Random Stellar Maps

[kurtis]

In the further adventures of my effort to generate a custom alt-trav setting, I am writing some random map tools. Now that I'm happy with alternate map dimensions, I'm moving on to actually generating test maps.

I'm posting in the MTU forum, because I am exploring various heresies that I find interesting. Tonight's highlights include:

- Not every hex containing stellar objects has a populated world. There will be a number of wilderness stars which have no permanent population. (Keeps the scouts busy. )

- Some hexes may contain more than one star system. For the purposes of game movement, however, all systems in the same hex share the same possible jump destintations by drive rating.

- The stellar density in whole may be a bit greater than traditional, while the number of significant population worlds will be somewhat reduced.

- The maps should be both natural looking and fun. I'm shooting for maps with more variety and character that also reflect 'realistic' wisps, whorls, and clusters.

My first task in map creation is to determine which hex locations contain objects and which ones are empty. Rather than simply count down each hex rank with a virtual 2d6, I wrote a somewhat more convoluted algorithm. Its nothing fancy or elegant, its pretty much a brute hack I wrote last night.

The idea is to promote formation of both stringy mains and clumply clusters, discouraging obvious artifacts like long straight lines or sharp corners.

Here's what the output of my test app looks like:

Sample Dot Map

Not too bad. Has a few garish 'shapes', but also has some nice forms too. I was fairly happy with this, but I also wanted to enhance the clustering/rifting that would bring more contrast to the map and make the remaining stringy bits stand out more.

I'll post more radical experiments in further posts, after I make sure my images are linking correctly...

 

[kurtis]

Okay, that sample was a decent effort -- a little more dense than canon, but still within the box.

Note the density entry. This indicates the base percentage chance of any hex being occupied by some kind of stellar object. It does not necessarily indicate the actual outcome of the map as a whole. I use a couple of different biases when rolling for the results on each hex, and the total map density seems to tip one way or the other based on very small tweeks to the density input.

The other input I used is what I cooked up to crank the maps up another notch. I wanted to encourage 'clumping' in both thick clusters and rifts. The sticky input is a biasing factor that encourages some hexes to be like their adjacent neighbors. Cranking up this sticky factor really boosts the constrast and had some interesting, if extreme effects. It also has a tweek dynamic with the density setting, so getting interesting results requires some balancing.

Here are a couple of samples using my test sticky setting:

<a href="http://www.imagehosting.us/index.php?action=show&ident=555655" target="_blank">

Medium Density</a>

<a href="http://www.imagehosting.us/index.php?action=show&ident=555678" target="_blank">

A Bit Thicker Density</a>

It definitely had an effect (a nice one, too), but wasn't as strong as I expected. All I needed to do, though, was crank the input up a bit higher:

<a href="http://www.imagehosting.us/index.php?action=show&ident=555717" target="_blank">

Medium Sticky</a>

<a href="http://www.imagehosting.us/index.php?action=show&ident=555725" target="_blank">

More Medium Stick</a>

Very nice! Just the kind of effect I was looking for.

But why stop here? Keep cranking...

<a href="http://www.imagehosting.us/index.php?action=show&ident=555735" target="_blank">

Very Sticky</a>

<a href="http://www.imagehosting.us/index.php?action=show&ident=555741" target="_blank">

Very Sticky II</a>

Kewal. This level of clumping suggests topography and completely changes the character of the maps. I don't know how realistic it is, or if I'd use this much contrast, but I sure like the way it impacts at a glance and suggests natural and politcal boundries. Pocket empires, ahoy!

So what happens when I "turn it up to 11"?

 

[kurtis]

<a href="http://www.imagehosting.us/index.php?action=show&ident=555765" target="_blank">

UberSticky(tm)</a>

<a href="http://www.imagehosting.us/index.php?action=show&ident=555768" target="_blank">

Son of UberSticky</a>

Wow!

Holy cow, that would be a different style of TU. I fear it doesn't represent the local spiral arm very well, but I really like these maps. I am now tempted to try a second pass algorithm that will soften the extreme elements while maintaining the bold strokes of these high contrast maps.

It turns out "11" for this test app is around sticky factor 30. Beyond this, weird biasing artifacts take over and ruin the maps.

I still have more work to do on this (manual edit tools, database backend, actual system and world details, etc.). One of the next processes I want to write will analyze an existing dot map in order to locate likely clusters, in which the member stars will share some properties (age, at least).

