AI-Generated Voice Acting for Real-Time Character Dialogue in Game
Year: 2025
Type: article
Abstract: : In many expansive video games, particularly MMORPGs and large open-world RPGs, it is often the case that voice acting is entirely absent due to the extensive dialogue and intricate storylines. As a result, players often rely on text subtitles, which may reduce immersion and limit accessibility. This study explores the expanding potential of AI-generated voice acting to enhance video games, circumventing the substantial expenses and protracted timelines associated with traditional voice recording methods. It analyzes how innovations in AI text-to-speech technologies, including Tacotron, WaveNet, and transformer models, facilitate the instantaneous generation of natural and emotionally nuanced character voices. Furthermore, the research evaluates the incorporation of these technologies into video games to deliver dynamic voiceovers for non-playable character dialogues, particularly in quests or side missions that generally lack voice acting. We examine the existing landscape of tools and platforms, investigate the technical and creative challenges, and highlight the prospects where AI-generated voice acting could transform narrative design. We also look into the ethical aspects, the benefits for localization, and how it can help indie developers by making voice acting easier to access. This study aims to kick off conversations about how AI can fix gaps in game storytelling, especially when traditional methods just won't cut it.
Author: Patil, Rijul Rajendra
Institution: University of Mumbai
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6893326170
FWCI:
Cites:
Cited by:
Related to: