Prosodically annotated speech databases

The corpus of prosodically annotated speech data that is exploited in project C4 consists of the following components:

Material gathered via digital satellite radio at IMS

• news stories recorded from the digital satellite radio at IMS
• a feature series recorded from the digital satellite radio at IMS

Material recorded by others but further processed at IMS

• acted speech from a multimedia CD-ROM
• material recorded elsewhere but processed by the project

Material recorded at IMS

• read short excerpt of a novell recorded at IMS
• material recorded under controlled lab conditions at IMS
• some read newspaper articles recorded at IMS

Processing of the database

all subcorpora are further processed to different extends:

• automatic word, syllable and phoneme alignment
• automatic POS tagging
• automatic ToBI labelling
• manual ToBI labelling

There are several utilities that ease acces to the database

• selection & search utilities