Fractal Terrains Error when using Incise Flow

[Khaalis]

Hi all. I am using Fractal Terrains and just got a map ready to start doing the process in the Ishra tutorial. However, any time I try to use Incise Flow (any settings), I get the following objects in the map.


[EDIT] For reference these are right down the middle of the map. Also to note, in case it matters, this is right down the "break point" in the map.


Has anyone seen this? Any known solution or do I have to start over?

[waldronate]

What you describe is a known problem. FT has a number of issues with processing things across the -180/+180 longitude line, because that's where FT's editing data has a discontinuity. The only way that comes to mind to fix those holes is to paint over them with a small brush. Start with Tools>>Actions>>Normalize Data before painting to ensure that the holes aren't composed of something nasty like infinities or not-a-number values because those two classes of numbers can't be painted over. Then use the Tools>>Paint Values>>Land Offset tool to get a paintbrush to paint over the holes. Use the mouse wheel to quickly change the tool size to a small one. The hold down the Shift key and click on a nearby point to pick up a land value close to what you need to fix the current hole. Finally, click on the hole and it should be filled fairly closely to the desired value. Repeat until the holes are filled.

If you choose to start over (using Map>>Next World and Map>>Previous World to reset the map editing data without ), I recommend using the Longitude value on the Secondary page of the Map>>World Settings property sheet to rotate that line around until it mostly intersects ocean. The mapmakers here on Earth did that and it's worked out pretty well.

FT unfortunately doesn't provide facilities to directly import and export editing channels and it also doesn't provide facilities to directly rotate the world in the software. At the moment, You're pretty much limited to exporting the height field to a third party tool as an image, editing it there, and then importing it again into FT ( see http://www.fracterra.com/ThereandBackAgain/index.html for an example of how to do this with the previous version of FT, but it should all still apply ). This technique pretty much loses the dynamic fractal part of the operation, but it allows you to keep the general shape of a world that you've invested a lot of effort into.

[Khaalis]

Thanks for the reply. It looks like I am going to need to start over. Normalizing the data and painting does nothing. Can't change the artifact. Tried to import it to Wilbur and got even stranger behavior when attempting to do an Incise.

Does FT accept an other type of import? Or is the best bet to create something in Wilbur first and then import it into FT? If so, what is the best way to get a decent (accurate looking) world in Wilbur?

[waldronate]

If you're starting over, you should be able to move the longitude seam as I described above. If you are willing to freeze the height field, then you can follow the steps of "There And Back Again" to get a height field image back into FT, at which point it should be amenable to painting (or you can paint it in Wilbur).

You mentioned "Tried to import it to Wilbur and got even stranger behavior when attempting to do an Incise." What sort of behavior did you see? Wilbur and FT have pretty much the same code for incise flow, so I am surprised that they behave differently.

FT wants to import some variant on the theme of binary raster image. It's possible to convert any number of data formats to such a format using tools such as QGIS or GDAL.

For interchange between FT and Wilbur, I recommend using the MDR format because it is a small header followed by a large floating-point binary raster image. There is the bug about a vertical flip required in Wilbur, but that's fairly minor.

As far as getting an accurate-looking world in Wilbur, it's not a simple task. Continental shelves are one problem (there is a description on how to do it at the Wilbur web site). A good spherical basis is another (again, there are descriptions of this at the Wilbur web site). Erosion and other things also need to be done. Fractal Terrains is a bundle of Wilbur features wrapped up so that they are easier to use, plus some extra goodies like dynamic fractalization, dynamic map projections, better Campaign Cartographer export, and other things. FT should do most of the things that you want in a basic planet terrain tool, if you can avoid some of the issues (by far the most ill-behaved thing in FT is the seam at the back of the world: perhaps seams in the back are sexy when found in stockings on nicely-formed legs, but they look pretty silly on a planet).

One thing that you might try is to export a few masks of your FT world (set the ocean to solid black and the land to solid white with no lighting). They you can use one of the Wilbur techniques to regenerate some of the landscape and it might look better than the original world devised in FT. It also might look worse, though, because the mask-filling techniques commonly used in Wilbur ( see https://www.cartographersguild.com/s...ad.php?t=29412 ) really only work on a range of scales that's somewhat less than whole-world.

