Some fractal terrain questions

[Naima]

could you also give me the description of the erosion precipitation window entries?
I can't find it in the manual .

[waldronate]

If I recall correctly, the Precipiton feature is newer than the manual. The general idea of the feature that that the system drops agents on the surface, which move downhill in the direction of steepest descent to a low point, moving some amount of elevation difference between cells as they do so. The total number of agents per pass is equal to the number of samples on the surface and each agent is dropped at a random location to start. Dropping them randomly prevents a certain class of artifacts that happened with an earlier implementation, but it doesn't guarantee that every point on the surface will be eroded equally.

At every step in an agent's life, it looks at the surrounding cells. How many cells it looks at is determined by connectivity: 4 (the four cardinal directions: up/down, left/right) or 8 (the four cardinal directions plus four diagonals). Of those surrounding cells, it finds the lowest one and adds Delta * (difference in height between current cell and lowest neighbor) to the lower cell. It then moves to the lower cell and repeats the process. This process repeats until the agent is in a pit surrounded by higher cells, until the agent runs off the map, or until it hits its maximum allowed number of steps. How the agent determines the edge of the map is controlled by the Wrap item: "None" means moving over any edge is off the map, "X Only" means that moving off the edge in the horizontal direction wraps around to the other side, "Y only" means that moving off the edge in the vertical direction wraps around to the other side, and finally "Both X and Y" wraps both horizontal and vertical directions. How many steps the agent can take is determined by the Max Length item: a value less than zero means that the system should pick a maximum number of steps, while values greater than zero will only let the agent move that number of steps.

That covers the Basics of the Erosion (Precipiton) Setup dialog. The Multi-Pass section is all about laziness. I became tired of repeatedly opening the dialog to get my desired final effect and I added the Passes feature, which repeats the above process the number of times specified. I'm not sure when the last time that I use the Blend or Noise features; they control the amount of blending between the uneroded and eroded surface for each step and the magnitude of noise to sprinkle on the surface at each step, respectively. If I ever get scripting working, this section will probably become a link to a parameterized script.

[Naima]

Thanks , here another small problem , I applied the heightfield computation noise in a selection with a feather of 2 ...
but it affecs also the outside ...

The changed values are :

size x y z 222 for all and no spherical evaluation , operation type set to multiply and size to 150

1.jpg

about the above description of the erosion , I am not sure I understood the meaning of :

how to make it more deep carving?

[waldronate]

I can't stop notice though that a lot of tools and erosion are in wilbur but not in fractal, why that ? Why not just bring all tools in FT too

There are a lot of features in FT that aren't in Wilbur, too (multiple image overlays, dynamic fractalization, and so on). It's partly product differentiation and partly that there are some things that work well enough for a static-resolution system, but translate very poorly into a dynamically-generated one. For example, the precipiton erosion feature: in Wilbur, the user can see exactly what they're going to get; in FT, the parts that can get eroded are only a small part of the final display. There's also a big fugly hiding in FT that most people don't notice right away, but that colors everything: editing data is defined in the Equirectangular projection, but most of the algorithms assume a flat equal-area map. Things in FT (rivers, erosion, editing, and so on) pucker strangely near the poles because much of the code came straight from Wilbur and Wilbur doesn't know or care about spheres. If I were starting out to write FT today I would implement things MUCH differently (not least because hardware is so much more powerful now thanit was back in 1997 when the FT ball started rolling). However, I'm a little stuck between adding new features and allowing people to keep their existing worlds. I try not to add features that highlight the ugliness of FT's spherical handling. If I ever get a few months where I can work on code (highly unlikely until the new baby is mostly grown), then I might work on fixing things. I'd probably sleep, though, so no promises.

and also why I can easily save maps in right size while in FT i have to cut and then enlarge the map couse it not saves the actual told dimensions but the window taking also the empty areas.

When you request that FT saves an image, it tries to save the viewed area of the world. What you see should be what you get unless you specify window (export) proportions different than the screen area. FT's Zoom Extents feature assumes that your main window is square and sizes things for that. I don't think that there is a "zoom to active pixels" thing in FT anywhere, but it would probably be a useful feature.

Also a very usefull thing woulkd be the possibility to pan around even when having open a window like for the incise flow function , so that while watching the preview , in other parts of the world , a lot of time I had to scrap what was done and go back because what looked good in a zone didn't in another, and I was in a zoomed in because from max zoom out is not possible to see details.

