Logo Utrecht University

Digital Humanities Lab


Automatic Analysis of Speech Prosody (AASP)

The application Automatic Analysis of Speech Prosody (AASP) was an idea by Prof. Aoju Chen (UIL-OTS), to make two approaches to classify speech prosody available to researchers without having to run scripts. Dr. Berit Janssen from the Digital Humanities Lab developed a graphical user interface, in which users can upload audio files and files which contain information on word boundaries within the audio files (TextGrid format). Then users can choose analysis in ToDI tone labels (Transcription of Dutch Intonation), or to find descriptors for the pitch curves of the speech melody (Functional Data Analysis). Ph.D. student Na Hu trained models for the ToDI labels, and advised on the procedure of Functional Data Analysis.