Multi-Frame Texture Bake

Requires the Max or Studio license

Multi-Frame Texture Bake fuses visible observations from one or more tracked camera plates into a permanent packed texture for static reflection, lighting, and shadow-catching geometry.

How to use it

  1. 1 Choose the receiving surface and a valid texture UV map, or enable Auto Smart UV.
  2. 2 Use the current camera or build a list of enabled cameras with independent footage slots, ranges, and frame steps.
  3. 3 Choose New / Replace or Add to Existing.
  4. 4 Set texture resolution, Smooth Blend or Best Frame, edge feather, view-angle and distance bias, self-occlusion, and padding.
  5. 5 Choose Lit Base Color or Exact Emission, then run the bake.
  6. 6 Save the blend file after the generated texture is packed.

Multi-Frame Texture Bake Panel

New in 1.0.7 · Max

Camera projection is only correct from the projecting camera's position. This panel goes a step further: it accumulates observations from one or more animated cameras (for example twelve camera positions along the track) into one persistent, packed surface texture. The surface then keeps the correct texture no matter where the camera moves, which is what makes accurate reflections and shadow catching on tracked geometry possible.

Main Controls

  • •
    Sources: the current camera, or an explicit camera list where each camera has its own footage slot, frame range, and step.
  • •
    Accumulation: New / Replace, or Add to Existing to extend a previous bake. Duplicate camera/frame evidence is never counted twice.
  • •
    Blending: Smooth weighted blending (footage alpha, image-edge distance, view angle, camera distance, optional self-occlusion) or Best Frame, which keeps the strongest observation per texel and reduces ghosting from moving subjects.
  • •
    Bake: the large bake button near the top; the result is a packed image, so save the .blend afterwards.

Extra Settings

Texture atlas options (automatic Smart UV generation, 256 to 4096 resolution, edge dilation), frame blending weights, permanent material output, and baked-output handling. Coverage blending keeps a controlled fallback for texels no camera ever saw: authored materials stay authoritative for uncovered regions, and when the mesh had no material, Automatic Base Color uses the coverage-weighted average of the observed texture (a manual Unseen Base Color is also available). Lit Base Color materials default to Roughness 1.0 and Specular IOR Level 0.0. After a successful bake the receiving mesh becomes a Cycles Shadow Catcher with Shadow ray visibility disabled (both with explicit opt-outs), and Remove Baked Texture restores the original materials and flags. A failed bake keeps the last successful bake and its object state instead of destroying it.

Typical flow: Heavy track → Point Cloud Surface (or 3D Point Generator → surface) → bake from the tracked camera across the range → place CG on the surface and render with real reflections and shadows. Full guide: Multi-Frame Texture Bake.

Technical notes

Smooth Blend weights footage alpha, image-edge distance, surface angle, camera distance, and optional self-occlusion. Best Frame keeps the strongest observation per texel and can reduce ghosting from moving subjects. Duplicate camera/frame evidence is not counted twice.

Lit Base Color defaults to Roughness 1 and Specular IOR Level 0. When the receiving mesh had no material, Automatic Base Color uses the coverage-weighted average of observed texture pixels rather than unseen atlas pixels.

After a successful bake, the receiver becomes a Cycles shadow catcher and Shadow ray visibility is disabled by default. Removing the baked setup restores the prior materials, shadow-catcher state, and ray visibility. A failed replacement leaves the last successful bake intact.

Docs Camera Footage Projection

Camera Footage Projection creates a live, non-destructive front projection from a tracked camera background image, sequence, or movie clip onto a mesh.

Docs Point Cloud Surface

Point Cloud Surface reconstructs a controllable 2.5D mesh from a Blender Point Cloud or mesh vertices. It is intended for floors, roads, walls, facades, and other surfaces that can be represented as one depth over a fitted plane.

Docs Planar Tracking

Planar Tracking solves the perspective motion of a flat surface such as a screen, sign, wall, card, or set-extension plane. You draw an editable four-corner quad over the surface in the MotionMaster 3D Track Workspace, track it forward or backward, correct drift non-destructively, and publish a native Blender Plane Track, a compositor Corner Pin, and a 3D plane or an animated camera.