VFX and compositing

Tracking, segmentation, roto, keying, simulation and compositing

• sourceViPE: Video Pose Engine
• sourceCorridorKey (neural green screen keyer)
• sourceLayerDiffuse (transparent image layer diffusion)
• sourceDiffuEraser
• sourceComfyUI-segment-anything-2 (SAM 2 nodes)
• sourceCutie: Putting the Object Back into Video Object Segmentation
• sourceTAPNet: Tracking Any Point (TAP-Vid, TAPIR, TAPNext)
• sourceSAM 3: Segment Anything with Concepts
• sourceSAM 2: Segment Anything in Images and Videos
• sourceCoTracker / CoTracker3
• sourceVideo Depth Anything (CVPR 2025 Highlight)
• sourceOpenFX image processing plug-in standard
• discussionWhen does video-to-video fail and image-to-video from a plate frame succeed for VFX inserts?
• discussionCan a fully commercially usable AI roto/track/inpaint stack be assembled from permissively licensed parts?
• discussionWhat colour and alpha contract should AI VFX tools use at the EXR boundary?
• discussionDo diffusion inpainters beat propagation methods for clean plates on moving cameras?
• discussionHow should AI roto/matting outputs be validated against production-grade mattes?
• sourceUnreal Engine: In-Camera VFX overview (official docs)
• sourceOpenImageIO (OIIO)
• sourceHow to Turn Your Own Footage Into Cinematic VFX with Seedance 4K
• sourceNetflix's Moria hybrid campaign by Vertex CGI with Higgsfield
• sourceHow BITT Animation Runs Hybrid AI Production on Higgsfield
• sourceAI vs VFX: Can Seedance 2.5 beat VFX? (full workflow)
• sourceNatron
• sourceOpenVDB
• sourceOpenEXR
• sourceLevector's Tango 'Libera tu cocodrilo': live shoot plus AI crocodile characters
• sourceDouble Shots produces music videos for pop idol Ivy with Cinema Studio
• sourceOpenCue (render management system)
• sourceThe VES Handbook of Visual Effects, 4th edition (previs chapters)
• sourceSasha Kasiuha on why the image still needs an author
• sourceFull AI Car Commercial Tutorial (Higgsfield)