View Full Version : WIP returned from the dead-Tavonni Arise

06-22-2009, 03:15 AM
So I'd been working on the Tavonni map by David "Hyphen" Jaques-Watson on the COTI list (http://www.travellerrpg.com/CotI/Discuss/) for a very long time when a hard drive death destroyed most of my work. I thought all I was left with was a questionable icomap projection and I had lost the all plate carée working maps I needed. I was very sad.

Today, as I was prepping a computer for sale, I found a treasure trove of Tavonni images on an old drive. These included the flattened version of the Tavonni map created by The Engineer and a lot of my early work. Sadly all of my recent experimentation with Wilbur died with the hard drive, but I needed to start over anyway(sour grapes?).

With all of that out of the way I'd like to get back to work on Tavonni, starting with this climate-modified variation on one of my earlier layer effects maps. I'll probably use this as a guide in Wilbur to build an HF.

See what you think.

06-28-2009, 11:30 PM
Got it into Wilbur. Some initial progress was made.
I saved selections for land, sea and some highland areas from photoshop.
After restarting in Windows(spit), I loaded my selections into Wilbur and used them to set initial elevations(Filter>Fill>Set Value...). I added a few higher points here and there in an unsuccessful attempt to break up flat areas. I know my error there, I shouldn't have used a round brush. Finally, I deterraced the entire mess(Filter>Fill>Deterrace...). So far what you see is a whole lot weaker than what I built with photoshop and layer effects, but I think eventually, once I get the hf completed, you will see the true power of the DARKSIDE! Erm, I mean the heightfield. Ahem...

06-29-2009, 12:38 AM
I always love your stuff so I'm looking forward to what you come up with.

06-29-2009, 02:18 AM
Mound > Deterrace.

06-30-2009, 04:01 AM
I'm going for a somewhat different effect. Should have left the highlands flat, I think.

It's really experimental, but I'd like to see if I could get more of a series of steep cordilleras effect rather than one gradual rise. Global maps are deceptive. The scale gets you. If you do an 8,000m rise over 600km, you only get a 1.333% grade. Less than 1º of slope. Although, in all honesty, even at high resolutions, you really can't get individual ridges in any detail, so maybe my effort is wasted.

By the way, the new distance algorithm rocks. A new version of Wilbur is always a great source of rejoicing in the Paddock house. :)

06-30-2009, 06:28 AM
The deterrace code generates a number of peculiar artifacts and doesn't really work as well as I had hoped.
Using the mound tool should make it possible to get a better representation of your area, I think. The thing I did above was with a simple linear profile but the intent was to show that stacked mounds can get a good effect analogous to mountain chains. You can set the mound edge-to-center profile to be whatever you like, including a sigmoid-type curve to ease in and round over the tops.
The most direct way is to use the tesselation tool, but it gets a bit cumbersome after a thousand or so points.

07-02-2009, 03:54 PM
I might try that next. I'm not so concerned with the odd little artifacts, they mostly go away with a little blur-and-smear, and can add some character. The discontinuity between the flat areas and the linear slope areas, that is almost impossible to remove.

07-02-2009, 07:35 PM
Great to see you back in the saddle, Su Liam!

07-05-2009, 12:49 AM
Yet another version. I gave up on the deterrace, although I suspect with a lot of work and artistry(anyone know a good terrain artist?) it has some potential.

The rivers are a nice addition, although it would be cool if the flow map could be made sensitive to the global shape(if nothing else, maybe some sort of cylindrical scrolling effect). Not just for the rivers, but for the incise flow. On the north pole, of course, I could just bury the collection of radial rivers under an ice cap, but I would be surprised if the rivers lined up properly across the dateline.

I did this with the mound operator. It does look good, but my vertical scale is all shot.

As an experiment I tried applying a 1 unit mound on top of another 1 unit mound. That seemed to work fine. I'd get, depending on how the two mounds were aligned, about a 2 unit mound. That was what I had expected.

Now I do it again with a fractal variation scaled to 0-1. I'm expecting, say, a mound ranging from 0 to 4 in height. Instead it's bizarre, like 12,000. Any help in figuring out scaling would be much appreciated.

Oh, there's apparently a bug in the latitude shader. Really just an ugly, streaky mess. Shame, really as I'm having a heck of a time trying to get an analogous effect with photoshop gradients and whatnot. On to planetGenesis, I guess...