In the meantime, I thought I'd share and see if anyone had any thoughts to share on scratch-made star maps.

 

[Malenfant]

Interesting. Somewhat high density of stars compared to Trav though, especially if you want 'fallow systems' where there are no habitations (or is this supposed to show all those too?).

Not sure about the Uberstickies though, a bit TOO clumpy I think.

 

[Spinward Scout]

I like A Bit Thicker Density, Very Sticky, and UberSticky. They look like viable star charts. Very Sticky II, and Son of UberSticky may have to many rift type areas that would make things hard to move around - unless that's your intention. Neat tool. How would you throw odd stellar objects onto that? Say, a four parsec long nebula (which I think is sorely missing in Traveller).

Dameon

 

[kurtis]

Originally posted by Malenfant:
Somewhat high density of stars compared to Trav though, especially if you want 'fallow systems' where there are no habitations (or is this supposed to show all those too?).

Yes. The idea here is to indicate that Something Is There. It could be nothing more than a common dwarf star or a cloud of gas or whatever. I may ask for some help when it comes to determining what the likely objects are. The question of habitation will be also interesting, but I'm not there yet.

Originally posted by Sir Dameon Toth:
Neat tool. How would you throw odd stellar objects onto that? Say, a four parsec long nebula (which I think is sorely missing in Traveller).

I agree! One of the very next things I want to do is add a process which analyzes the existing dot map looking for likely clusters, based on the nearby dot density. Going by GT:First In, you can then roll (or pick) an age category for any common group of stars. If the age is young enough, it will indicate a likely nebula, where the stars have thick pre-planetary dust and gas halos (I'm sure Mal could better tell us what to expect).

One of my goals is to make space itself more interesting in gameplay. Navigation hazards and points of interest, etc.

 

[Malenfant]

Well realistically, as I pointed out in my JTAS articles, it's likely that if we mapped out every star around Sol (including all the red and brown dwarfs that are out there, most of which we currently don't know about yet) then every hex would contain an object. I think it's quite likely that for a realistic universe (as realistic as a 2D universe is at least) every hex on a hex map would be occupied by something.

 

[Sifu Blackirish]

Man, the things you find when you look randomly...hope this isn't too long an interval before responding.

I like non-TU maps, but the dice rolling can get pretty tedious. so, I sort of turned things on their head. I take a map of a state, and superimpose a sector grid on it, making cities equal to stars. US roads and insterstates remain as trade routes.

See my signature for how one version of this works. Surprising what effects you can get from transposing RL locations into TU style maps...

 

[archhealer]

I love the idea of non-populated systems. I have a lot of them IMTU. Especially around the fringes of the cluster (It's the OTU, with an added-on cluster that's hard to reach- see the "Finding Solitude Cluster" thread).

I even have a (small) number of stars that are just out there, all alone, without so much as a planet. Useful, I suppose, only for navigational purposes or hiding!

 

[kurtis]

I've been meaning to get back to this thread at some point -- I'm still working on my mapping tool, but I switched the whole project over to the .NET 2005 beta platform, and I've been busy with coding geeky stuff like data adapters and such for the backend database.

Once I've got the database and dotmap bits redone, I want to come back to this topic and talk about distribution of clusters and specific stellar objects.

DLO, that's a pretty neat technique! Excellent outside-the-box thinking.

 

[Straybow]

(3D)

 

[Merxiless]

I like this project. It's interesting.

Long ago, I tried to determine the best way to squash a sphere into an LBB Subsector or Sector Map. Of course, it can't be done.

Scouts was a great book. Also GURPs Space was very good for ideas, along with Space Opera by FGU, Universe by SPI, and a few more.

There are a few guys that wrote a program a half dozen or more years ago called Cherry View, that map real world stars out to about 100 LY radius from Sol (after that, it loses definition, a bit, and only the brighter stars are mapped.)

I was turned on to another mapper called Celestia two days ago, from reading the threads here, where the one gentleman posted his pic of Regina. In a word, Stellar display.

In then end, it comes down to the (rough) fact that in any given 15-16 LY radius, there are about 50 stars in this area of the Orion Arm.

So, it ultimately comes down to not how many stars, stellar objects, etc. but how many significant ones.

Nebula, and all the rest of non-system hexes are great and add flavor. Due to the way jump space is set up, Eithher those things are ignored, or the places where jumps can't go are occupied by these objects (how it works IMTU).

