FT3 weird terrain generation issue?
So I'm coming back to FT after 6 years, and I'm working on generating a science fiction world, and I like the general map I've made... but on zoom in it looks like the terrain got run over by a truck, with tire marks. The effect is too regular and patterned to look natural. Is there any way of preserving coastlines and general mountain range placement without having a stripy lace effect?
Comments
You might be able to save an altitude selection from your desired world and paint into that selection on a new world, but that's a lot of work to get right.
I really need to be able to zoom in on relevant sections of the map for illustration purposes.
It may be that my install is buggy, but for example shift does not bring up the pan tool (which the tutorial suggests should happen), which means I'm switching tools constantly in order to move around.
If I shift the map over to a better display of the continents (so that, for example, a continent is not broken at the edge of the map,) things get really, really weird when I incise flow. It would be a logical function, and perhaps it exists, to be able to tell the map, "shift this thing over and treat this as continuous." I just spent a couple hours fighting a section of map where raise and lower didn't work at all, but if I put on lower roughness, the terrain would go up (sometimes dramatically) and if I did raise roughness, I'd end up with a circular pit. This is... not good. Yes, I can shift to a different layout, but it wastes a tremendous amount of my time to find a planet I like, tweak it until I'm ready to work with rivers, and then discover that it's just flat out unworkable because it's not behaving properly around the edges.
I'm talking about maybe shifting to better palette layouts, and better defaults. Explaining things better at the point of the tool (on hover maybe?) rather than sending people out to the manual constantly. I've got 20 years of graphic design experience in multiple programs, have worked in 3d environments, have an outstanding spatial sense, and this is not behaving in intuitive and logical ways. I taught myself graphic design in Pagemaker and Photoshop, took a magazine from pasteup to completely digital at a time when even Adobe didn't know they had the capability of placing images within a PDF in a print-functional way, before InDesign allowed us to put PDFs within layouts, and I am comfortable learning new programs, but this one is really, really out of date. We know a lot more about functional UI now, and I'm no expert in coding, but this feels deliberately obtuse, and that makes it a much less practical tool.
And, you know, maybe pulling it from a native Windows XP program to Windows 10. But I'm biased. I'm popping back and forth between Google Earth and FT3 and everything about how Google Earth moves is really logical and this is... not. I mean, there's no reason to have to grab the orthographic to spin the globe, but when I do, if I pull right and left, it goes right and left, but if I pull up it goes down and if I pull down it goes up and that's just not good design.
Maybe only having it redraw the section you're working on rather than the whole thing? Or hey, we've got these amazing computers that are like 100 or 1000 times more capable than the machines it was designed for, maybe real-time updates? I've got 12 gigs of RAM. I did increase the threading and memory, so at least it's going faster but most modern programs will detect settings and give options for things like resolution and redraw rates.
I literally just paid for an upgrade from FTpro to FT3... and it doesn't feel like much of an upgrade and I'm pretty sure I was happier with it when I was messing around 7 years ago. So I'm willing to put my money where my mouth is, but right now I'm really, really frustrated, because I'm trying to design a map for a science fiction world I'm creating, and this SHOULD be able to do the job, but right now I'm thinking I'm going to let go of the map idea completely and move on to other parts of the project because it is not worth my time to keep fighting it.
I'm pretty sensitive to things that waste my time like this because I have rheumatoid arthritis which affects my hands, and wasted clicks and correcting mistakes that shouldn't have happened in the first place cause me physical pain. I don't mind drawing in a few mountain ranges, but I shouldn't have to spend a couple hours fighting a bad map seam by trying to manually fix what never should have been broken in the first place. The whole point of having procedural generation is to avoid that minutiae.
(And if we're really getting pie in the sky here on the modernization... plate tectonics? global warming? I can dream, right?)
I do my map works in CC3/CC3Plus. I make layer and sheet named bmp. Insert the png on that layer and sheet. Then use it as a drawing guide. You can hide the layer to check your work.
* It's a 20 year old code base with a UI framework that doesn't see many updates from the vendor. The code base was adapted from another project, which led to certain elements not working in an optimal way for what FT is trying to do. The seam at the back of the world is one example: it's an artifact of FT3's editing and computation system being based on what amounts to a paint program. It handles wrap in X, but not in a particularly elegant fashion. And updating the UI is a bit touchy. FT3 got a fairly significant internal upgrade from FT2 because I upgraded the UI framework by a major version. However, Microsoft isn't really updating the MFC framework and trying to switch over to another framework is pretty much a complete rewrite of the software.
* There are approximately 0.05 programmers working on it (on average). The means just one guy doing the code, and getting to work on it pretty sparsely at that. First priority is bug fixes and only then do enhancements get some time. I had hoped to get a new version out this year, but it's not looking like it's going to happen.
* (most importantly) FT has an installed base. That means that I can't redo things as I would like (especially in the procedural parts). I know a whole lot more about how to do things and there are many more libraries available now than there were with Windows 95. However, changing the underpinnings of the system to allow those new elements to take effect while still preserving legacy behavior is a, well, I'll call it "hard". A traditional software development joke says "God created the world in seven days because He didn't have an installed base".
