Echoes of Innovation: AI in Dubbing and the Evolving Landscape for Voice Talents and Professional Studios

Abstract

Recent advancements in artificial intelligence (AI) have significantly impacted various industries, and the field of audiovisual translation is no exception. This research delves into the transformative effects of AI on dubbing practices through a comparative case study. Specifically, it examines the process of dubbing audiovisual content from English into Arabic, juxtaposing traditional studio voicing techniques with Respeecher, a cutting-edge AI-powered automated voicing solution. The research investigates the technical and prosodic challenges inherent in both approaches, exploring the intricate aspects of synchronisation and the conveyance of emotions. By scrutinizing a dubbing clip, the study critically assesses the quality of voicing achieved by human talent in a studio setting compared to that generated by AI software. The study employs a combination of established frameworks to evaluate voicing quality effectively: the Perceptual Evaluation of Speech Quality (PESQ) (Rix et al., 2001a) and the Revised Mean Opinion Score (MOS-R) (Lewis, 2001). While the PESQ assesses the quality of speech signals by comparing the synthesized speech to a reference human recording and evaluates objective signal property, the MOS-R captures subjective listener opinions on audio quality and enhances evaluation by accounting for contextual and emotional factors, ensuring a holistic assessment of dubbing quality. The findings presented offer valuable insights into each method's capabilities and limitations. While AI brings undeniable efficiencies and cost savings, it also faces challenges in replicating the nuanced expressions of human emotion. On the other hand, despite being resource-intensive, traditional studio dubbing continues to excel in terms of prosodic authenticity and emotional richness. This study contributes to the ongoing discourse on the evolving landscape for voice talent and studios in the age of AI. It provides a glimpse into the future of dubbing practices shaped by rapidly advancing AI technologies. The implications of these findings are far-reaching, particularly for stakeholders in the media and entertainment industries, as they prompt a thoughtful reconsideration of strategies for integrating AI into creative processes while maintaining the highest standards of quality and artistic integrity.

Date of Award	2025
Original language	American English
Awarding Institution	HBKU College of Humanities and Social Science