As posters above said, pretty much every hex is occupied, (mostly by a dim Class M, or a pair of Ms, if you go with the flat 8 X 10 subsector. If you instead assume that each block is 3-D somehow (without the attendant problems of stacking them, and trying to map that), then the 50% value for standard mapping works out about right, giving each hex a value of +5 to -5 for Z-axis height, or 1-10, etc. for height, with each hex at 50%, looking down from the top.

I did the calculations some years ago, but the exactness of it didn't really matter to me, because really, it starts to turn into work, at that point, and the fun fades.

None of these are exact, mind you, more or less back of the napkin in the starport lounge type stuff.

Lately, I have taken to doing jpeg maps of a stellar area, using artist-type programs to draw spiral arms that look decent, with perhaps a few bright stars.

I then convert it to a dithered B/W color, and clean it up. It works just as well as anything else, except that it's then a problem of blocking it out into subsectors and mapping it by hand.

In the end, I think it's best to generate what your story needs, map those worlds, and then go from there.

But I agree, mapping and generating is great fun.

I haven't seen an approach of this clumping behavior used anywhere, so I am intrigued.

I look forward to reading more about this project.

 

[Merxiless]

Here are samples of what I have been doing, lately, to prep for my upcoming 2006 CT GRiP Campaign (Neither of these are finished, by any means):

Graphic:

Dithered B/W dot map, derived from Graphic:

The tiny, miniscule reddish area in the second one is the dot map for an 8 X 10 subsector (which I will use Galactic to map, then gal2cc to make into a traditional Subsector display with trade routes and such.

Of course, I will expand and label all of this later this week.

So, just another way of doing it.

 

[Ptah]

Merxiless, that graphic is amazing. Is that a Hubble telescope image or did you create it? Also, will that be your star map?

All for your ideas here. I did the calculations years ago as well, basically the average density of stellar objects was about 1 per cubic parsec IIRC.

It has been a few years (okay maybe 8 years) since I tried 3D star maps, stacking, dots conencted by bars off of a median plane. Worked well at small scale (20-30 systems) but rapidly became burdensome. Would love to try it again, I imagine there are some programs out there that could help make such a map for game purposes. Any suggestios along those lines would be great, and the program doesn't have to be free.

 

[Merxiless]

Well thank you. No it isn't from hubble, it's my original work.

The above is perhaps 1/20th of the entire galaxy I am using. I used it, as it was toned down a bit in color, the rest of it is well, I was going kind of crazy with the artwork there for a while.

Here is the whole galaxy (Very Large, and a bit gaudy, I admit, I did it some years ago. The colors are by no means accurate to any sort of realism, it was more or less an art work experiment):
http://www.landofetarnon.net/pardalis/GalaxyMap.jpg

The Known space map in this thread, and the galaxy graphic above was generated using the following:

Universe 1.62, from Diard Software (freeware, registered for a few bucks gets you the ability to do planets.

http://www.diardsoftware.com/

However, planets are much better done with Lunarcell, as a plugin for photoshop, or paintshop. Then cut and paste.

I'll put some of those pics up on my Planet Pardalis site later this week.

I have a pic at diardsoftware that they accepted as a demo of the program. (A map of an RPG campaign for another game system. They run art contests there, too, or at least, used to.

But Universe is great to do stars background, and lens flares, along with most of the stars, and some nebular gases / arms.

It's not very robust, all it does is do stars, planets nebulae, etc, so it is astronomy art, period, but, for what it does, it is Excellent, A+. It takes a few days to get the feel of it. Not because the learning curve is steep, the program itself is simple to use, you don't even need to read any kind of directions. It is just seeing how what colors overlayed with what tools to get what effects. And working with it a bit. For what it does, it's powerful.

After that, A few days of paintshop pro 5.0 to finish out and touch up the rest.

The Galaxy pic took a month, same deal.

As to the stars elevated above the plane, how big of an area are you looking for? Subsector? I mean, if you want, I can block out a grid in 3/4 view, and toss some stars in.

If you have a listing of coordinates, with soloes, binary pairs and stellar types, I could do a draft of something like that in a few days or less.

Otherwise, I'd recommend Cherry View.
I have to find a link for it.

I didn't write it. And it's old. But it has a grid function, sort of what you describe, with a 3-D spin. It has some bugs, and I haven't used it in some years, but eh, it's a tool for use.

If you want to hack off a slice of my galaxy map, feel free.

And yes, I will use the above Known Space map as my star map, and the dot map as my campaign systems locations.