A significant additional difficulty with people trying to use all of the editing features is that you need to understand that you're adjusting an underlying model rather than directly doing stuff. It's definitely a little weird and takes getting used to.
From your comments about the roughness behavior, I would guess (quite possibly wrongly) that you are painting a world from scratch, perhaps starting out by using the set altitude tool to get some constant altitude. The roughness tool is a scaling factor on the underlying fractal function. If you're pushing up roughness on an area where the underlying fractal function is negative, it will suck that part of the world down quickly. Similarly, if the
The pan tool should activate when the space bar is pressed, not the shift key. Using the Shift key should cause an eyedropper to appear with paint tools to allow you to pick up a nearby value.
The editing seam in FT3 is fixed at the -180/+180 degree longitude boundary, regardless of your center of projection. It's a bit of a silly thing, I agree, but avoiding painting in that area is really the best option if you can at all avoid it. We did that here on Earth for the most part.
I'm not sure what you mean by "pulling it from a native Windows XP program to Windows 10." Can you elaborate? I do agree that The navigation tools in Google Earth are quite nice and fairly discoverable; analogous tools have been on the wish list for a while. And the problem with the globe direction is an known bug that is in line behind many friends.
FT does try to redraw only the section that you're painting at the moment. When you release the mouse button, it triggers a redraw of the whole display because there may be little areas that it missed and I felt that it was better to have a correct and consistent display than a fast one.
FT3 is a 32-bit program that uses software rendering. As such, it can only access 4 GB of memory at a time and has little use for the enormous computing capacity of a modern graphics processing unit. Porting FT to use a 64-bit instruction set is possible, but not trivial. As I described above, procedural generation is very sensitive to effects like numeric precision, which means that any attempt to leverage all of the computational capacity of a modern system will require adding new subsystems while keeping the old ones exactly equivalent to what they are now. I am of the opinion that grafting that sort of thing onto the side of the existing system will take far more development resources than are available in any reasonable time frame.
I understand your frustration. You have a vision in your mind and the tools aren't getting you to where you want to be. That's never good. I apologize for the aggravation and for the pain that it causes you. FT does a lot of things quite well, but it also does a lot of things very badly. If I were setting out to build a new system like it today, I assure you that it would behave quite differently because the tools and techniques available today (also the ones that I'm aware of today) are very different than they were 20 years ago.
Plate tectonics is a long-time resident on the wish list (it's just about drinking age in the US, if I remember correctly), as is a better climate model and a better way of painting climates. I have a solid handle on how to do the climate things to a much better approximation than they are now I have at least a good understanding of how to do the tectonics stuff. Unfortunately, that 0.02 programmer-equivalents don't get a lot done in a given day and these sorts of things take a big chunk of focused time.
My current settings are:
Wilbur Ridged Multifractal
Highest peak 33000
Lowest depth 30000
Circumference 25377.2
IDK what the world seed is, roughness is a little lower than average, land size is whatever setting gives me two connected and at least one separated continents.
I'm not trying to paint the land from scratch, I'm generating continents and then fixing where the mountains are so that the planet looks like plate tectonics are happening.
The big error happens when I fill basins--there end up only being basins on the seam in ofset, and then there's this regular line of basins along the seam that makes any globe usage impossible. So I have a choice between tidy X wrapping or rivers, and that's frustrating. I'd like to be able to output a cylindrical projection and, for example, wrap it around a sphere in Second Life like I would any other planet map (I have a model in SL of the planets using NASA cylindrical projections and megaprims, it's pretty awesome as a teaching/perspective tool.
I don't have CC3 and am reluctant to get it because I don't need city planning capability right now, I need a globe.
Hitting shift does not pull up an eyedropper.
In the tutorial, it says "Painting Toys
Holding down the Shift key temporarily switches to the Pan tool to let you move around the world without having to switch tools"
So that maybe should be changed?
It seems like this program is suffering from a Catch 22... it's not getting much development because it's old, it's not getting much interest because it's out of date, it's not getting much interest so no one wants to develop it.
But I can't possibly be the only writer wanting to create globes. The price point would be good if it were more robust.
Google Earth has a developer engine... I'm just no programmer (I'm very good at designing structure, implementation is not my forte.)
I think FT2 Ridged Multifractal gives nice results, and i haven't seen any hiccups in the worlds I've created with that setting.
I haven't seen the fill basins problem across the world seam, but I do admit to avoiding that column of the worlds that I get in FT, even if I have to rotate the north pole longitude position to move the fractal parts away from it.
FT does support a sphere mapping mode for applying to balls in other software packages. File>>Export World>>Sphere Map Image will generate an image file with no trimming required.
I changed the problem description in the tutorial on my web site. I'm not sure that there's much I can do about copies of it on other sites.
FT suffers from neglect mostly due to lack of programmer resources. I'm sure that ProFantasy would be happy to sell more of them if they had a regular stream of new features to advertise and/or ways to regularly introduce new content for it in a manner analogous to their annuals for CC3.
The short answer then, is that yes, moving the North Pole will move location of the seam.