Satellite workshop: A Common Prosody Platform (CPP)

A Common Prosody Platform (CPP) for Comparing Prosodic Theories

Speech Prosody 2016 Satellite Workshop

Boston University, Saturday, June 4, 2016


Doug Whalen (Haskins Laboratories, City University of New York)

Yi Xu (University College London)

Santitham Prom-on (King Mongkut’s University of Technology Thonburi, Thailand)

Description and call for participation:

Despite fast advances in prosody research in recent years, many fundamental issues remain unresolved, and some are still under heated debate. A major reason for the theoretical difficulty is the lack of means to directly compare theories and models of prosody. To address this problem, a Common Prosody Platform (CPP: is being developed to enable close comparison of prosodic theories and models. CPP hosts major theories and models of prosody that have been converted into quantitative models that are trainable on real speech data. The trained parameters can then be used in theory-specific predictive synthesis of prosody. The synthetic prosody can be evaluated by comparison to natural prosody in term of numerical accuracy and perceived naturalness and contrastiveness. The quantitative models developed so far include AMtrainer for the AM theory, Fujisakitrainer for the Fujisaki model, TADAtrainer for the task dynamic model and PENTAtrainer for the PENTA model.

This one-day workshop will serve two purposes. The first is to introduce CPP and the models currently hosted, and the second is to discuss theory-driven computational modeling as an emerging research paradigm involving theories and models both within and beyond the current version of CPP. The workshop will therefore consist of a) introduction and tutorial of CPP and b) presentations of other computational approaches to speech prosody. We welcome participation by colleagues who want to either learn more about CPP or present their own computational work on prosody modeling, especially if it is theory-driven.


Morning session: Introducing CPP

2 hours of tutorial presentation, 2 hours of poster presentation and hands-on exploration.

Afternoon session: Talks and posters

Oral presentations plus open call for posters on prosody modeling, either with CPP or with other approaches

We expect to provide an onsite lunch


School of Education, Boston University (across the street from the Questrom building) 

Submission and acceptance of abstract:

Submissions are now closed.

Workshop Program:Schedule now available


Attendance is free of charge, however in order to gauge the level of interest we would appreciate expressions of interest via as early as possible in

  • attending the workshop either for a half or full day,
  • submitting an abstract for oral (talks of up to 20 minutes) or poster presentation

****IMPORTANT ****

Accommodation in Boston University Housing:

If you are attending the main SP2016 Conference and staying in on-campus housing, you may be able to extend your stay to include Friday and and/or Saturday nights. Please email for more information.

For those NOT attending SP2016, rooms may be available for Friday and Saturday if there are sufficient numbers. If you are interested in this option, please contact us as soon as possible at so that we can make the necessary arrangements.