IMSCONV database

Corpus for investigating convergence in spontaneous speech dialogues

IMSCONV database

Antje Schweitzer, Natalie Lewandowski


IMSCONV has been created to investigate phonetic convergence in German spontaneous speech. The database consists of 46 dialogs of approx. 25 minutes length each, between previously unacquainted female subjects. 22 dialogs were in a unimodal setting, where participants could not see each other, while the remaining 24 dialogs were recorded with subjects facing each other.

The database was automatically annotated on the segment, syllable and word level using forced alignment with manually generated orthographic transcriptions.

It was designed for investigating phonetic convergence and its correlation with personality traits and scores of mutual liking and esteem, but in general the data can be used to investigate various aspects of interlocutors' behaviour in free, spontaneous dialogs, possibly considering social aspects.

We are planning to make the database freely available for non-commercial use in the near future.

To the top of the page