Speech Graphics – Automatic Lip Sync from Voice Recordings





Speech Graphics Ltd is an award-winning spinoff from the University of Edinburgh’s world-renowned School of Informatics and the Centre for Speech Technology Research. The founders Michael Berger and Dr. Gregor Hofer (above) are speech technology experts with 20 years of combined experience in speech animation. The facial animation in this video above was generated completely automatically from audio.

Rockstar Games’ best-selling detective thriller L.A. Noire featured extremely detailed and realistic facial graphics and animation thanks to Depth Analysis’ MotionScan technology. The problem with MotionScan’s technology, however, is the sheer amount of storage space it requires.

Speech Graphics’ solution is a much quicker and less data-intensive process than MotionScan which requires the actor to sit in a chair and perform their script while recorded by 32 high definition cameras.

There are rumours that Rockstar Games’ GTA V is set to use and advanced facial animation technique for better realism in character interaction.

For more information and demos, check out www.speech-graphics.com