[Khaalis]

First, thanks for the responses.

If you're starting over, you should be able to move the longitude seam as I described above. If you are willing to freeze the height field, then you can follow the steps of "There And Back Again" to get a height field image back into FT, at which point it should be amenable to painting (or you can paint it in Wilbur).

I didn't really want to as I finally got landmasses in FT I was happy with (after a LOT of painting). I find that the random worlds tend to have a lot of ring formations or long strings. However, if I turn up land mass I get huge masses of land with massive mountains that are centered in the mass.

As to some of your specifics, please understand this is all new so I'm lagging a bit on the learning curve.
* How do you freeze the height field?

You mentioned "Tried to import it to Wilbur and got even stranger behavior when attempting to do an Incise." What sort of behavior did you see? Wilbur and FT have pretty much the same code for incise flow, so I am surprised that they behave differently.

Ok so this is what happens when I try to incise in Wilbur with the same file when following the "Fun with Wilbur" Tutorial, so thanks for pointing out that I can't use that. This is what happens when trying to change the height span.
Attachment 113236

FT wants to import some variant on the theme of binary raster image. It's possible to convert any number of data formats to such a format using tools such as QGIS or GDAL.

I'm not familiar with either. Are you saying I could save the map in an image format and then just convert it to a binary?

For interchange between FT and Wilbur, I recommend using the MDR format because it is a small header followed by a large floating-point binary raster image. There is the bug about a vertical flip required in Wilbur, but that's fairly minor.

I've been using MDR. What I've been trying to do is follow tutorials. The plan was:
1) Generate the world in FT using the CG and Ishra tutorials. However, to get what I'm looking for requires a lot of Prescale Land Offset painting.
2) Transfer to Wilbur (There and Back again)
3) Edit in Wilbur (Fun with Wilbur tutorials)
4) Back to FT to work on Climate and biomes.

As far as getting an accurate-looking world in Wilbur, it's not a simple task. Continental shelves are one problem (there is a description on how to do it at the Wilbur web site). A good spherical basis is another (again, there are descriptions of this at the Wilbur web site). Erosion and other things also need to be done. Fractal Terrains is a bundle of Wilbur features wrapped up so that they are easier to use, plus some extra goodies like dynamic fractalization, dynamic map projections, better Campaign Cartographer export, and other things. FT should do most of the things that you want in a basic planet terrain tool, if you can avoid some of the issues (by far the most ill-behaved thing in FT is the seam at the back of the world: perhaps seams in the back are sexy when found in stockings on nicely-formed legs, but they look pretty silly on a planet).

One thing that you might try is to export a few masks of your FT world (set the ocean to solid black and the land to solid white with no lighting). They you can use one of the Wilbur techniques to regenerate some of the landscape and it might look better than the original world devised in FT. It also might look worse, though, because the mask-filling techniques commonly used in Wilbur ( see https://www.cartographersguild.com/s...ad.php?t=29412 ) really only work on a range of scales that's somewhat less than whole-world.

I am unsure what you mean by exporting masks from FT. Do you mean just save them as an image and then use Gimp to paint the sea and land into black and white?

Would it be simpler to generate a basic png in Gimp, import it to Wilbur to create a binary file then import that into FT?

[Khaalis]

Ok, so I did the following:
1) Save an jpeg of the map from FT
2) Opened that in Gimp and traced the continent layout to create a black and white map
3) Used Photoshop to generate a basic hightmap and saved as a grayscale png
4) Opened the png in Wilbur and followed the steps here (http://forum.profantasy.com/comments...cussionID=7479) to generate an MDR
5) Imported the MDR into FT and got this:


This put the seam in the middle of the ocean - Yay!

Do you think this is good enough to handle starting to paint again?


[EDIT] - I no longer have prescale land offset options. I'm back to starting from scratch or trying to figure out what you meant with some of the other suggestions.

