PDA

View Full Version : Geofiction Base Maps using Wilbur



a2area
01-23-2010, 02:03 AM
I played around with Wilbur (my first time) and our geofiction world's fractal last night and got some really nice results (I thought) for churning out uniform base maps for players. I also used Illustrator for creating nice paths and of course Photoshop for composing. Combining this with part of the key I made previously for Torentine's political map (if players so choose to use it) gave a nice formula for produce high quality maps in a minimum amount of time.

I'm pretty excited about the realism factor that Wilbur can give to fractal terrains. However, there was some question of the scale of the erosion/drainage being too small (more continental than regional and showing too little detail, river incises) Maybe some of you veterans can help me with this.

The "fun with wilbur" tutorials gave me the basics then it's pretty much play around til you get what you want without altering the original terrain too much.

Here are a couple examples with no cities added yet with a western usa map at the same scale for relief comparison.

RobA
01-23-2010, 05:56 PM
Hi a2area!

I would like to suggest posting maps as jpg or png files, since bmp files can not be viewed in the browser.

-Rob A>

waldronate
01-23-2010, 06:17 PM
Wilbur can only generate data at a detail level proportional to the image size. The incise flow and precipiton algorithms generate their most plausible results at between 1 and about 40 meters per pixel and the rendering algorithm gives nicely visible results in the same range. Unfortunately that means that a Wilbur map would need to be on the order of 25,000 pixels wide and high for a 1000km wide map to give nicely plausible results. You can do that with the 64-bit version, but you'll need lots of memory and patience. A 25k square map takes a bit more than 5GB to represent and about the same amount for each of the undo levels - some algorithms require several work surfaces to do their job so it's just not practical at this point in time. Give it a few years, though...

a2area
01-23-2010, 06:42 PM
ahh... i see that now.. i will change them.

a2area
01-23-2010, 06:48 PM
it doesn't have to be THAT precise. Just close enough to give believable results at this scale.. i don't know what "this" scale is in proportions as i haven't figured it out but you can see the 500 km scale on the map and get an idea. It just has to look plausible by whatever means.. perhaps it can be "faked" a bit with the incise flow .. i will just have to give it a go.

This actually may look all right as is though. Any thoughts on the detail to this scale?

waldronate
01-23-2010, 07:40 PM
Wilbur's flow algorithms are pixel-based and become visually implausible when pixels get too large. Most systems that incorporate a physical model have the same problem. What I'm (badly) pointing out is that continental shapes are an interplay between the strictly local processes of erosion and water flow working against the large-scale process of mountain building. Wilbur has useful tools for the erosion part (subject to the strict limits of the model), but it is up to you to use the tools in Wilbur for the mountain building part.

I can't say much about getting good results at a continental scale because I haven't yet come up with a good way to do it. Wilbur's erosion algorithms give good results where the ground slope is more than a few degrees. However, most continents have a slope much less than that. The Mississippi River, for example, drops 0.28 miles over a distance of 2300 miles. That's an average slope of less than 0.007 degrees (the last 750 miles drop only 0.01 miles, which is far, far less of a slope). This small slope is well outside what Wilbur can handle effectively.

One thing that you might try is to use multiple layouts instead of a single one. For example, place a platform at low altitude, then do mounds for mountainous areas followed by some noise and erosion. The mountain texture is hard to get plausible with small-sized maps. See tha attachment for an example of an implausible large-scale map done using this process.

su_liam
01-24-2010, 12:18 AM
I've had a modicum of success with continents using Wilbur. It takes a lot of work, and an artistic eye(my tripping point), but it can be worked.

Sigurd
01-24-2010, 02:29 AM
So a quad core processor with 12+gig of memory and a 64 bit version of windows 7 could take the software to its logical limits?

Ya know I need a justification for my next computer. :)


Sigurd

a2area
01-24-2010, 03:50 AM
So I think the plan is to render some bump maps of various real world regions at the same scale and then go to town with wilbur recording each step until i achieve success. I want to do it in a sensible # of steps if possible. I'll post my results.

waldronate
01-24-2010, 11:33 AM
So a quad core processor with 12+gig of memory and a 64 bit version of windows 7 could take the software to its logical limits?

