😳4D Face reconstruction to generative pipelines…
We’ve busy building away, exploring some interesting new ways to create…. Combining 4d face reconstructing, with automated character reference and motion transfer allowing for more control in motion transfer generative camera
moves and recapturing performance from different angles. Very early days with this tech but the potential really is mind blowing. Beginning to get some really interesting scene based test working, keen to really stress test this beast out, let’s hope it holds up at scale🤞
R&D making me hungry 🤤
Testing how far we can strip back CG and still get strong results. Giving AI more freedom creates great outcomes.
Simple proxy meshes for caramel control + basic particles for falling chocolate chunks help when precision is needed.
This was using Kling 2.6, been getting some impressive results out of Ray 3 modify also, which has so much more flex but needs a bit more fiddling to get around quality issues. Finding adding some subtle motion blur, grain, refocus/sharpen really make the difference
2026 is gonna be a crazy year!
More experimenting with Al motion capture, holds up pretty well! Feeling like camera moves aren't translating fully so will need some re compositing over original and some flaws with the mouth for sure.
But damn it’s a big step forward
@keenanisalive Ahh great thanks, apologies I was massively over thinking this, pipeline wise the displacement can be done normally in Blender etc, result looks great, thanks for the tip!
Testing a bit of a new pipeline going from from @Blender into @LumaLabsAI then on to @Magnific_AI to upres keyframes and on to moonvalley Marey for keyframe control. Early days but can see a lot of potential!
R&D into some video AI relighting techniques using Switchlight 2.0, Topaz, Blender & traditional compositing work. Potential to be a super powerful pipeline 🚀 #ai#VFX#switchlight#Topaz#blender