Skripsi
PENERAPAN RETRIEVAL-BASED VOICE CONVERSION UNTUK MEMBANGUN ASET SUARA KARAKTER MONSTER PADA GAME
Audio in video games plays a crucial role in creating an immersive and engaging gaming experience. However, indie game developers often face resource constraints, particularly in producing high-quality character voices. Voice cloning technology based on Retrieval-based Voice Conversion (RVC) offers an innovative solution by enabling accurate voice replication and transformation using minimal data. This technology provides flexibility for developers to create unique character voices without requiring direct voice actors, making it more cost- and time- efficient. This study aims to explore the application of RVC technology in generating monster character voices for video games, using a case study with voice clips from the anime Goblin Slayer. Through the voice conversion process, this study evaluates the performance of RVC models using objective standards such as Mel Cepstral Distortion (MCD) to measure voice transformation and Mel- Frequency Cepstral Coefficients (MFCC)-based analysis to assess the similarity between the original and converted voices. The results demonstrate that RVC technology is capable of producing voices significantly different from the original, enabling the creation of audio characters tailored to the narrative needs of the game. This technology offers an effective solution for indie game developers to create innovative and high-quality audio assets at low cost. This study highlights the importance of voice cloning technology in supporting competitive indie game development in the market.
No other version available