The ALST Project: Technologies for Audio Description

Abstract

Matamala describes the ALST project, which aims to research the implementation of three technologies (speech recognition, machine translation and speech synthesis) in two audiovisual transfer modes (voice-over and audio description). Matamala focuses on the experiments related to audio description within the ALST project: firstly, tests in which speaker diarization and speech recognition systems were used to generate an audio description transcript; secondly, experiments in which the effort involved in creating, translating and post-editing an audio description were compared and, finally, tests in which end users assessed audio descriptions created using a text-to-speech system.