No, the alpha channel was the depth cue.  The other channels were the actual image, i.e. object brightness was affected by texture, lights, etc.

Now, if only I could get it to work beyond the simple cases.  Maybe this is just a wish-list for Poser 8.