Quote from: Volker Harun on August 07, 2007, 04:29:58 PM
Postwork:
I reduced the 32-Bit OpenEXR to 16 Bit (Thanks BigBen).
I used some slight sharpening (unsharp mask 95%; rad. 0.5)
Copy of the Background, gaussian blur, radius is 1% of image width. Set to soft light.
Curves to adjust brightness.
You might be interested to try a sharpening technique I picked up out of a graphics mag one day.
Duplicate Background, set to hard light, high pass filter. Reduce the high pass filter radius until you can just see some edges in the filter preview (or check the image preview). I tested a while ago on aTG2 image and found a radius of 0.5 was enough. (anything less than 1 means the image was pretty sharp to begin with, so I don't bother much)
And back on topic
Here's a render test from my experimentation with jmeyers' clip last night. A few observations:
Most of the wonderful clips here are created on fractal terrains. When applying them to TERs with fractal detail added things can get messy because the fractal detail is a lot more complex than you average power fractal node. Things usually get screwed up on the cliffs, where these clips are at their best.
The image below includes a TER without fractal detail, a couple of low displacement power fractals and a strata shader before jmeyers' clip. Adding the extra power fractal helped a lot and is one of the key tweaks.
The other problem I had was when the second surface of the clip was on its own. (see the foreground of the last two images in
my previous post) I brought their slope distributions closer together and duplicated the fractal breakup of the first to use for the second, using differences in the colour offset to give the second surface a slightly greater distribution.
So far so good, and I haven't touched the scaling yet to really screw things up
I'll have another go tonight and post a TGD tomorrow.
PS. this clip makes great outcrops if you let it loose on slopes down to 15°. not so good for my terrain, but someone else will have a ball with it.
[edit] Added 2 more images of the latest version I had on the train this morning with a slightly different approach to using similar distributions for the two surfaces. Added a view of the north face which is probably the steepest slope in the terrain. This is the first time I've looked at a steep slope with this, and it looks quite promising.