FT uses modal dialogs to do most of the user interaction. Changing how the UI works to allow pan and zoom tools to operate on the main display but no others would be a major architectural change. A smaller change of adding a preview window to the various features with an option to update the main view from the preview would be slightly less work, but only slightly. The problem is that FT would need to keep a local copy of the world state to work on and the renderer would need to... Anyway, it's really ugly to implement and would take a very large amount of effort to try to force the current framework to allow for that sort of thing (you'll notice that there's not a preview on the FT version of Incise Flow: it's for very good technical reasons; you may also notice that the incise flow dialog has an odd blank spot at the bottom where Wilbur's has some controls - very suspicious, if you ask me).

[waldronate]

Thanks , here another small problem , I applied the heightfield computation noise in a selection with a feather of 2 ...
but it affecs also the outside ...

The changed values are :

size x y z 222 for all and no spherical evaluation , operation type set to multiply and size to 150

1.jpg

The dashed line showing the edge of a selection is the mid-point of the selection values. If we assume that a selection's intensity ranges from 0 (0%=no effect) to 255 (100%=full effect), then the dashed line shows at the place where the selection is 127 (50%). That means that even though your selection may show just a tiny area selected, there may still be parts that can receive up to half of the effect. In Photoshop, there is a warning if there are no pixels in the selection that cross the edge threshold; Wilbur doesn't have the warning. Wilbur also doesn't have the feature that allows you to view the selection as a color on top of the main image (Quick Mask).

about the above description of the erosion , I am not sure I understood the meaning of :

how to make it more deep carving?

With the precipiton erosion, what you get in a pass is what you get. If you're doing multiple passes, setting the Blend amount to greater than 100% will give more weight to the newly-eroded terrain when blending, but strange artifacts appear. Using Edit>>Fade to Prior with a value less than -1 will also give interesting effect, but cause strange artifacting. The incise flow operation is controllable, but the precipton one is much less so.

[Naima]

I have done another test and I put size to 1 1 1 and scaling to 10 ... this is the result ... Not sure why I can't get just a slight fractal noise added normally on the whole surface I used multiply because I wanted that some areas got slight under and others slight upper , to create breaking patterns and have less straight fluvial erosion as the % noise is too uniform and also is too much dotted .

1.jpg

[waldronate]

Why use multiply if you want noise added? Why not just use add noise and then offset down to bring the surface back down? For example, using Filter>>Noise>>Fractal Noise with an Operation of Add and an Amplitude of 100 will add numbers from 0 to 100 to your surface followed by using Filter>>Fill>>Set Value with an Operation of Subtract and a value of 50. These two operation have a net effect of adding fractal noise ranging from -50 to +50 onto your existing surface.

[Naima]

I will try that ...
meanwhile other issue ...

Now I exported the map as mdr from wilbur and loaded into FT ...

The height range is pretty huge in FT , I mean I have peaks of 16000 m and -45000 in under sea ... If I want to bring all among the levels of -9000 and +9000 in FT how could I do ?I have tried to set that into the world parameters but that sort no effect couse I guess works only for fractal worlds and instead this is a loaded one .

Also I was playing with lightning and color , but when I go to select the color its pretty hard to select when the colors are close and also there is no way to save the selection after having decided what colors indicate what ... or am I wrong?
Also about saving a proper sized image so to have the map exactly the size I want and not the window size in FT how can I do ?

[waldronate]

Binary Worlds in FT can't be edited in FT.

I don't understand your question about selecting colors in FT. It can be hard to pick an individual color when you have defined many colors, true, but a simple technique is to avoid specifying many colors in the first place. FT interpolates colors, which gives many more effective colors than you selected.

The "Select Coloring Scheme" tab on the Lighting Color window provides a way to save and to load color schemes. You will need to have installed FT in something other than the Program Files directory for things to work correctly because I made a mistake many years ago of keeping user data in the install directory.

Use File>>Export>>Sphere Map Image to write an equirectangular-format image using the current color scheme. I don't think that it's possible to write an image of a precise world size (e.g. Mollweide at 7000 pixels wide) using FT.

[Naima]

ok and how can I do to rescale back the mountain ranges that come from Wilbur ?