(Texture needed) Rlaan Antimatter Refinery

[Link]

I've created a model for the Rlaan antimatter refinery. I don't have a texture yet because I suck at texturing, and I don't have any subunits or docking ports yet because I thought other people would probably make better decisions on the placement of such. But I do have the model itself, and I thought I'd present it to you. It may need some cleaning up and/or fixing, but I've got the basic thing ready.

Without further ado, files can be found at http://users.penguindevelopment.org/Lin ... r-refinery
(Direct link to .wings file).

I included a lot of fins since I imagine a lot of heat et al needs to be radiated by an antimatter refinery, and such a structure could use a load of shield controllers as well.

I'm releasing it under the LGPL. I hope it's of use, and makes it into the game in some form. Enjoy!

[chuck_starchaser]

First I'll wait to hear whether there is interest in it. If so, I can produce a set of ambient occlusion and static lights bakings, for use in the texturing process; but you'll need to decide about the sub-units and their placements first, as the subunits should cast shadows; --they need to be part of the baking process.

[Turbo]

I like the model. Sadly, I can't help right now because
* I'm busy with a video project
* I make decent textures and can modify existing wrapped textures, but have yet to figure out model-wrapping from scratch with consistent results

Before a texture artist takes this on, I suggest you post a smaller image of it here in the forum (so we don't have to click 2 links in succession), and explain what we are looking at. That will help the texture artist's vision match yours. For example:
* What is the scale? That helps with choosing docking ports, scaling windows if any, placing subunits, etc.
* What are the pointy bits (antennae, heat sinks/radiators, docking towers?)

I also have some additional thoughts about the station's place in the game:

* Most pilots, upon approaching this structure, will fly through the hole in the middle just for fun. But, the toroidal shape suggests that the refinery is a huge magnetic bottle to contain the antimatter. In such a case, a ship flying through the middle might suffer some "nebula" effects on their sensors, maybe some shield damage, and perhaps anger the station owners, "Request for flyby is denied for safety reasons."
* If someone were to destroy the refinery, the explosion should be much larger than normal.
* Have you considered creating an interior for this station? The game currently has only a few planet and a few station interiors. I have a new planetary base with interior on my design board, if I ever I finish my video project.

Once the basic texture is done, do take advantage of Chuck's offer. "Baking" takes a texture to the next level and is a skill more rare than texture wrapping.

Turbo

[chuck_starchaser]

Indeed. I know nothing about the Rlaan, or how antimatter is refined, but my general feeling about this model is of incompleteness. Lack of much detail. Specially if it's meant to represent a large station. My personal principle is that no matter how big a ship or station is, there should be a whole range of detail sizes down to at least a few things being inch-sized, like a few hand grips or rails near hatches. The very tiniest things catch the eye and make everything else appear very large. Light fixtures with wires coming out. Antennas. Recessed hatches or windows. Whatever. I don't know what the Rlaan use, but there need to be at least some tiny things for psychological perception of magnitude. But for that you need to decide the size of the thing and scale it so that grid units represent real measures, like a meter per grid line, or ten meters per grid line; and to work with a solid idea of the size of the part of the thing you're modeling. The way it is, my perception is of it being about 100 meters in diameter; but with a few very tiny greebles I could be persuaded that it is much larger.

[jackS]

Ditto what chuck said - more detail would greatly enhance the model.

The basic gist of a rlaan station is there: the empty core, the fundamental radial symmetries of the primary station body, the fins, the spiky projections. So, on that level, it's a great improvement, since we don't have any (real) Rlaan station models, and this clearly is one.
However, without some additional complexities of geometric detail in the main body, the main body lacks character (if it's toroidal, it's going to look like a donut no matter what, but there are a lot of ways to vary a donut to change its character ( consider 1 vs 2 vs 3)), and, to again reinforce what chuck said, without some smaller components, the sense of scale will rest entirely on the texture, which may be asking a lot.

