Depth information from 2.5D map has too large of extremes.

[rygoody]

So here is a texture I sculpted in 2.5D

[] If you notice, the vertical collums are pure white. I need to get that fixed so the vertical collums still have detail in the height map. It is most likely doing this because the vertical collums are on a separate layer and the vertical collum layer is much more pulled out than the background layer. So all the details get condensed into white. I tried to use the 'displace Z' function in the layers menu to make the vertical collum layer as close to the background layer as possible and this had absolute zero effect on the height information. So I don't know what is going on there. Any ideas?

[aurick]

You’ve identified the reason. And you’re using one possible technique for the solution. Most likely, you just haven’t displaced Z enough. Move the Transform palette to one of the trays and open up its Info menu. Now mouse over the canvas so that you can see the depth of the back layer. Then mouse over the lowest edge of the front layer. The difference between the two values will tell you how much you need to displace Z.

Another option is to do two Grab Doc operations – One for each layer while the other is turned off. You could then combine them in Photoshop.

[marcus_civis]

You could also use the MRGBZGrabber tool to grab the canvas. This will grab all the layers at once.

EDIT: actually that doesn’t. I can’t think why I thought it did… :o

[rygoody]

even when generating an alpha for each individual layer it still comes out all white

and I used the z displace to move all the z values between 0 and 30, it still came out solid white

this seems almost like some kind of bug…

or is there another setting anywhere to adjust?

[marcus_civis]

I’m not sure what is happening here. Can you share the document?

[rygoody]

Here is the file
http://www.4shared.com/file/233311219/257ff418/cememtMetalVerticalA.html

[marcus_civis]

Thanks, that’s useful.

The two vertical bars are quite far apart in depth (as well as far forward) and that throws the whole thing out further. Although the depth information is all there it is concentrated at each end of the grayscale so looks almost complete white or black.

To get a better spread for the values I duplicated the vertical bar layer then erased the extra bar on each so that I could adjust the z depth separately.
You can get the file here:

http://www.benbus.co.uk/tmp/cememtMetalVerticalA_rev.zip

To get a good depth map you will need to select Layer 2 then merge before doing a GrabDoc. Grabbing each separate layer won’t give as good results as the clipping plane is taken into account in the vertical bar layer. In your original document this is what makes the bars nearly pure white, as the limit of what can be stored in a 16bit grayscale is reached. The depth info is still there but is degraded - you can see this if you do an Alpha>CropAndFill with the Basic Material.

HTH,