Hi,
What is the best method to create spherical terrain?
Thanks
Posted 23 January 2013 - 09:23 AM
Edited by C0lumbo, 23 January 2013 - 09:38 AM.
Posted 23 January 2013 - 09:43 AM
My favorite is a sculpty tool, which every good 3D program has. This makes work fast and accurate with some practice. I have occasionally been forced to use a plane mesh and begin with extrusions - sometimes even having to measure altitudes - and refining that with the sculpty tool. For each terrain element to look completely unique and finish the map in a reasonable time, then a sculpty type of tool is a must, preferably within a terrain editor.
Doing simple work by coding is doable but takes more time initially. Once you get a noise algorithm going, then it can actually be very effective visually, especially if you are skilled in shaders. It can also cover many square miles (or kilometers) in short order. However, highly detailed work is many multiples more time consuming to make terrain, especially if it must be hard edged ( can handle full physics ), by coding instead of using a 3D program such as a terrain editor.
Edited by 3Ddreamer, 23 January 2013 - 09:44 AM.
Posted 23 January 2013 - 10:12 AM
There is a lot of warping (the grid shape and area is smaller at the corners of the square than at the centres), but that is tolerable for my particular project.
Are you using the improved equation, to minimise warping?
Tristam MacDonald - SDE @ Amazon - swiftcoding [Need to sync your files via the cloud? | Need affordable web hosting?]
Posted 23 January 2013 - 10:20 AM
Edited by eppo, 23 January 2013 - 10:23 AM.
Posted 23 January 2013 - 10:39 AM
There is a lot of warping (the grid shape and area is smaller at the corners of the square than at the centres), but that is tolerable for my particular project.
Are you using the improved equation, to minimise warping?
Nope, I'm just doing the obvious normalise-to-distance.
Will give that a go, thanks!
Posted 23 January 2013 - 10:53 AM
Posted 23 January 2013 - 02:38 PM
Have a look at this: http://freespace.virgin.net/hugo.elias/models/m_landsp.htm
It describes creating a randomly shaped planet.
What a fascinating article, got me very intrigued for a while there, thanks. ![]()
Posted 24 January 2013 - 04:13 AM
There is a lot of warping (the grid shape and area is smaller at the corners of the square than at the centres), but that is tolerable for my particular project.
Are you using the improved equation, to minimise warping?
Just tried this out. The linked article didn't have a side-by-side comparison of the naive approach (normalising each vertex) and their improved approach so I'll post a couple of screenshots so the improvement can be seen in case it's useful to anyone.
Here's the naive approach:

And here's the improved method:

So I'm quite happy with that result!
Posted 24 January 2013 - 09:26 AM
So I'm quite happy with that result!
Aye. It's a fairly subtle change, but it is a very big improvement in reducing distortion.
Tristam MacDonald - SDE @ Amazon - swiftcoding [Need to sync your files via the cloud? | Need affordable web hosting?]
Posted 24 January 2013 - 11:40 AM
Have a look at this: http://freespace.virgin.net/hugo.elias/models/m_landsp.htm
It describes creating a randomly shaped planet.
I have to try that one out sometime. What I have to wonder is with all the resizing of the slices, how does it take care of the problem of empty spaces that are exposed from resizing the slices.
Posted 24 January 2013 - 01:20 PM