Filmmakers and animators will soon be able to create facial expressions and micro-expressions for their characters with significantly less effort. This is promised by the new AI tool “Act-One,” which Runway is now introducing as part of its Gen-3 Alpha video model. The feature aims to replace complex motion capture procedures with simple video recordings.
Until now, productions required specialized equipment such as marker systems and elaborate face rigging to create realistic facial animations. According to Runway, Act-One works with footage from standard cameras. The tool transfers not only basic facial expressions but also fine details such as eye lines and movement timing. Runway claims that creatives can achieve cinematic results even with limited budgets. Additionally, a single actor can record voice and facial performances for different characters, which can then be transferred to various generated figures.
To prevent misuse, Runway says it has implemented various safety measures. These are designed to block unauthorized depiction of public figures and verify rights to used voices, among other protections. The tool is being gradually rolled out to users with sufficient credits in the Gen-3 Alpha model.
Sources: Runway, VentureBeat