Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[mikemayday]

Hello!
Being both a huge fan of mountains and Outerra, I offered some time ago to provide some general ideas on how a little bit of geology might make the Outerran landscapes a lot more interesting. Small disclaimer: it is not my intention to stress the shortcomings of the current landscape. I actually think lakes and rivers are a priority rather than geology. So all the comparisons are stricly for clarity's sake. The ideas below pertain to our good old alpine biomes - like the Tatras and, of course, the Alps.

I believe just these four elements would take us from 10% to 90% in terms of satisfaotry mountain simulation:

gully (couloir)
arête
Colluvial (scree) cone
rock face

Obviously right now Outerra takes the smooth heigtmap, trows some noise, sharpening, rocks on top of it and places grass where appropriate. This creates rather uniform looking mounds. Here's a comparison of a photo I managed to take in the Tatras with the same place in Outerra.


I've outlined the basic formations on the photo: (slope is basically where the surface stays the same as in Outerra):


Obviously the "rock faces" (for lack of a better word) and gullies are the more frequent the higher the steepness of the slope. Colluvial cones, on the other hand, gather at the local bottoms - they are VERY frequent and there is always a distinct border between them and the rock faces that they always neighbour.

One more comparison for good measure:


Here's some colluvial cones and gullies in a non-alpine environment:

This shows very well how the tip of the cone is always in a gully - because the formation of the latter creates the former.

And here's a nice photo of the Slovakian Tatras:


It looks vary chaotic and varied, but if you examine it closely in terms of shape, the difference between Outerra and reality are the four elements I've mentioned: gullies, colluvial cones, rock faces and the arrete - the sharp line of peaks at the top. The rest is vegetation.
EDIT: well, Outerrra does actually create arretes to some degree. So there's that!

Some additional info:
Slopes with northern exposure get less sun and therefore they have less vegetation and more erosion (more gullies, colluvium and rock faces).
Colluvial cones can therefore be bare (lots of rocks and small boulders) or covered with soild and small vegetation.
The other factors are the local climate, local height and geology. The higher and colder, the less vegetationd and more erosion - formations will be sharper, larger, more frequent.

Naturally, different geoligical struture (and orogenesis) will produce different landscapes - like the Grand Canyon or the Chinese stone forests. However, simulating the generic alpine mountains seems like a good first step to me. What do you guys think?

[KW71]

"Obviously right now Outerra takes the smooth heigtmap, trows some noise"

Actually OT does way more than just that:

http://forum.outerra.com/FMX/presentation/

[cameni]

Probably the simplest way to achieve better looks is using more detailed elevation sources, that will provide slope variety. But looking at the pics, I'm thinking how we could generate the regions procedurally, to make up for the missing detail on those continuous slopes coming from the smooth base mesh interpolation. Maybe some kind of delaunay mask to parcel it and bias the noise amplitude ...

[mikemayday]

KW71 - well, as far as I can see the thing I didn't mention are the addition of rocks and peaks by Outerra, but those are done randomly (with adjustable parameters) based only on the local steepness and global height, not taking into account the surroundings.

cameni: Yeah, I think being able to achieve realistic landscapes from 90m samples would make Outerra a much more powerful tool. I have, sadly, no technical knowledge on the subject (neither from the geological nor from the programmer side).
But gullies can be placed pretty much randomly (based on local parameters of course, as written above), and then you place a cone at the _local_ bottom with its tip at the gully- the larger the gully the larger the cone. Doesn's seem like much of a problem on what and where to place - so coding it would be the entirety of the job - however difficult, I cannot tell.

[bomber]

I think this is a very good presentation.... Well done.

I also think that increased hight maps won't solve the problems described... I do think that logical procedural code is the right way forward and I think on having solved it once it would become quickly apparent where with a bit of tweaking it could be used elsewhere like for instance sand dunes in the desert...