CAVA home (British Sign Language Corpus Project)

Background

The CAVA (human Communication: an Audio-Visual Archive) repository contains rights-cleared primary audio and video recordings made by UCL’s human communication researchers and their associates. The British Sign Language Corpus Project (BSLCP) section of the CAVA repository contains video recordings of BSL video data from 249 deaf signers of BSL from 8 regions around the UK: London (L), Bristol (BL), Cardiff (CF), Birmingham (BM), Newcastle (N), Manchester (M), Glasgow (G) and Belfast (BF). The participants are mixed for gender, age group, age of BSL acquisition, social class and ethnicity. There are 2 levels of access. The Open Access data includes narratives (n) and lexical elicitation (l), and the Restricted Access data includes interviews (i), conversation (c) and some combined narrative followed by conversation (n-c) – for more detail see BSLCP Activities. All participants were filmed in pairs for all activities, with three cameras: one on each participant, and one on the pair.

Example Resource titles and what they mean:

BF9+10i = Belfast participants number 9 and 10 in the interview activity

CF17n = Cardiff participant 17 in the narrative activity

Accessing BSL Corpus data on CAVA

In order to view narrative or lexical elicitation data in your browser, simply search or browse CAVA, and you will be able to view the video clips without logging in. Downloading the Open Access (narrative or lexical elicitation) data to your computer first requires agreeing to some terms and conditions. Viewing or downloading the Restricted Access (interview or conversation) data (either viewing in your browser or downloading to your computer), requires a user license (request a user license here).

Depending on demand and the details of your particular request, the process of obtaining an End User License may take several weeks. After you have obtained your End User Licence, please log in here.

Updates

As of 2011, the BSLCP dataset consists of video clips only. Annotations, transcriptions and/or translations will be made available on this site periodically, after they have been completed and checked. The annotation guidelines for the BSL Corpus Project roughly follow those by Trevor Johnston for the Australian Sign Language Corpus Project: http://www.auslan.org.au/about/annotations/

For more information, follow the links below: