In Katana we have a visualization mode for depth (which is invoked on a button push).

- sets the input colorspace to the space defined by ROLE_DATA

- scans the image for min/max depth values in a depth-specific manner.
(ignore depth < 1e-3 are ignored. ignore depth > 1e10  for f32. ignore nan/infs.  for f16 adjust maxdepth to 65K)

- our monitor supports a post-display transform, which allows for a view-min, view-max, and view-gamma

We set the view-gamma to 1.0, map the min-depth to 1.0, and the max-depth to 0.0.   This 'inverted' view matches the historical result of tools such as prman's 'sho', which does a decent job at depth visualization.

-- Jeremy