I was thinking a dual quad with 48GB myself but only because I can't justify the $40K for the quad socket quad core with 144GB...

a2area
01-24-2010, 08:05 PM
Here is what a couple more incised flows did applied at different settings to 3 different elevation groups. I think it gives a better feel for this scale and I am happy with it for the purpose. Most of all what i'm looking for is relative consistency from map to map that will stand up at a wider scale (continental) so it wouldn't be in my best interests to start making the landscapes too unique. And I suppose that is where the "fiction" in geofiction comes in! Unique natural featurs that could plausibly fit into this overall landscape don't necessarily have to be featured on the map in perfect topographical scale as much as i would love that.

Maybe someday Wilbur and Fractal Terrains will have a baby (-:

.... one question i do have that i couldn't find a satisfactory answer to:
I have been outputting gray-scale image from FT to open in wilbur.. works fine... BUT, I would like to have wilbur know where sea level is automatically instead of having to mess with the scale or range which wilbur doesn't remember the next time you open the file. I would simply make my color scheme in FT conform to this... output... input to wilbur and boom.. perfect sea level.
How do i do this? Seems like it should be obvious like (ie) "set sea level : RGB value" or something.
It doesn't make sense to have black as the default sea level when you have negative elevations (depths).

su_liam
01-25-2010, 12:56 AM
First point. Images don't contain real elevation data, just numbers fro 0 to 255(0 to 65,536 for 16-bit), so you have to use the Mathematical>Span filter to tell Wilbur what span of heights the image represents. You can find a tutorial here (http://www.ridgenet.net/~jslayton/ThereandBackAgain/index.html) to tell you how to save the data in a full precision binary format. You really want to avoid greyscale image files for elevation as much as possible.

Second point. You are now required to put up a tutorial on exactly how you did this:

Here is what a couple more incised flows did applied at different settings to 3 different elevation groups. I think it gives a better feel for this scale and I am happy with it for the purpose..
I think you might be using something similar to what Mr. Slayton does here (http://www.ridgenet.net/~jslayton/FunWithWilburVol1/index.html), but if you've come up with something that works better for small scale(large area) terrain, then me wantee!

waldronate
01-25-2010, 02:32 AM
Maybe someday Wilbur and Fractal Terrains will have a baby (-:

I sure hope not because Wilbur gave birth to FT 0.0 12 years ago...


I would simply make my color scheme in FT conform to this... output... input to wilbur and boom.. perfect sea level.How do i do this?

In FT's File>>Save As menu there is a file type called "Wilbur MDR Files (*.mdr)" that outputs the current view as an MDR file. I recommend this file type for use with Wilbur. I also recommend http://www.ridgenet.net/~jslayton/ThereandBackAgain/index.html as a tutorial for using Wilbur together with FT.

a2area
01-25-2010, 03:20 AM
I sure hope not because Wilbur gave birth to FT 0.0 12 years ago..
UGH.. well that would be incest then, huh? hah.. then why can't you use wilbur like a sphere and erode the whole dang planet... or can you and i'm just ignorant... which i am since i'm a wilbur newbie.

Thanks for that .mdr advice.. i had no idea. AND that link to 'there and back again.' I have randomly found some of his pages via the wilbur page and google search. If you know of a better index of his tutorials other than what is at http://www.ridgenet.net/~jslayton/wilbur.html i'd love to find it.

Although i am bit confuse about this tutorial here http://www.ridgenet.net/~jslayton/FunWithWilburVol5/index.html I can see he's going from wilbur to photoshop and i think i get the concept but am losing him at the last 2 pics.. he says he's applying more erosion so it has to be in wilbur.. then is he again going back to photoshop and compositing it for us to view? To a newcomer it almost makes it look like it's all being done in wilbur (the 'refilling' of the lakes) but in reality i think most of it is just masking in photoshop using the outline of select basins. Am i right here?

Well, i've got a lot of experimenting to do. Thanks for the help, i'll try to be more resourceful digging up info before asking.


and yes su_liam, i did use that tutorial you linked to.. i did a different setting for each of the low, med and highest terrains.. then went back one more time overall to finish it off nicely. I have i all written down so i will most definitely do a tutorial on how i get my final satisfactory results when i get to that point.

Thanks again.

waldronate
01-25-2010, 10:42 AM
The erosion and flow algorithms in Wilbur are all planar algorithms (one of the reasons why they look reasonable only over a limited range of scales). They generate plausible results only for equal-area projections. The typical Equirectangular projection where latitude and longitude are equally spaced isn't this sort of projection and you get odd puckers near the poles when you try to apply such algorithms. In FT, for example, the river routing algorithm looks mighty odd when it's in the polar 10 degrees or so.

The solution for this problem is use an equal-area projection, but then you have the problem of how to wrap at the edges. The most general solution is to provide a connectivity map that specifies how every pixel is connected to every other one. There is no good way to generate this connectivity map except with knowledge of the map projections.