[Link]

I imagine this thing being huge - some 24km from one radiator tip to the opposite one. That's a kilometre per grid line. Since antimatter is so extremely volatile, I imagine huge vacuum cells are needed to store it safely. Anyway, I'm adding docking bays and greebles right now, though I'm not sure what kind of greebles would suit the Rlaan well.

As for what everything does - it's a Rlaan base; most Rlaan themselves probably haven't the faintest clue what everything does .

Also, erm, if I were to add a load of greebles, I'd of course add to the polycount an awful lot. Any limitations as to how far I can go? The initial model had some 3.5k polies, but even small greebles easily add 10 polies, and since the station is radially symmetrical, that's a LOT of polies added.

And re. the subunits - an antimatter refinery will of course need to be armed to the teeth (because if it goes boom, it goes boom real good), but how many turrets would be reasonable for a station of these proportions?

As for the actual process of refining antimatter, I would say inverse gravitics would be a start; use antigravity to push the antimatter into a ball with a heavy core and lighter shell, and then use some mechanism to isolate the compounds that are or aren't needed. Perhaps a pulsed graviton beam to deflect particles, or a proton beam to selectively annihilate the parts that aren't favourable.

[klauss]

If you merely add greebles, we can keep the current mesh as LOD.
You could also model greeble packs (groups) separately (as an addon to the base mesh), that would be attached as a subunit. That would help lodding and scalability a lot.

[chuck_starchaser]

Also, erm, if I were to add a load of greebles, I'd of course add to the polycount an awful lot. Any limitations as to how far I can go? The initial model had some 3.5k polies, but even small greebles easily add 10 polies, and since the station is radially symmetrical, that's a LOT of polies added.

ROFL. 3.5 K is not enough to make a fighter look like more than a toy. Most videocards nowadays can handle millions of polygons at full frame rate. You won't have enough time in your lifetime to make a model poly-heavy by adding greebles. Poly counts go up dangerously fast when people do things like draw an intricate profile, then spin it around full circle in 100 steps, then copy-paste that 100 times around. But hand-made greebles take too much work to do to be able to catch up with gpu specs in a lifetime

EDIT:
Best greebles are things that come out a bit, like antennas and poles and suspended pipes. Greebles that are too close to the surface can be done with the normalmap, for one thing; but besides that, you'll find z-buffer precision issues make them disappear when you zoom out. So, for very superficial greebles that I cannot do with the normalmap, what I do is actually cut and displace the big surface, rather than float them.
Best policy is to make stick-out greebles and then give them "bases" or connecting pipes or wires, rivets, whatever, using the normalmap. Then the geometry seems less "disconnected" from the main mesh, and the normalmap features look like they have more "purpose". Another advantage is that seeing actual geometry greebles associated with normalmap features makes the latter more believable.

[safemode]

Why would the RLAAN use antimatter anyway? their ships are living, the energy source they use would have to be some kind of derivative from an energy source available to living things.

Rather than this creating antimatter, perhaps it should create a food source that the RLAAN ships can digest directly into energy. The food source could be a highly stable form of dense matter, created on these stations. When combined with the digestive enzymes on the ships, the dense matter becomes unstable, and takes on mass. It does this until it reaches critical mass and collapses on itself, The resulting explosion is a miniature version of a pulsar. This powers the ships

It would be cool then to have a RLAAN base that basically attaches to special asteroids found to contain this raw material it uses to process into the super dense stable matter. Kind of like a spider that grasps the asteroid. The station would then devour the asteroid and move on.

[chuck_starchaser]

I know nothing about the Rlaan, like I said, but you're talking about chemical energy, and chemical energy is not enough to power starships. If the Rlaan ships are living, that may mean they use chemical "food" for their basic maintenance functions; but they must have discovered antimatter, or something of the sort, if they are a space-faring race that has reached beyond their birthplace system.

[safemode]

The thing with antimatter is that more energy is required to create it than you get from it. It might make more sense that the energy is derived from harnessing the effects of gravity. A dense enough piece of matter could implode under it's own gravity. Much like a star going neutron nova. The energy released is massive. We can pretend that the matter is mined as food, and that special enzymes cause the matter to absorb protons until it reaches critical mass.

