Where APLS data comes from
On this page
Sociolinguistic fieldwork
The audio files in APLS come from sociolinguistic fieldwork conducted in Pittsburgh in 2003–2005 by Barbara Johnstone and Scott Kiesling. For more information on this fieldwork, see this page
From audio data to APLS
The audio files in APLS have been transcribed by research assistants according to a specific set of conventions that facilitate analysis in LaBB-CAT. (This takes a ton of time and effort!) After a transcription file is uploaded with an audio file to APLS, APLS generates numerous layers for the transcript, using dictionaries with compositional representations of words (e.g., morphological parsing), machine learning models (e.g., the [Hidden Markov Toolkit][htk] for determining time-alignments of individual speech sounds), and other techniques. Additionally, metadata about the participant and the transcript are uploaded to APLS.
On this page
- TOC