Topic: Post-Editing in Audiovisual Translation: where is it taking us?
The rapid advancements in machine translation (MT) and artificial intelligence (AI) have significantly impacted the field of Audiovisual Translation (AVT), particularly in subtitling, dubbing, and speech recognition. Consequently, post-editing has evolved into a vital skill for AVT specialists. While MT can enhance efficiency, it often struggles with linguistic accuracy, cultural nuances, and the technical constraints of AVT, necessitating human intervention to ensure quality.
This presentation explores the role of post-editing in AVT, examining its challenges, opportunities, and practical applications. Key topics will include the limitations of MT in AVT, such as contextual misunderstandings, text condensation, and synchronization with audiovisual elements. To illustrate the challenges of post-editing and assess students’ reliance on machine output, an experiment was conducted with a group of master’s students. They were given a short video extract with machine-generated subtitles containing several introduced errors, including mistranslations, incorrect segmentation, and unnatural phrasing. The experiment aimed to measure how often students blindly accepted machine output versus applying critical post-editing skills to correct inconsistencies. The results revealed varying degrees of dependence on MT, highlighting the need for training that enhances students’ ability to critically evaluate and refine automated translations rather than passively accepting them.
Senior lecturer and assitant researcher at ISCAP-IPP in Languages and Cultures. Master of Anglo-American Culture from the Faculty of Arts and Humanities, University of Lisbon and PhD in Translation and Paratranslation from the University of Vigo. Member of the research unit in Languages, Communication and Education at CEOS.PP and Transmedia Portugal. Research interests lie in the field of Audiovisual Translation, with a focus on Subtitling, Dubbing, Accessibility and Inclusion.
