FIRST IT WAS confined to hard surfaces and solid objects; then it conquered water and fire; then fully-CG characters became a feasible, convincing option. Today, there is little that the VFX industry can’t do with enough time and money.
The problem is finding both. Photo-realistic people were a final frontier for years, but even that is now achievable, at least with a big budget and in short bursts. Think the blend of Alyla Browne and Anya Taylor-Joy as the young Furiosa, or the de-ageing of Tom Hanks in Robert Zemeckis’ upcoming Here (both AI-assisted). “Digital humans work fine if they’re not doing much — looking moody and soulful is way easier than delivering Henry V’s St Crispin’s Day speech,” says Paul Franklin, a two-time Oscar-winner for his VFX work and one of the founders of London-based VFX giant DNEG. “But getting a completely rounded emotional performance from a completely synthetic digital human is still a way off.” Franklin also thinks that traditional motion-capture suits, with all their reflective dots and reference points, will disappear in the next few years, because machine learning can capture 3D movement from normal video images. James Cameron managed underwater performance-capture, at considerable expense, for Avatar: The Way Of Water; so this could bring his budgets down to under the GDP of many countries. Many traditional VFX jobs and processes, in fact, may be replaced by AI. But eventually there could be an artistic pay-off to this turmoil.