It's an idea anyway. Antimatter just seems too expensive. Containing it, creating it. since it doesn't exist in that form already, you arn't going to get more energy annihilating it than you put into creating it. My idea utilizes the energy already in matter, you're not creating it. Then Special enzymes cause the matter to "cold fuse", implode, then explode as a mini-nova

[chuck_starchaser]

The thing with antimatter is that more energy is required to create it than you get from it.

That's true of any form of energy storage.

It might make more sense that the energy is derived from harnessing the effects of gravity. A dense enough piece of matter could implode under it's own gravity. Much like a star going neutron nova. The energy released is massive. We can pretend that the matter is mined as food, and that special enzymes cause the matter to absorb protons until it reaches critical mass.

Enzimes are too big to deal with subatomic particles. Anyways, you're talking about the source of energy, which is a separate issue from storage. Antimatter is a way to store energy with high density.

It's an idea anyway. Antimatter just seems too expensive. Containing it, creating it. since it doesn't exist in that form already, you arn't going to get more energy annihilating it than you put into creating it.

Again, efficiency is a separate issue. As far as containing antimatter, that's the million dollar question. I would envision that one day we may figure out a way to build chrystals of matter atoms with antimatter particles in stable suspension. Then antimatter suspension chrystals could look like table-salt, and only release their nuclear energy when hit by a laser or something. Pure speculation, but you get the idea.

My idea utilizes the energy already in matter, you're not creating it. Then Special enzymes cause the matter to "cold fuse", implode, then explode as a mini-nova

You lost me now

[safemode]

definitely not for this thread, but the dilithium crystal idea is no more unbelievable than enzymes that can fascilitate cold fusion. Much like how they cause some things to gain or lose electrons at low energy, these special enzymes would do the same for protons.

the idea is, the atoms would gain so many protons, that their mass would cause an implosion, this would result in a micro-nova. Thus the energy put in is minimal, due to the enzymes that allow cold fusion at low energy. The energy retrieved is great, due to the massive conversion of matter to energy as the atom collapses from it's own gravity.

I just think that antimatter and living ships dont mix all too well, but hey, it's all made up anyway. back to texture talk

[jackS]

I have said it, written it, posted it, etc. before, but apparently it needs repeating.

Rlaan ships are not alive.

Living organisms are:

• used by the Rlaan in the production of materials and components that will later become parts of ships

• used by the Rlaan as primary components in a select few subsystems, primarily waste disposal/environmental recycling and distributed repair stations

• used by the Rlaan as PAIs on drone craft and other autonomous entities

[Link]

The easiest form of antimatter storage, I would say, is a strong AG field combined with a magnetic field, a vacuum and cryogenic temperatures. I'm not sure how the repulsor beam works, but such technology fine-tuned and implemented into small-ish tank walls could keep antimatter stored at high density in a ball shape. A maser could then blast small amounts of matter off the shell of the ball, and a strong magnetic field would tunnel the antiions into a nearby controlled annihilation environment. Of course, antimatter refineries will need loads of empty space as a buffer for any unexpected event; a tiny droplet of antimatter straying from its position could annihilate any of the nearby matter. Needless to say, if that matter happens to be part of a containment field controller, chaos ensues.

On topic, though, I'm still wondering what sort of greebles would go will with a Rlaan ship; they don't strike me as being the type of having visible communications disks and the likes. Other than some pustules, cupolas, raised turret platforms, and other forms of nondescript arbitrary protrusions, I'm having trouble finding a type of greeble that stays in line with the Rlaan feel. A lot of greebles would be so incredibly tiny I'd have to blow up my model by factor 10 for me to be able to work on them, and they wouldn't be visible unless you were practically touching the hull. I'm especially having a lot of trouble with the blank polies on the top of the station; what could I put there that still kept in line with the Rlaan look and feel?

