I've seen tutorials on how to split 16bit displacement maps into two halves and then recombine them inside of P6 to create true 16bit displacement - Stonemason is one of the particular few out there who uses this to an advantage right now. I'm not sure exactly how it's done, so I'm not sure how stupid this next question is going to sound. Keeping that in mind, is it possible to somehow use the same technique to bump P6's own procedural nodes up to doing 16bit displacement?

Procedurals are all 32bit floating point to begin with, so no tricks are required here. Plug 'em in and render!

Message edited on: 03/14/2006 21:58

Hmm....So if P6 can handle 16bit displacement through procedural nodes, what's the issue with using image maps in 16bit?

Image maps are loaded in at 8 bit resolution. This means that there are only 256 possible values for the displacement map leading to "ridges" in the bumps. By using 16 bit displacement maps - there are a possible 65K values - in other words, a more detailed displacement. For some displacement maps - 8 bits is fine. However, the displacement maps that give you wrinkles, pores, scars, etc need the 16 bits to give you the full range of displacements without looking chunky. --EK

Alright. I understand that part....but what I don't understand, is why would the designers build an engine that's able to use 16bit displacement, but then only load image maps at 8bit resolution?

Because most image formats are only 8 bit to begin with. 16 bit formats have only recently (in the context of how long Poser has been around) become popular. Hell, even GIMP only supports 8 bit atm. --EK

The trick is actually quite simple. If a 16bit displacement map is converted to a color map, using the lower 8 bits for pure red and the higher 8 bits for pure green, you can use Component nodes to separate the two bytes. Multiply the output of the green component by 256, add the output of the red component, and voila, a 16 bit displacement map. You could even go to 24 bit by also using the blue channel.

"You could even go to 24 bit by also using the blue channel." okay that is an interesting idea. what would the math nodes be to do that, and how do you seperate out the 3 channels. Do them in Photoshop??

Attached Link: http://www.stewreo.de/poser/scripts/loadZBrushDisp.py

how do you do it without Zbrush tho?.....

Attached Link: http://pixologic.com/zbrush/class/roadmap.html

aargh I don't have it ethier.. I was trying to find out how to do it without zbrush but using Photoshop...

There is a trick in using Photoshop and RAW format files to get what you're after Khai. I think Quinlor came up with it first, but I'm not sure. It's hiding somewhere in the Rendo archives. --EK

the RAW format never worked for me : & I've never heard of sydl's method,I'd like to see it in action if anyones tried it. if you dont have zbrush: Get-jolly mentioned another technique in an IM,I've yet to try it myself but the JollyTroll is using 16 bit displace & was made without Zbrush. "In your friendly paint/compositing package you can separate out the displacement maps, by clamping the levels, the two maps can then be recombined in Poser."

Gif format. Most Gif tools are capable of breaking down into 8 bit format, which poser supports for displacement maps.

gif? 8 bit is no good for displacement

PNG format is the one to use really with poser 6.

Attached Link: http://www.renderosity.com/messages.ez?Form.ShowMessage=1740708

I'm afraid that with Poser 6 I've had some odd things happen with PNG files, which look to be a failure to properly transfer from bitmap proportions to square UVspace. It may be down to the paint program I used, but I never had the problem with Poser 5.