There is a projection that I learned from Dave Allen that is very roughly equal-area and has simple connectivity rules: start with a square map and declare the upper-right corner to be the north pole, the lower-left corner to be the south pole, and the diagonal from upper-left to lower-right to be the equator. Connectivity is then simple (when you go off the top you come back in on the right; going off the bottom you come back in on the left) and quick to compute with no connectivity map required. I need to implement this algorithm in the various flow computations in Wilbur one of these years so that folks can do whole-world erosion and river computation.

The last 3 pictures in the Fun With Wilbur Vol. 5 tutorial have all been composited in Photoshop. I probably should have made that clearer in the tutorial.

The index on the Wilbur page has all of the available tutorials Wilbur tutorials. Please always feel free to ask if something is unclear. It's how we all learn, especially when there are too many features and not enough documentation. Many times user questions have led to useful features and explanatory documentation. Wilbur is a manifestation of my unwillingness to finish anything and inability to focus on one thing for too long so it may be difficult to actually use.

a2area
01-25-2010, 01:00 PM
There is a projection that I learned from Dave Allen that is very roughly equal-area and has simple connectivity rules: start with a square map and declare the upper-right corner to be the north pole, the lower-left corner to be the south pole, and the diagonal from upper-left to lower-right to be the equator. Connectivity is then simple (when you go off the top you come back in on the right; going off the bottom you come back in on the left) and quick to compute with no connectivity map required. I need to implement this algorithm in the various flow computations in Wilbur one of these years so that folks can do whole-world erosion and river computation.
Sounds interesting.. would make a bizarre looking map in rectangular form but i see the logic in the approach for the purpose of avoiding excess data overlap. I saw a wrapping approach for Jhendor but i believe it was done from two equidistant polar projections. I don't think my computer could handle whole world erosion yet anyhow at a good resolutin (0:


The last 3 pictures in the Fun With Wilbur Vol. 5 tutorial have all been composited in Photoshop. I probably should have made that clearer in the tutorial.
Good, i thought that was the case but then i wasn't quite sure.

I'm loving Wilbur and am cool with the limitations and realize that, like fractal terrains, it's got a lot of hidden (and apparent) power. I have been using fractal terrains passively for a few years but just recently am starting to get to know it better. I can see that together they make sense once you get the feel for the processes. Thank you for the help and i'm sure there will be more questions along the way!

Experiment time...

su_liam
01-25-2010, 02:01 PM
Connectivity is fairly straight forward in cylindrical equal-area projections the Lambert cylindrical equal-area projection (http://en.wikipedia.org/wiki/Lambert_cylindrical_equal-area_projection) is the general case. Gall orthographic (http://en.wikipedia.org/wiki/Gall–Peters_projection), Hobo-Dyer (http://en.wikipedia.org/wiki/Hobo-Dyer_projection) and Behrmann (http://en.wikipedia.org/wiki/Behrmann_projection) are merely special cases.

A good place to look for technical information(formulae and the like) is here at Wolfram Mathworld (http://mathworld.wolfram.com/CylindricalEqual-AreaProjection.html).

waldronate
01-25-2010, 04:13 PM
The problem with all cylindrical projections from a raster processing perspective is that they make at least two points on the globe a line in the raster. That is, there is a non-unique mapping between points in the world and points in the raster. A projection like the sinusoidal projection eliminates the non-unique mapping problem, but the connectivity on the raster gets screwy.

Dave Allen's approximation isn't a conformal or equal-area projection, but it's not terrible (its trivial edge connectivity and pseudo equal-area pixels makes it excellent for raster computations). It also has the advantage that the cardinal direcitons on the work map don't correspond to cardinal directions of a typical projection.

There are many possible solutions for compromise between rigor and ease of use. Dave Allen's solution is simple in the sense that it's only linear math, is a unique mapping, and its connectivity can be implemented with a couple of if statements. I haven't seen a better one yet and I've been looking for one for years. Having said that, I've known about this for 10+ years but haven't gotten around to implementing it anywhere. Sheer laziness on my part.

su_liam
01-25-2010, 07:57 PM
The "pointy poles" argument? I gotta give you that one! Even if the projection isn't very attractive, it would probably be better as an underlying projection for the editing tools in FT than the current equirectangular one. Wow, that was a sentence...

