Designing a new platform for AD analyses. The VIW project

Audio description (AD) is an intersemiotic translation in which images are translated into words. Research on AD is recent, and very few projects have approached the creation of a corpus of AD (TIWO, TRACCE). However, two frequent problems are found: processing multimodal content and dealing with copyright. This is probably why there is not a freely available corpus of AD. Inspired by Chafe’s Pear Stories project, VIW aims to develop a platform which will build an ad hoc corpus of ADs. Departing from a shared stimulus, a wide corpus of ADs in multiple languages will be developed, allowing for contrastive studies. In our paper we will be presenting the project rationale, the platform architecture and its development. Key aspects that will be discussed are the creation of corpora from authentic versus created materials, the need for open access copyright-free platforms, as well as the challenges of processing multimodal multilingual data.