[chuck_starchaser]

I'm still wondering what sort of greebles would go will with a Rlaan ship; they don't strike me as being the type of having visible communications disks and the likes. Other than some pustules, cupolas, raised turret platforms, and other forms of nondescript arbitrary protrusions, I'm having trouble finding a type of greeble that stays in line with the Rlaan feel. A lot of greebles would be so incredibly tiny I'd have to blow up my model by factor 10 for me to be able to work on them, and they wouldn't be visible unless you were practically touching the hull. I'm especially having a lot of trouble with the blank polies on the top of the station; what could I put there that still kept in line with the Rlaan look and feel?

Well, if this thing is 24 km in size, even if the antimatter refining systems take up 90% of the space inside, there's still enough room to house a million people. Of course, these are Rlaan, not people; and a million of them may not be needed. But still, you're talking about a city's population. What does a city need? Transportation of goods. So you need docks for cargo ships. Transportation of people, er... Rlaan, so you need docks for Rlaan civilian transports (smaller but with less "cranes" and more "tubes", perhaps). You need dockings for small craft, visiting VIP's. You need docks for launching base defenses. There you've got 4 types of dockings to implement. With 4 of each, you got 16 greeble clusters. You need base defenses. Communications... Maybe Rlaan don't use radar dishes? I don't know. But they must use something. Anti-meteorite shielding over sensitive sections? More heat radiators? Shield emitters? Is the station molded or cast into a single piece, 24 km in size? Or is it made of smaller, replaceable pieces? Etceteras.
And if the recycling and whatever systems are biological, surely there must be a use for sunlight; so greenhouse-like things perhaps?

(I don't know why people want gigantic ships and stations all the time. What would be the problem with an antimatter refinery being a kilometer or two in diameter? Certainly there is NO way in the world to convey the size of a 24 kilometer thing... Hell, there's hardly a way of conveying even a few hundred meters worth of stuff via a mesh that has to load all in one piece.)

[jackS]

24km across is a bit on the big side (one may assume the main body in this case to be perhaps mainly empty, but, as chuck mentioned, there's still enough surface area for a few cities worth of inhabitants) -- currently, the largest model in the VS dataset is the human starfortress, at ~13Km in it's maximum dimension. I guess I wasn't paying close enough attention, because I should have brought this up already -- I'd suggest a shrink factor of 2 to 4 on your planned diameter to bring it more in line with the scale of the largest stations. Don't worry -- It'll still be a behemoth

[Link]

Hmm, goes to show how much I know about the scale of things in VS. Alright, I'll adjust some of the parts to reflect the smaller station. How does 8.2km from fin tip to fin tip seem? The actual station body is smaller than that; some 5.2km from fin base to fin base.

I've expanded on the model a lot; files can be found at http://users.penguindevelopment.org/Lin ... efinery/2/ (model: http://users.penguindevelopment.org/Lin ... y_v2.wings ).

Images:
Whole station from a distance

Whole station, lower part from a distance

Escape pod launch tubes with turret spire in the middle


Inside the regular docking bays, showing entry structures in the middle, cupola at the back and turret pedestal at the top (on the ceiling)

(Purpose undetermined) spike quad with turret pedestal in the middle

Unprotected docking port for large ships

Turret pedestal on fin joint

(Landing light?) protrusions above docking port


The model has about 12k polies now.

[chuck_starchaser]

Good work! You can keep going. 12k is a joke.
Let me give you an example: This PU ship, the Cutter...
http://wcpedia.com/dw/lib/exe/fetch.php ... ame_01.jpg
http://wcpedia.com/dw/lib/exe/fetch.php ... ame_02.jpg
http://wcpedia.com/dw/lib/exe/fetch.php ... ame_03.jpg
is about 70k and no problem at all in-game. (Its size is only 125 meters length, btw.)
In my book, 50k polies should be the bottom limit for something kilometers in size. No videocard will have trouble with 50k. I had a 150k poly model I was working on for another mod; and no problem at all with my OLD videocard. I'd guess 1/2 million triangless might start to get heavy on some videocards, with FSAA on.
You could put A LOT more greebles inside the docking bays, given that these are areas the players will fly into and look at closely. And if those cuppolas are going to be made of glass, you could model what's behind them. And don't forget lights. Light fixtures. When it's time to start texturing, you can give them an emissive material and then do a radiosity bake, so that interiors AND exteriors appear to be realistically illuminated by multiple lights. By the way, you could model ONE docking bay and make it a sub-unit, with its own LOD, and then call for that sub-unit 8 times in units.csv.