[waldronate]

I have a nasty habit of wandering down jargon-cluttered lanes of thought, so sorry about that.

To answer some of your questions, a little background is required. Altitude in FT are based on a fractal function of arbitrary complexity combined with some image maps (the editing data channels). The altitude at each point on the surface is calculated as

altitude=(fractalFunction+prescaleLandOffset)*land Roughness+landOffset

It's a tiny bit more complex than that because there is a calculated scaling factor to get your specified low and high values as well as an exponential operation before roughness to get continental shelves, but the above equation is the basis of altitude computations. You can simplify the altitude computation by setting roughness to zero, which removes everything from the computation except the landOffset channel. You can copy the altitude values into the landOffset channel and set roughness to 0 with one step: Tools>>Actions>>Burn In To Surface. I refer to this as "freezing the heightfield" because you're freezing the contribution of the fractal function at a single resolution, which is the resolution set for editing in Map>>World Settings:Editing.
You can get a similarly frozen height field by importing a binary file via File>>New as described in the There And Back Again tutorial and the using Tools>>Actions>>Burn In To Surface. Freezing the height field using Burn In To Surface has the nice property that it gives you smooth interpolation across altitude values. If you'd rather have a faceted appearance, use the Raw Height Field checkbox on Map>>World Settings:Secondary.

As far as the odd incise appearance, I would say that you're likely using too large of an Amount and possibly of Blur on the incise.

Your plan for interchange between FT and Wilbur seems like a sound one. I recommend taking the world that you have, save it as a special MDR, load it into Wilbur and then paint out your pits in Wilbur.


When I was referring to masks in Wilbur, I was refering to using the techniques described in the CSU Johnsondale tutorial ( https://www.cartographersguild.com/s...ad.php?t=18120 and also on the "Fun With Wilbur" thread ). While you could output an image and use a paint tool like GIMP to make masks, it's much simpler to make FT do the grunt work:

1) Use Map>>Lighting and Color to bring up the Lighting and Color property sheet.
2) Select the Altitude tab and set Colors under both Land and Sea to 1 (just one color for each).
3) Click on the Sea Level color patch in the Land section and select white (color 15; holding down Shift while doing this gets you the full Windows color picker instead of the CC3 one).
4) Click on the Lowest Depth color patch in the Sea portion and select black (color 0).
5) Uncheck any checkboxes that read "Blended" or "Shaded".
6) Click Apply. Your map should change so that any land in your map is white and any water is black. This image is your mask.

Save that mask as your coastline mask. If you want a mask for mountains, pick an altitude and lower and (or raise water) by that amount and save a new mask. Then proceed as normal in Wilbur.

[waldronate]

A frozen height field has no concept of fractal function or of roughness or of prescale offset editing.

You should be able to export a special MDR from FT to represent your world with the seam shifted, then use File>>New in FT to load that file right back in. You still won't get prescale editing or roughness editing because your height field is frozen, of course.

[Khaalis]

Ok, so another question. I did the following.

1) Made the black/white coastal mask.
2) Loaded it in Wilbur.
3) Added Fractal Noise to get this.
Attachment 113267

4) Flipped the vertical and saved as MDR
5) Open FT and load new world from MDR file
6) Instead of getting what I expects, I get this instead:
Attachment 113268

I feel like an idiot as I am assuming I am missing something I shouldn't be. Why is the fractal map in MDR not being imported into FT or is it just not possible to go back to FT after building in Wilbur from a Mask?

If not, what Tutorial do you most recommend for actual painting of terrain in Wilbur? Most of the ones I've gone through are all random creation rather than actually being able to say lower or raise land offset.

PS: If I haven't said it yet, THANK YOU for all the help you've given!

[Khaalis]

[EDIT] On a hunch I tried to reload the MDR file in Wilbur and get the error:

Error Reading File
{location}\Test MDR1.mdr
May not be of type
MDR Surface (*.mdr)

Click OK and get:

Error opening file '{location}\Test MDR1.mdr:
Unknown return code (-2)

So now I am totally clueless as to what I did wrong.