I've been researching a 'displacement layering' technique. Most TGTP images I see (including my own) still have that characteristic smooth look on rocky surfaces. Hopefuly this will be a way around it.
Here are the results. Its nothing fancy, just a proof of concept image. There are 3 layers of displacement here, each with a different scale.
EDIT: I just noticed there are some shadowing problems, likely the known displacement shadows bug.
Nice,
I love the detailing in the rocks, cant wait too see where it gose
Nice work,
Zionner
yes, the left side really has me impressed, great job :)
there are some shadow issues on the left side of the rock....I am currently doing the same thing on fake stones....
very nice job!
good on close look but not when you are looking far away
Uh, yeah. Now I see where you've been. Great work!
Quote from: dhavalmistry on November 27, 2007, 05:57:20 PM
there are some shadow issues on the left side of the rock....I am currently doing the same thing on fake stones....
very nice job!
yea, I noticed that... its got to be a bug.
This looks great. Great job with the lighting, surfacing, and displacement.
Hey, your planet looks very good as well.
hey...did you use functions for this??
God dhavalmistry why would he use functions? The very nerve
edit: Eh what happened to MF's post?
Quote from: dhavalmistry on November 28, 2007, 03:30:38 PM
hey...did you use functions for this??
NO! >:(
Just kidding! ;D
Actually I used both powerfractal and image shaders, though they could be done with functions as well, or any of them by themselves. In this image, the displacements are mainly image map shaders (since I can make exactly what I want in Photoshop in mere minutes), and the colors were from power fractal shaders.
There are 2 key points to the theory:
1. Using the general to specific rule: Start out with larger scale displacements, and then step down to smaller and smaller displacements. The way I have the nodes setup, the displacements are cumulative.
2. Use negative displacements to "cut away" or "depress" the surfaces, which seems to work very well. However, positive and negative can be used if desired. The fun part is watching your rocks and surfaces change as you dial them in!
Unfortunately, I am stuck at work and don't have a clip file to post. I'll post one when I get home.
My next tests are using the "compute normal" after each displacement (to reset the true surface), but this is killing render times as expected. But I'd rather longer render times for more realism.
Quote from: Will on November 28, 2007, 05:22:45 PM
God dhavalmistry why would he use functions? The very nerve
edit: Eh what happened to MF's post?
Sorry will, I deleted that post as I accidentally posted it before I was done, LOL. The new one is just below yours.
I used this same technique with clouds last night and the results seem to have promise.
Cool. I've been working on texturing as well. This is good work...again, I say.
Quote from: moodflow on November 28, 2007, 05:37:40 PM
I used this same technique with clouds last night and the results seem to have promise.
Quote from: moodflow on November 28, 2007, 05:33:00 PM
1. Using the general to specific rule: Start out with larger scale displacements, and then step down to smaller and smaller displacements. The way I have the nodes setup, the displacements are cumulative.
you stole my technique...you THIEF!!....lol....just kidding...I started using that after Volker's tutorials on powerfractals....
Quote from: dhavalmistry on November 28, 2007, 06:46:01 PM
Quote from: moodflow on November 28, 2007, 05:33:00 PM
1. Using the general to specific rule: Start out with larger scale displacements, and then step down to smaller and smaller displacements. The way I have the nodes setup, the displacements are cumulative.
you stole my technique...you THIEF!!....lol....just kidding...I started using that after Volker's tutorials on powerfractals....
Heheh yea, but I figure we'll all stumble upon the same new techniques as time goes on. ;)