Starting to look professional.

[klauss]

Chuck, don't forget onboard GPUs do exist, and are generations behind discrete ones. Maybe you can have huge first lods, but you need a... say... ~3k LOD at the very least.

[chuck_starchaser]

Sure; I guess LOD's must be a must... I've yet to do my first one, tho...

[pyramid]

This model (get wings mesh here) exceeds by far Pyramids_Modeling_Capabilities (= None). After loading the mesh into Blender, I have realized that the mesh should be only partly smoothed, which is close to mastering Magic for me.
Therefore, it would be appreciated if a volunteer could pick on the model and bring the mesh into useable state maybe even combined with texturing.

Any volunteers, Fendorin?

[chuck_starchaser]

I loaded it into blender, put the edge-split attribute on, selected sharp edges (135 degree threshold) and marked sharp.
That did the trick.
However, this model is completely ruined. Someone applied a boolean union of the torus and the spikes, and that
created a mess of a gazillion triangles all over the place; plus, if you look closely, some of the smaller features that
were booleaned-in produced higlhly non-planar triangulation, which shows as artifacts in the shading...



There's even overlapping, coplanar polygons in many places, like around the entrance door.



http://wcjunction.com/temp_images/rlaan ... y_v2.blend

If you give me time, tonight I can replace the torus with a new one.

[pyramid]

Thanks for picking this up, chuck. Experienced modeler speaking (while to me it's still magic). Sure you've got all the time in the world to make this mesh somewhat acceptable.
What does the edge split attribute exactly do? And how does "edge splitting" help the smoothing? In other words, what is the relation between edge splitting and smoothing?

[chuck_starchaser]

Well, smooth-shading is the NATURAL shading mode, for the gpu. In order to show sharp edges or flat shading, the only way a gpu can do it is by having the vertices "split"; --that is, having two or more vertices in the same place, but with different normals.
In the old days, the only way to do this in Blender was to select all the faces belonging to a common "smooth group" and pressing Y to split them. They became disconnected from adjacent polygons, and so the boundary around the smooth group became sharp.
The problem with that was that the mesh became very hard to work with: If you moved some polygons in one smooth group, there'd appear a crack along the split; that is, the other polygons across the split wouldn't move with them.
So, the Blender devs decided to make it so that a mesh can be kept in one piece, yet allow you to have sharp edges by "marking them sharp". But they made this feature dependent on a new mesh attribute they called "edge-split". Probably the reason was for backwards file compatibility. What it does it it keeps the mesh internally in one piece, but splits edges marked sharp on export, and for display purposes.
When you add the edge-split attribute, it has choices of automatic splitting by angle and/or manual splitting of edges marked sharp.
You DON'T want automatic, by angle. Only from edges marked sharp.

So, the way you do smooth groups, nowadays is,

1) Select all in edit mode, W -> Set Smooth. (always)
2) Object mode, add attribute, edge-split, from edges marked sharp
3) edit mode: select edges you want to look sharp and Ctrl-E mark sharp.
4) While you're at it, you probably (in many cases) want to mark those same edges as seams (Ctrl-E mark seam), AND...
if you are going to produce a high poly, subdivided mesh with sub-surf for normalmap and ao bakings, which you will, you also want
those same edges "CREASED" (Shift-E -> [2] [Enter]). Otherwise sub-surf doesn't know not to try to smooth-subdivide across sharp
edges, --don't ask me why...

So, I usually do all three in one really fast succession of keystrokes