Big Ben's terrain sets

Started by bigben, July 01, 2007, 09:27:30 PM

Previous topic - Next topic


Hi All

[edit]I'm slowly planning an overhaul of my websites which is going to take a while, but I'll use this post to announce additions to my terrains before then. [edit] and while slowly overhauling it, I'm now moving it as well. New terrains will be posted to, but don't bookmark anything yet as the site is in a state of flux. [/edit]

Changes to the previous versions include:

  • Changing terrain projection to UTM to simplify tiling
  • Additional images for creating masks for surfacing or population control
  • Files are now compressed in 7zip format. I use ZipGenius (free) but most other decompression utilities should be able to handle them as well.

I haven't included the "extra-wide" terrains partly for space reasons but mainly because combining the terrains currently requires the terrains to be flattened. This may change later. To give you an idea of sizes, the Grand Canyon set is a total download of around 400Mb (about half of which is taken up by the landsat images). The files are compiled from a collection of around 10Gb of data.

There are no images associated with the 40km terrains because the images for the 80km terrains are the same resolution as the previous 40km images. Other than that, you're pretty much at my mercy for now as to which files/sizes are released. This is the stuff I prepare for my own use.

Additional information is included on the web pages.

Terrain Sets

Before downloading these, be aware that there is some stepping in the very flat regions of these terrains, probably due to my conversion processes. This will probably not change until I can afford a newer version of GlobalMapper. I now have the latest version of GlobalMapper. There may still be some stepping in terrains on my new site but this will be largely due to stepping in the original data.


Posted a test animation of the Grand Canyon terrain (1.8Mb) using a quick modification of a 100m resolution landsat image as a texture map. The terrains are not intended to have the camera this close at the outer extremities but it gives you a good feel for the terrain.

While my original intention was to provide padding for higher camera positions, the ease of positioning tiled UTM terrains like this would also make it fairly straight forward to have high res terrains along the camera path with low res terrains for padding.

Harvey Birdman

Go, Big Ben!


I'll check it out as soon as I get a chance. I'm a bit swamped at the moment.


Ben, this is cool, since it puts the canyon in perspective.  I've been there and it's huge and yet now I see how long it is.  So, it's bigger than I imagined.  It's almost unsatifyingly fast, but that's realistic if we were truly flying over it in a jet plane.  Otherwise, I would have liked to saunter around abit inside the canyon looping around and seeing the sites.  "Hey, look, see the donkeys on that trail down there!"


Great job.  You've learned loads.
So this is Disney World.  Can we live here?


While looking up reference images I read a lot of interesting stuff on theories about how the canyon was formed. This influenced my choice on whether to release the 160 or 205km terrain.  The latter includes a bit extra to the north that really illustrates the river carving through a terrain that is increasing in altitude, and then dropping off suddenly (and then increasing in altitude again).

The animation I posted was only every 1/8th frame of the actual script, but for demo purposes it wasn't really worth doing more.


Added a single SRTM3 terrain of Scotland centred roughly on Ben Nevis. 

Being a single terrain you can turn off "Flatten first" to provide a curved terrain. The "squashing" of the data from reprojecting to UTM has artiticially increased the horizontal resolution which is why I produced a 50m resolution terrain rather then the original resolution of 90m.

Included a link to my reference info from an old Ashunder challenge.


Quote from: bigben on July 15, 2007, 11:36:43 PM
Included a link to my reference info from an old Ashunder challenge.

Perhaps it's time to revisit that now that TG2 is publicly available? If you'll recall Matt came up with a pretty good approximation of the atmosphere using a then-unreleased "TGD" build. It's come a long way since then, in particular the clouds are much better, and if this DEM is free of holes then we might be able to do a lot better...

- Oshyan


The SRTM3 data set has had the holes filled... and from memory there were a lot in this area in the original data. Looking back at the old TG0.9 stuff you get a good appreciation of just how much anti-aliasing and fractal detail alone do for this low res data. It will be interesting to see how the new atmosphere/lighting model can be used to simulate dawn lighting like this.


The new atmosphere should help a good deal along with *GI*. GI alone may mean the difference for realism really as the original was quite subtly lit by indirect illumination I think. Also the automatic detail enhancement/addition on imported terrains ought to help as you said, and yes the original did have a lot of holes back then. So is that a yes to trying this again? ;D I know you're still working on the next iteration of your "Four Seasons" animation as well as providing terrains, utilities and other great stuff for the community. But since there is no contest going on right now it might be a cool thing to start up again. If I find time I'll try to get that going, unless you have the inkling...

- Oshyan


Well there's so much to learn, I'm mainly using TG2 for brain fodder at the moment. I learnt a lot during the original challenge and it was one of the reasons why I made this terrain. It's very applicable to my four seasons project as lighting with a low sun was definitely one of the most difficult aspects of the first animation, and at least I have a reference pic (and a memory of the original situation... or is just that it's going to snow down to 0m today)

I'll also be looking at rocks with lichen for this one as well.

Here's a test render I did yesterday with the fractal detail beefed up just to see what would happen, and a landsat overlay to help interpret the image for masking.


Wow... that snow is incredibly realistic!  I love the overall detail here.  If only you could now apply that level of realism to where the landsat overlay is, and I think you would have a darn good image. - A great Terragen resource with models, contests, galleries, and forums.


Thanks, although the terrain is a little "square" to really show the snow at its best. I used a modified version of cbalaska's snow ( There'll be more snow in the final render to match the reference image, but it should be relatively straight forward to get some realistic distribution with a combination of altitude/slope/face restrictions.


Hi Ben,

The links to the TER and SlopeMap on the Ben Nevis page appear to be wrong, although it's still possible to get to the TER via the directory listing.

Just because milk is white doesn't mean that clouds are made of milk.


Thanks. I'll fix those tomorrow when I'm back at work.


Posted 2 new terrains to complete the 3 Peaks. (Ben Nevis, and now Snowdon and Scafell Pike) Landsat images still to come.

They're only SRTM3 resolution. For now I'm using them to compare terrain displacements on different TGDs at this resolution.

Google Earth provides a good reference for Scafell Pike, but unfortunately Scotland and Wales only have low res imagery. This terrain would also be good for practising some basic surfacing on as it's pretty much just grass, rocks and heather (and maybe snow).. no trees to complicate things ;)

I'm going to repeat some early experiments I did on masking non-flattened terrains using some extra masking tricks I've picked up since then. From memory I managed to at least get 2 terrains merged successfully so this would make it possible for me to add a 400km terrain to each set. If I can get more than two merged nicely then I'm going to go nuts with my other terrain sets but I'll wait and see first  ;)

PS. Added the sea to the TGDs provided to make things a little easier. At least I now know that the sea level is consistent at -0.5 x terrain spacing for my workflow. I've set the water level at 1m above this.