UA Forced Alignment Workshop

Workshop for Transcription & Alignment of Linguistic Data

We are offering a 2-day workshop for training in transcription and alignment of speech. It will take place on the 27th and 29th of January. The workshop is funded by a grant from the Confluence Center, so there will be no charge for participants.

Any undergraduate who wishes to make a phone call to his/her grandmother, to be included in the corpus, please contact malcah@email.arizona.edu or anac@email.arizona.edu

The advantages of preparing your own data for detailed analysis using forced alignment are well-known.  The system is now used in many Linguistics Departments, as well as in Computer Science, Sociology and Communications departments. Using a preliminary transcript and its sound file, the FAVE system (http://fave.ling.upenn.edu/FAAValign.html) outputs a time-aligned phonetic transcript into a Praat (www.fon.hum.uva.nl/praat/download_mac.html) sound file. Automatic boundary-marking of phonemes reduces time-consuming boundary-marking by hand, facilitating the linguistic analysis of a much larger number of tokens, permitting the researcher to code accurately, and to recheck ‘outlier’ tokens [of phonological, prosodic, or other discourse phenomena] efficiently, saving researcher-hours: For example, Labov et al. (2013) found the forced alignment increased the number of vowel tokens they could analyze from each interview from 300 to 9,000!

 

The first day of the workshop (1/27/16), in Communications Room 309, from 9-11, will be spent on preliminary transcription training.  Students will transcribe a segment of a corpus of their own choosing. 

Check out the conventions Here.

 

The second half of the workshop (1/29/16), from 9-12 and 1-3, in Gould-Simpson 942, will permit a guest researcher, Eric Wilbanks from NCSU, to train students to use FAVE (Forced Aligner of Vernacular English) or FASE (Forced Alignment System for Español) to align their preliminary transcripts with the PRAAT wave-form. He has run several previous workshops. Software for other languages will also be discussed. Students will learn how to use the computational software to facilitate their own research needs. We advise that you arrive with PRAAT installed on your laptop! 

Added instructions are found  here.

Additional information can be found on Eric Wilbanks' website here.

 

This training will provide the opportunity to become paid transcribers for monolingual [English] or bilingual [English-Spanish] telephone conversations now being recorded here in Arizona, and will also facilitate the students’ own future research needs.

 

If interested, or for further inquires, please contact us.

We are hoping to see you at the workshop!

Malcah Yaeger-Dror (malcah@gmail.com) and Ana Carvalho (anac@email.arizona.edu), Co-Investigators

Powerpoints from this presentation can be found here and here.

 

Intergenerational Telephone Conversations