VFX roundup flags production shifts

A CG/VFX news roundup published April 7 highlights ongoing shifts in visual‑effects work—think AI assisting rotoscoping and cleanup, real‑time rendering shrinking iteration cycles, and toolchain integration changing how images get made and shipped. (youtube.com)

Visual effects work is being pulled earlier in the pipeline, not just sped up at the end. A March 2026 industry roundup pointed to the same pattern across new tools: masking, asset building, and look development are moving closer to “make it now, fix it later” speed. (youtube.com) One of the clearest examples is rotoscoping, which is the frame-by-frame job of cutting a person or object out of a shot like tracing around it on thousands of photographs. Boris FX said Mocha Pro 2026 added Matte Refine Machine Learning to clean up fine hair, soft blur, and fast motion blur after the first mask is made. (blog.borisfx.com) Another new tool, ONYX AI Matte 2.5, does a similar job from rough boxes, guide points, or even a text prompt like “person” or “car,” then tracks that object through the shot. CG Channel reported on March 24 that it can track up to 32 objects in most modes and 64 in text mode inside Nuke, Fusion, and DaVinci Resolve. (cgchannel.com) That changes who spends time where. Instead of an artist drawing every edge by hand, the software makes the first pass and the artist spends more time correcting bad frames, checking hair edges, and deciding what should stay in the shot. (blog.borisfx.com) (cgchannel.com) Real-time rendering is pushing the same shift from another direction. Twinmotion, Epic Games’ real-time visualization tool, is built to show lighting, materials, and camera moves immediately instead of waiting for a long offline render like sending a photo to a lab and picking it up later. (twinmotion.com) That matters because iteration is where schedules usually get crushed. When a supervisor can move a light, swap a material, or test a camera path and see the result right away, more decisions get made during layout and previs instead of after a heavy render farm run. (twinmotion.com) (youtube.com) Asset creation is also being folded into the same loop. Autodesk said on March 4 that its new Wonder 3D tools inside Flow Studio can turn text prompts or reference images into editable 3D characters or props, then remesh and texture them for downstream use. (blogs.autodesk.com) Flow Studio itself matters here because it was already aimed at automating technical setup around live-action shots. Adding text-to-3D and image-to-3D means the same platform now reaches further upstream, from putting a digital character into footage to generating some of the digital pieces in the first place. (blogs.autodesk.com) The integration story shows up inside the old core apps too. Autodesk’s 2027 releases for Maya and 3ds Max both added an in-product Autodesk Assistant, which means fewer trips out to browser tabs and forum posts when artists are stuck mid-task. (help.autodesk.com 1) (help.autodesk.com 2) Put together, the production shift is less “artificial intelligence replaces visual effects” and more “more of the pipeline becomes interactive.” The March roundup’s tool list — Mocha, ONYX, Flow Studio, Twinmotion, Maya, and 3ds Max — all point to the same destination: fewer handoffs, fewer waits, and more decisions happening while the shot is still fluid. (youtube.com) (blog.borisfx.com) (blogs.autodesk.com)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.