I really need to look David Allen up.

a2area
01-26-2010, 05:14 AM
it's coming along (0:

su_liam
01-26-2010, 03:36 PM
Tres bien! I've been trying to do stuff equivalent to the isthmus of Panama for quite some time with little success. The erosion effects look a little bit masked at higher elevations. This is good; I didn't know Wilbur erosion could be masked. Could swear I've tried that before without success...

I am, like Palpatine, watching your progress with great interest. Great interest, indeed!

Thorf
08-16-2013, 10:55 AM
The discussion in this thread is of great interest to me, as I am now working through a similar process to develop a world using Wilbur teamed with Photoshop.

What I've found is that perhaps unsurprisingly the higher the resolution you work with, the better results you get. I started with a test model of 6000 x 3000 representing the entire world, and ran it through precipiton erosion and incise flow procedures until it started to look about right. From the start this was just a test, but it really brought home to me the "scale" of the erosion actions, and the fact that I would indeed need to go much higher than that resolution to get a world that actually looked like a world instead of just a region.

Next I moved up to a 21600 x 10800 model, working this one through to close to finished too. The increase in resolution posed a number of problems, not just because of the time it took to complete an action but also because of the sheer scale. I found that mountains and hills worked out well - far better than at the lower resolution, although still too big to make a realistic world. But the plains were featureless expanses of nothingness. Lessons learned, my next model starts out with much more texture in the plains for the erosion to wear down but still leave something interesting behind.

But before working through the next world model, still limited to 21600 x 10800 for practical reasons, I decided to zoom in on the most important region and develop it separately. Incidentally, I reprojected it into a stereographic projection so that the shapes would be true (or at least truer), because the latitudes involved are distorted in the usual equirectangular projection. For reasons I don't understand, I was unable to get the file working right at the 14404 x 11492 resolution I started out with - fill basins was acting very strangely and corrupting the map. Cutting out the surrounding sea until it got to 9998 x 9999 solved the problem. The area depicted is about 5000 km across, and my starting design had mountains, hills, and textured plains to work with. It went rather well.

All my tests have reinforced the idea that working on smaller regions at high resolution is the answer. This seems to agree with Waldronate's comments in this thread, which is very reassuring. For me there are two things to bring away from this: for my purposes, it may not be necessary to get down to real scale of 25000 pixels to 1000 km; and it is probably easier and more practical (or in fact necessary) to work on the world piece by piece as necessary rather than trying to do everything in one fell swoop.

On the issue of scale, it seems to me that with a fantasy world it is acceptable - indeed, it's standard practice - to start with a reasonably rough map and slowly add detail as things develop. It's nice to have as detailed a base to work from as possible, of course, but it's just not practical to do everything in high resolution from the start.

At the same time, working with Photoshop allows height maps to be worked on in smaller chunks and recomposited back into the master map. This is absolutely necessary in the case of the poles, which really need to be done in a separate projection from the rest of the map. I use Manifold to accomplish this; I started out with G.Projector, but it doesn't work with high resolution images and is less able to round trip an image (e.g. from equirectangular to stereographic then back to equirectangular). My current plan is to work on slightly overlapping areas and slowly build up higher resolution maps of the world. Eventually these can be composited back into the world map - or not. I haven't decided yet.

lostatsea
08-16-2013, 04:52 PM
WOW Give this guy an award for thread Resurrection.:) 3 years 7 months and 21 days !I picked up some new tips for my own technique evolution !!

waldronate
08-16-2013, 05:06 PM
I'm much more impressed with the patience needed to go for an image of that size than I am with the feat of thread necromancy. I don't have anywhere near that level of patience. Of course, if I fixed the Wilbur implementation of a few (dozen) things, things would go MUCH faster.

Thorf
08-17-2013, 09:48 AM
It's really not that bad, as long as you know what you want to do. You just set things up, start the erosion and then leave it for a while. Yes, it takes a few days (or more) to get to the end result, but that's not really a big problem.

Precipiton erosion is especially easy to just set and forget, because it takes a loooooong time. I was setting it for 10 passes and it would still be going 12 hours later... Incise flow is slightly more irritating because you have to wait for the computer to chug through it all before it even displays the dialog box. But the good thing is that once the dialog comes up, you can change the settings and hit preview and it updates within 30 seconds or so even on a huge map. This is really useful.

By the way, I took notes of all my settings and saved backups at various points throughout, so I may be able to write up a tutorial akin to the wonderful Genesis of Israh tutorial eventually.

Waldronate, any changes you want to make to Wilbur would be wonderful. But I'm very happy with it as it is right now. :D It's a real life saver.