maschinelle Sprachverarbeitung
Pfaffenwaldring 5b
70569 Stuttgart

Currently, I am working the the Infrastructure Project of the SFB 732.  We are responsible for a number of aspects, all centering around the data collected in the SFB. Within the project, one of the topics I concerend with is bringing together the output of the text-based tools and annotations available at the IMS with our speech data (and its automatic or manual annotations). To this end, I support my incredibly clever colleagues Kerstin Eckart and Markus Gärtner, who are computer geniuses (and much more), by bringing phonetic expertise into the team.

Markus developed a powerful search tool for linguistic data, ICARUS. Its prosody module, which we worked on, recently, brings together output from parsers, taggers, human annotators, ... with the acoustics of intonation. We used a parametric intonation model, PaIntE, to make characteristics of the tonal contour available in the search. For instance, you can search for "prototypes" of accents (and instances similar to the prototype) in specific configurations of the linguistic context. You can directly play the utterances, see the PaIntE values, export the result of your search, and in the future you will also be able to correct annotations or change other aspects. 

From July 2006 to July 2014 I worked in the project Incremental Specification of Focus and Givenness in a Discourse Context (A1) of the SFB 732  investigating the semantic-prosodic interface. We annotated both information structure and intonation, to build up corpora for statistical analyses of correlations between these two aspects. For instance, we released the DIRNDL database.  Moreover, we aimed to intergrate pitch accents as well as their discourse context into a usage-based model of intonation. We examined the effects of lexical frequency and frequency of contextual configurations on pitch accent realisations. The results we gatherd present a challenge for traditional autosegmental theories of intonation while a usage-based account predicts them correctly.

From July 2010 to October 2011 I also worked in a DFG-project on Prosodic Phrasing that investigates the status of prosodic phrasing in auditory and visual sentence processing. Prosodic phrase boundaries can influence the syntactic structure as well as the processing of argument status. One aim of the project is to identify the exact timing and functional make-up of prosody processing by measuring event-related potentials (ERPs). However, we also looked at phrasing in a corpus-phonetic approach. Specifically, we looked at the idea of prosodic balance, that is, the idea that speakers tend to produce prosodic phrases of equal structure.




Research Intrests

I love digging into data and finding relationships between different aspects of Speech and language. When intonation is involved, I am most happy :-), but I am also intersted in other aspects of prosody, in prominence, segmental features and also in extra-linguistic features.

My PhD thesis was concerned with  the question whether intonation can be lexicalised, and how intonation and the context in which tonal events occur can be modelled in a usage-based theory of speech production and perception.  I looked at effects of word frequency on pitch accent shape. They exist! The results fit well with the idea of Exemplar Theory, therefore I outlined how an existing exemplar model can be modified to deal with such phenomena. Now I am keen to find out more about how the production and perception of intonation works and how models of language and speech processing could integrate intonation...

  • WS 03/04. Tutorium zu Intonation und Informationsstruktur (Klaus von Heusinger)
  • WS 04/05. Tutorium "Labeln von Intonation" zum HS Spracherkennung (Grezegorz Dogil)
  • WS 07/08. Phonetik (Schule für Logopäden, Tübingen)
  • WS 08/09. Phonetik (Schule für Logopäden, Tübingen)
  • WS 11/12. Parts of "Topics in Laboratory Phonology"
  • WS 13/14 R-Sessions of "Computational Skills for Computational Linguists"
  • WS 14/15 R-Sessions of "Computational Skills for Computational Linguists"
  • WS 16/17 R-Sessions of "Computational Skills for Computational Linguists"



  • Katrin Schweitzer, Kerstin Eckart, Markus Gärtner, Agnieszka Falenska, Arndt Riester, Ina Rösiger, Antje Schweitzer, Sabrina Stehwien, Jonas Kuhn: "German Radio Interviews: the GRAIN release of the SFB732 Silver Standard Collection (accepted at LREC 2018).

  • Kerstin Eckart, Markus Gärtner, Jonas Kuhn und Katrin Schweitzer: "Nützlich und nutzbar für die linguistische Forschung: Sprachtechnologische Infrastruktur" (in preparation) 


  • Katrin Schweitzer, Michael Walsh, Antje Schweitzer: "To see or not to see: Interlocutor visibility and likeability influence convergence in intonation". Interspeech 2017, Stockholm, Sweden.


  • I took leave in 2016.


  • Katrin Schweitzer, Markus Gärtner, Arndt Riester, Ina Rösiger, Kerstin Eckart, Jonas Kuhn and Grzegorz Dogil (2015): Analysing automatic descriptions of intonation with ICARUS. Proceedings of Interspeech 2015, 6-10 September 2015, Dresden, Germany, Pages 319-323.

  • Nadja Schauffler & Katrin Schweitzer (2015): Rhythm influences the tonal realisation of focus. Proceedings of Interspeech 2015, 6-10 September 2015, Dresden, Germany. 

  • Niamh Kelly & Katrin Schweitzer (2015): Examining Lexical Tonal Contrast in Norwegian Using Intonation Modelling. Proceedings of the 18th In­ter­na­tional Con­gress of Pho­netic Sci­ences, 10-14 Au­gust 2015, Glas­gow, UK.

  • Frank Kügler, Bernadett Smolibocki, Denis Arnold, Stefan Baumann, Bettina Braun, Martine Grice, Stefanie Jannedy, Jan Michalsky, Oliver Niebuhr, Jörg Peters, Simoin Ritter, Christine T. Röhr, Antje Schweitzer, Katrin Schweitzerm, Petra Wagner (2015): DIMA - Annotation Guidelines for German Intonation. Proceedings of the 18th In­ter­na­tional Con­gress of Pho­netic Sci­ences, 10-14 Au­gust 2015, Glas­gow, UK.

  • Markus Gärtner, Katrin Schweitzer, Kerstin Eckart and Jonas Kuhn (2015): Multi-modal Visualization and Search for Text and Prosody Annotations. Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (ACL-IJCNLP 2015) System Demonstrations, Jly 26-31 2015, Beijing, China.

  • Nadja Schauffler, Antje Schweitzer, Katrin Schweitzer und Petra Augurzky (2015): Avoiding melodic clashes in pitch accent production: a corpus study. In: Leemann, Adrian, Marie-José Kolly, Volker Dellwo, Stephan Schmid (eds.), Trends in Phonetics and Phonology in German-speaking Europe.








Student Theses






Member of the local organization committee of Labphon 13 (Stuttgart, 2012).
Member of DIMA - Deutsche Intonation und Modellierung.
Member of Clarin F-AG 6 (Sprache und andere Modalitäten).
Reviewer for Frontiers in Psychology.
Reviewer for Interspeech 2015.
Chair for Phonetik&Phonologie 2009.
Chair for Interspeech 2015.




International Cooperations

  • 2014/2015 Niamh Kelly, University of Texas, USA

  • 2012 Dermot Lynott, University of Manchester, UK

  • 2010/2011 Sasha Calhoun, University of Wellington, New Zealand / University of Edinburgh, UK

  • 2009 Matthew E. Adams, Uriel Cohen Priva, Stanford University, USA



And now for something completely different ...


Projekte, die ich besucht habe und unterstützenswert finde