ABSTRACT
This chapter examines how multimodal corpora (MMC) can be exploited for the purposes of audiovisual translation (AVT) research. It discusses the architecture of these resources, consisting of audiovisual texts and their translations, and explores various aspects pertaining to the design, compilation, processing and maintenance of these corpora. The chapter also examines the difficulties that are endemic to MMC. It suggests different avenues for the development of such computer-held resources in terms of their contribution to AVT research. The chapter explores the current and potential capacity of MMC to yield new insights into AVT. MMC can hold certain types of audiovisual texts that are translated. Along with films, several TV genres, theatre performances, videogames, interactive virtual assistants, pedagogical materials and audio books are likely to be of commercial or political interest, and hence likely to be translated. Research on multimodality has been mostly restricted to face-to-face interaction, without much consideration of visual or acoustic elements in the surrounding situation.
