ANR Normals Prep

Overview

Normal passes come in many forms: packed 0 to 1 or signed -1 to 1, camera or world space, Y up or down, Z up or Y up. Normals Prep reads the pass on its Normals input, says in Status how it read it, and hands on one form: Nuke camera space, X right, Y up, Z toward the camera, -1 to 1. That is what ANR Normals Matte and ANR Normals Light want.

Detect and Lock works out the encoding and which axes point the wrong way. A world-space CG pass is turned into camera space with the Camera you name in Camera Node, frame by frame, so an animated camera works.

The Clean Up section pulls soft AI normals onto the plate's edges, smooths their grain without crossing outlines and steadies flicker. Depth and Detail can take the large shapes from an ANR Depth Prep and the fine detail from the pass, or add fine relief read from the plate's own light and shade.

With Normals From set to Depth Input the node makes normals from a Depth Prep's distances, and with Plate Only from the picture alone. Output also offers checking views (a clay render, an orientation view), a curvature pass and a height map.

Inputs

1. Sourceoptional

   The plate. Edge Snap, Smooth and Plate Detail use it; without it those are off.

2. Normalsoptional

   The normal pass to read: ANR Depth and Normal Creator with Output on its Normals choice and Write To on RGB, a CG render's N shuffled into rgb, or any normal map. Not needed with Normals From on Depth Input or Plate Only.

3. Depthoptional

   An ANR Depth Prep (Output on Depth). Used for Normals From Depth Input, for Shape From Depth, and by Detect and Lock to check every axis, X included. Its lens is read through the wire.

4. Motionoptional

   Motion vectors in rgba (forward u v, backward u v), for example ANR AI Vector Tracker with Output on Motion Vectors, so Deflicker follows moving pixels.

Output. With Output on Normals (the default), camera-space normals in RGB, -1 to 1. Other choices give them packed 0 to 1, in world space, or as a clay render, orientation view, curvature, height map or the plate's detail alone.

Quick start

1. Connect the plate to Source and the normal pass to Normals. For better axis detection, also connect an ANR Depth Prep to Depth.
2. Press Detect and Lock and read Status: it says the encoding it found and any axes it flipped.
3. Set Output to Clay to check the normals by eye: the grey render should look like the plate's shapes lit from the front left (Clay Light Azimuth and Clay Light Elevation move that light).
4. For a world-space CG pass, set Space to World Space and type the name of the render Camera into Camera Node.
5. For AI normals, keep Edge Snap and raise Smooth a little if they look grainy; set Edge Snap to 0 for a CG pass.
6. Set Output back to Normals and connect the node to the Normals input of ANR Normals Matte or ANR Normals Light.

The panel

The node's panel as Nuke shows it, with the values of a new node. Click any control to see what it does; tabs and sections work as in Nuke.

All controls

Tips

• Normals from ANR Depth and Normal Creator are already signed and in Nuke's camera space: Encoding Signed (-1 to 1), Space Camera Space, no flips.
• Without a Depth input, Detect and Lock can settle Y and Z but not X. Check that walls on the left of frame key as facing right in ANR Normals Matte, or connect a Depth Prep and press Detect and Lock again.
• Output > Orientation shows sideways surfaces red, upward ones green and those facing the camera blue: a quick check that floors read as up.
• Shape From Depth with a Depth Prep keeps the pass's fine detail but takes the large, often wobbly, shape from the depth; Shape Size sets where shape ends and detail begins.
• Plate Detail adds fine relief such as skin, bark, fabric or brickwork from the plate's light and shade; turn on Dark Is Raised when cracks and pores are dark.
• For flickering per-frame AI normals, raise Deflicker, and connect motion vectors to Motion on moving shots.

Good to know

• The node runs no AI model and needs no graphics card.
• World Normals output only differs from Normals when Space is World Space, turned with the Camera in Camera Node or with Camera Rotate.
• Height treats the frame as seen head-on: good for relief on surfaces facing the camera, not a depth map. A floor seen edge-on comes out as a cliff.
• Deflicker and Shape From Depth work on the Normals input only, not with Normals From on Depth Input or Plate Only.

Works with

ANR Depth and Normal CreatorMakes a depth pass, and with MoGe-2 also normals and positions, from any plate with AI, straight into depth.Z, N or P with the plate untouched.ANR Normals MatteMattes from which way surfaces face: every floor, every wall facing left, rims turning away from camera, edges and creases.ANR Normals LightRelights a plate through its normals: up to three new lights with shadows, an HDRI environment, a matcap, or refraction of a background through the surface.ANR Depth PrepTurns any depth map, AI or CG, into one clean, calibrated distance pass that the other ANR depth nodes read, with edges snapped to the plate and flicker removed.ANR Interactive LightLights an element with footage of a fire, lightning or a screen, through its normals and depth, with shadows from its own shape or from smoke.ANR AI Vector TrackerAI motion vectors for every pixel of the plate: live for vector blur and retimes, or baked to smart vectors and ST maps that carry a paint fix through the shot.