General description

In this workshop, we will be using the DARLA interface to force-align audio data with transcribed text. DARLA is a free online forced-alignment tool which provides a way to easily extract formant measures along with phoneme and word level alignments from large amounts of audio data (e.g., from sociolinguistic interviews) much more quickly and efficiently than if we were trying to do everything by hand.

We will go through all the necessary steps needed to get the most accurate alignments, look at common issues in the workflow which might cause the process to crash and then run some simple Praat scripts to process the data further (depending on what you might need it for). By the end of the workshop, you will be able to take your own audio data, force-align it through the DARLA interface and extract the information you need to further analyse your audio data. This workshop assumes you have some basic knowledge of Praat, but we will go through everything you need to know about Praat if not.

Please register by email to Prof. Schleef (erik.schleef@sbg.ac.at), copy to Cynthia Dermarkar (contact@hpsl.uni-freiburg.de).