A Novel Lip Synchronization Approach for Games and Virtual Environments

This paper presents an algorithm for the offline generation of lip-sync animation. It redefines visemes as sets of constraints on the facial articulators such as lips, jaw, tongue. The algorithm was comparatively evaluated with 30 healthy participants by presenting a set of phrases delivered verbally by a virtual character. Each phrase was presented in two versions: once with the traditional lip-sync method, and once with our method. Results confirm that the suggested solution produces more natural animation than standard keyframe interpolation techniques.