View Full Version : not enough detail
09-19-2006, 10:16 AM
I'm stuck on this for a few days now, so I finally decided to ask for your help. I'm doing a landscape render (aerial view +/- 30m above the ground, so there's a need for quite some detail) of a site which is about 2km to 1.5km... Nevertheless there are a lot of small details that have to remain vissible like little canals, slopes, dikes... I started to model the terrain in 3ds max, put a gradient map on it, so I can render a heigtmap of it... I finish the heightmap in photoshop (since there are some imperfections in the 3ds model) and use it as the base for my terrain in vue. No matter what I try, the canals keep looking 'pixelated' when I render them. I cracked up the resolution of the terrain mesh to 8000x6000 (resolution of the heigtmap is something like 8000x6000 pixels too). Here are some screenshot and testrenders to illustrate the problem.
09-19-2006, 12:51 PM
Check out this thread and see if it helps.
Check this tutorial..Possibly another way...
09-23-2006, 11:18 AM
first of all: The terrain should have a height resolution of 16bit, and it must be a tga file, because Vue can only import tga files in 16bit. (not the normal image tga, but the pov-ray)
I think, constructing the terrain in 3DSmax is not a good idea, because you can not get the 16bit height resolution out of it.
But if you have done it so, use only the special sizes, 1024,2048,4096 and squared. When you import the heightfield, do not forget to change the grid before the import to the image size.
I think, this simple canals should be no problem, and the image looks, like you have not set the grid before the import.
In general I would recommend, to use special tools like GeoControl or Leveller to create the terrain. Both have tools, to manipulate the terrain by hand.
09-23-2006, 06:08 PM
Wow, your explanation looks very promissing, but there are a few "why?'s" left :)
Why do i have to use the pov-ray tga? What's the difference between that and the 16-bit tga I get when I save my render in 3ds?
My 3ds-workflow to create the grayscale image is: modelling the terrain; Apply a uvw-modifier and apply a gradient ramp going from black to white as a texture.
Why using the special sizes? And why squared?
By changing the grid, do you mean when having a 1024x1024 image, the grid size should be 1024x1024?
Thanks for the tools-tip. I'll take a look at those.
I really appreciate your help, and it's not that I don't believe what you write is right... I just try to understand "why" you do all this. Thanks again!
09-24-2006, 07:01 AM
The tga: TGA is a normal image format, what means, every 8 bit for green, red and blue.
The pov-tga is defined as 16 bit heightmap, where the red and green channel is used. 16 bit means 256*256. So if there is a height of lets say 16000, then the 16000 is devided by 256. This result is put in one channel, and the difference between the 16000 and the division in the other. I do not know, if the 16 tga of 3DSmax use this encoding, if it does, you can use it in Vue, else you will get staired results.
The sqared format is typical for terrains: they are generated by a subdivision process. So most terrains have this squared row of pow of 2. If you take another size, the heightmap is resized to fit to the Vue sizes. You have no control over the resizing algorithms, and so no control over the results.
Grid: Exact, the grid is most important for the details and should be the same as the heightmap, else the heightmap is resized.
Developer of GeoControl (http://www.cajomi.de/GeoControl/geocontrol.htm)
09-24-2006, 07:01 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.