IMPLEMENTASI LIPSYNC ANIMASI 3D BERBASIS TEXT-TO-SPEECH DENGAN PARAMETER VISEME PADA FONEM BAHASA INDONESIA DI UNITY MENGGUNAKAN ICLONE 8

Lipsync animation in 3D character models is employed to provide a more human-like and realistic impression. However, accuracy remains a challenge when applied to the Indonesian language, as most systems utilize English-based viseme parameters, leading to visual discrepancy. This study aims to implement a 3D character lipsync animation workflow using iClone 8 and Unity by optimizing viseme parameters specifically for Indonesian phonemes. The method employed is Research and Development (R&D), comparing Condition A (automatic) and Condition B (manual). The test results indicate that manual optimization significantly improves visual articulatory accuracy, particularly for bilabial phonemes requiring full closure and vowels requiring specific adjustments. Furthermore, the application of fine-tuning techniques on animation curves is proven to produce smoother and more natural motion transitions, thereby minimizing the uncanny valley effect. The final implementation in the Unity game engine demonstrates that the animation aset run stably at 60 FPS without latency, proving that this method is effective for digital content needs.