From VFXPedia

Auxiliary Channels

Z-Buffer

Each pixel in the Z-Buffer channel contains a floating point value that represents the relative depth of that pixel in the scene. A Z-buffer channel can be used to perform depth merging, depth of field and fogging. Normal merging requires one layer to be selected as the foreground and another as the background. Depth merging uses the Z-buffer value for each pixel to determine which is in front of which.

The drawback to this approach is that an individual pixel can only have one depth value, that of the frontmost object touching that pixel. If the pixel has color data for two items at different depths in the scene, there may be a fair bit of aliasing in the merged image.

Most 3D applications will take the depth value from the object closest to the camera when two objects are present in the pixel.

Coverage

The Coverage channel is used to indicate what fraction of the pixel is covered by the frontmost object touching that pixel. It may appear similar to the alpha channel, however transparent areas of the object will still have full coverage, and internal edges (where an object overlaps itself) may be opaque but only have partial coverage.

Depth composites in Fusion that are based on images that lack Coverage and Background RGBA channels may not be properly anti-aliased, except in very simple cases, as these channels provide information that is essential for processing foreground objects separately from the background.

Background RGBA

This channel contains the color values from any object behind the foreground object, as described for the Coverage channel.

This channel is needed (along with Coverage) to separate the foreground object's color from the background object, to allow clean depth merging or masking of an object.

Object ID

Several 3D applications are capable of assigning integer ID values to objects in a scene. Each pixel in the Object ID channel will be identified by that ID number, allowing for the creation of masks and mattes.

Material ID

Several 3D applications are capable of assigning integer ID values to materials in a scene. Each pixel in the Material ID channel will be identified by that ID number, allowing for the creation of masks and mattes based on textures.

Unlike the object ID channel, most of the 3D applications that support material IDs will allow segments, or portions of an object, to receive differing material ID values.

U Texture and V Texture Channels

These U Texture and V Texture Channels contain information about mapping coordinates for each pixel in the scene. This is used to properly apply warping to images and textures wrapped around the object.

X Normal, Y Normal and Z Normal Channels

These X Normal, Y Normal and Z Normal channels contain information about each pixel's orientation (the direction that it faces) in 3D space.

X Normals represent the horizontal component of the pixel's normal vector. Y Normals contain the vertical component. Z Normals contain the perpendicular component (in or out of the screen).

Pixel Velocity

The Pixel Velocity channels indicate the vector of a pixel's motion from frame to frame. They can be used to apply a fairly realistic motion blur to an image as a post process without having to resort to slow optical motion analysis. These channels are used by the Vector Motion Blur and Vector Distortion tools. There is also a third-party plugin available that can use this data (visit www.komkomdoorn.com).