Thanks for your nice comments.

If you're looking for some simple custom work, similar to any of the above, if you can get me a sketch, or link me to something similar, I can do it for free, just will take a few days.

 

[kurtis]

Bravo, Merxiless!

I particularly like the way the dot map derived from your color image. It creates an indelible impression of natural topography, which is exactly what I'll be shooting for as well. Tasty!

I bet sampling some choice photographic images could generate some potentially useful dot-scapes as well, using this technique.

Regarding 3D, I was hardcore about wanting to pursue Traveller in 3D several years ago, but I have become much more philosophical about it in recent years. As much as I like doodling for the fun of it, I have to make my time count -- these sort of projects have to be practically useful for near-term gaming. For now, I am content with pursuing a small degree of elegance in 2D.

If I knew my campaign space could be absolutely contained in an area of 8 or so cubic subsectors of space (something like 20x20x20 parsecs), I might consider taking up the 3D torch again.

 

[Merxiless]

If anyone wants to keep track of updates and such to the above, I'll save bandwidth here, and just leave the link to my new site. I'm going to try for launching the campaign sometime in January, after I get the mapping and such done. So there's still a lot of work to do. I'll be working on the Subsector tonight and perhaps Sector tonight, I think.

http://www.guildportal.com/Guild.aspx?GuildID=55289

 

[Merxiless]

Agreed, I really like this clumping idea, it very much reminded me of what I was doing, though my approach is do the pic (which is the randomness) then do the dots, and the clumping does the reverse, sort of.

In the end, both are great techniques, clumping would more readily lend itself to .sec file generation for Jim V's Galactic, I'm sure.

As to the 3-D, yes, back in the day, I was very much of the mind that I wanted it to be utterly realistic, that is, no sound in space, etc. All kinds of experimentation with various rules sets, and game systems, and real world research into stars, etc.

Then after 10 years of having players say, we want to go to X, and me having to fire up a 3-D spin program, and them usually getting lost, I relented and went with the standard, time tested (though innacurate, sure) 2-D map a la CT.

The players jumped for joy, at now knowing where they were, and what the borders were, and who was at war with who, not having to deal with Z axis any longer.

And I realized it is more about the story and the day to day scenartios each week, not 1000's of worlds that the PCs will never visit. It has to be used in game or have the potential to be used, or it tends to become a time-wasting frill.

So, agreed. 2-D is much easier, and in the end, it's all flavor anyway. A world is a world is a world. I figure it might be that the jump routes are topographical nodes in 3-D space, such that you cannot go from system A to B, unless you follow the route, etc.

Thus the 3-D pic whatever it is, might be wildly different from the 2-D routes map (which is what I consider the Spinward marches style maps to be.. travel route maps)

But as we see, it works.

And the offer of maps is up for anyone here. I like making them, when I am not building the campaign.

 

[Merxiless]

Here's a link to a Cherry View site that is still up:

http://members.nova.org/~sol/chview/

The CHView files described in the above site are located here:

ftp://chview.nova.org/users/sol/chview/

 

[Ptah]

Regarding 3D, I was hardcore about wanting to pursue Traveller in 3D several years ago, but I have become much more philosophical about it in recent years. As much as I like doodling for the fun of it, I have to make my time count -- these sort of projects have to be practically useful for near-term gaming. For now, I am content with pursuing a small degree of elegance in 2D.

If I knew my campaign space could be absolutely contained in an area of 8 or so cubic subsectors of space (something like 20x20x20 parsecs), I might consider taking up the 3D torch again.

Agreed and ...

As to the 3-D, yes, back in the day, I was very much of the mind that I wanted it to be utterly realistic, that is, no sound in space, etc. All kinds of experimentation with various rules sets, and game systems, and real world research into stars, etc.

Tried the same. And also agree that:

I figure it might be that the jump routes are topographical nodes in 3-D space, such that you cannot go from system A to B, unless you follow the route, etc.

Thus the 3-D pic whatever it is, might be wildly different from the 2-D routes map (which is what I consider the Spinward marches style maps to be.. travel route maps)

But as we see, it works.

I also tend to the view that not all stars off the galactic plane get mapped into the 2D representation.
It has to do with higher order non-covariant Tensor variations in the gravity waves emanating from the central rotating singularity at the core of the galaxy. Such variations create chaotic non-attractor states in space-time that render any potential jump points unstable in the Heisenberg limit. Yeah, that's it.