NC

NCSOFT today introduced ‘Zero-Shot MultiVerse TTS’, an innovative AI model designed to enhance the entire process of game character voice production.

Text-to-Speech (TTS) is a voice synthesis technology that converts written text into spoken language. NCSOFT’s MultiVerse TTS excels in high-quality speech synthesis, requiring just a 3-second voice prompt. This capability allows a variety of speech style transformations, cross-lingual voice generation with exceptional speaker fidelity, and efficient multi-lingual voice production.

NCSOFT plans to integrate the MultiVerse TTS model extensively in game development to optimize the voice production process. By utilizing limited voice resources, the model facilitates the efficient creation of rich, AI-driven character voices while significantly reducing the time and costs associated with conventional production methods.

A standout feature of the MultiVerse TTS is its ability to function as a single model, capable of managing multiple languages and functions. This enables the generation of multi-lingual voice content at a fraction of the operational cost compared to competing models, while maintaining high-quality output.

NCSOFT has reaffirmed its position as a global leader in AI technology by publishing a research paper on the MultiVerse TTS model at the Empirical Methods in Natural Language Processing (EMNLP), one of the most prestigious AI conferences worldwide.

With a strong focus on AI research and development, NCSOFT is dedicated to advancing multi-lingual voice technology for global game releases. By the end of this year, NCSOFT aims to develop voices for 100 game characters using MultiVerse TTS and is also working on a control feature that adjusts NPC voices based on their personalities and situational contexts.