Motion Depth Blur

The motion depth blur plugin simultaneously applies the velocity and z-depth pass generated with 3-D renders.

Overview

The main problem with motion and depth blur in 2-D is that each operation invalidates the velocity and z-depth information generated with 3-D renders. This is due to the fact that certain pixels will no longer be associated with their origin z-depth and velocity as these values may have been blended with others. To avoid the invalid data, both motion and depth blur can be applied simultaneously using the value assigned to the original pixel by the renderer. This produces a much more accurate result than two separate tools.

Limitations

Unlike 3-D motion blur, 2-D motion blur is working with partial data since nothing is known about pixels behind pixels. So if a fast moving object starts in front of a stationary object there will be gaps in stationary objects surface. This means motion blurring in 2-D is best when motion is small or objects are rendered to separate images and merged after motion blur is applied.

Also since 2-D motion blur uses only a single velocity value for each pixel while acceleration is ignored. This may cause problems if an object has a rapidly changing velocities.

Inputs

• Input a velocity image or uses the channels embedded in the image
• Control the scale of motion in the x and y direction
• Handle antialiased depth maps using the alpha channel from a rendered image
• Process normalized or absolute pixel velocities
• Control depth blur using the same controls as Depth Blur
• Disable depth blur and only apply motion blur

Results

A sample input that contains a depth and velocity channel.

The motion blur resulting from the moving teapot in the center of the image.

The motion and depth blur are applied to the scene with focus placed on the bottom left object.