High!
I just programmed a macro to generate simple cylindrical maps from
pigment
patterns... see p.b.s-f..
Here a first demo image, at 500 by 250 pixels, which means 43'12" per
pixel;
to get an impression of the original sphere, I also attached views of
both hemispheres
of the "planet".
It worked well... but at least when using a hetero multifractal function
pattern, its
awfully memory-consuming:
Statistics for E:\Povscn\evaltest.pov, Resolution 500 x 250
----------------------------------------------------------------------------
Pixels: 125000 Samples: 125000 Smpls/Pxl: 1.00
Rays: 125000 Saved: 0 Max Level: 1/5
----------------------------------------------------------------------------
Ray->Shape Intersection Tests Succeeded Percentage
----------------------------------------------------------------------------
Box 125000 125000 100.00
Bounding Box 2009009 125000 6.22
Vista Buffer 7569417 5400035 71.34
----------------------------------------------------------------------------
Function VM calls: 251000
----------------------------------------------------------------------------
Calls to Noise: 1255000 Calls to DNoise: 627510
----------------------------------------------------------------------------
----------------------------------------------------------------------------
Smallest Alloc: 25 bytes Largest: 1004024
Peak memory used: 1409426307 bytes (!!!!!!)
----------------------------------------------------------------------------
Time For Parse: 0 hours 10 minutes 58.0 seconds (658 seconds)
Time For Trace: 0 hours 16 minutes 40.0 seconds (8 seconds)
Total Time: 0 hours 11 minutes 6.0 seconds (666 seconds)
...as you see, this comparatively small map of 125,500 pixels (i. e.
small box objects)
used a cool 1.4 gig, mostly swap space of course.
So I asked first myself and now you whether there is a chance to do this
with an isosurface rather than with "pixel boxes", so that the map is
procedural like
the original pattern...
And, yes, is there a way to get more believable mountain ranges (i. e.
not so
many closed ring shapes, but long ridges, like the Andes on Earth),
perhaps
even with a valley network by using function patterns?
See you in Khyberspace!
Yadgar
Post a reply to this message
Attachments:
Download 'planet_physical_eastern_hemisphere.jpg' (47 KB)
Download 'planet_physical_western_hemisphere.jpg' (56 KB)
Download 'cylmap.jpg' (42 KB)
Preview of image 'planet_physical_eastern_hemisphere.jpg'
Preview of image 'planet_physical_western_hemisphere.jpg'
Preview of image 'cylmap.jpg'
|
