CT-2000 - LIHW Wake Industries Corporate Shuttle

[rivalin]

Ok, few different things, I've decided to do a bit of contributoryness for VS so I've got some ideas/questions

1. I was thinking what might be cool would be to do a quick video for the game intro just showing the VS logo in a quick animation, as you tend to see in commercial games, to make it look a bit more pro and a little less 1997. I know theres video support in the game now but does the retro weirdness with the way the game opens from a launcher and stuff like that make it difficult to play a little video clip on game opening? Also, I've searched everywhere for the VS font and havn't found it, can someone direct me to it if it isn't too much trouble? (I'm awful at navigating SVN so a direct link to where to download it would be useful)

2. Please can someone update the Wiki , you're not going to attract many artists with the level of confusion there, I found myself a nice, open project, the Prytanis, and started work on it, and after several hours of prelim work I stumbled across a year old thread stating that it had been modelled not once, but twice, over a year ago. I don't mind, but it's just shooting yourselves in the foot, because others certainly would mind wasting their time because of a lack of project organisation.

3.I'm thinking of adapting the modelling I've done so far for another race, with a more conventional mechanical look. I'm thinking of it basically being a private jet type spaceship, chartered to transport a small number of passengers from one planetary starport to another (thus reentry capable) swiftly and in style. I'm wondering which of the races will have lots of private corporate type stuff, with a bit of flair in the construction. I don't know the races well enough to know the art styles but I'm thinking a cross between style and functionality.

Here's the (now useless) Aeran style concept art, and the basic model, I know there's not much, but I'm just looking for some basic ideas.






PS please can someone who knows the game really well fill in the artstyle guides, just a bit, I doubt anyone's impressed by the
--EMPTY 07/05-- etc tags on the articles.

[pyramid]

I did some editing around the 3d models page, and it should be much more clear now which ships are still open and what concept art already exists. You seem to have found a good compromise of improvement and the existing concept of the Prytanis and I like your model very much. At the end it's with jacks to approve it or assign it to another ship type.

It's true that the artstyle guides are pretty desolate at the moment. I just hope that when jacks is more libre from his assignments he'll be able to focus a bit more on the many open aspects of the universe.

EDIT: Btw, I just found one reference to the Prytanis. By the second one do you mean the post from Fendorin that eventually ends up being the Emu model?

[rivalin]

EDIT: Btw, I just found one reference to the Prytanis. By the second one do you mean the post from Fendorin that eventually ends up being the Emu model?

Thanks for the reply, and yep it was just the one post by Fendorin but someone chimed in about half way through that there was a model that had been done a year or so before that by someone else, presumably Fendorin's model was the one that became the Emu.

ps if anyone has that VS font to hand or knows where abouts it can be downloaded that would be helpful.

[pyramid]

For the fonts you'll ned to checkout vega-proj/webpage. The file contains the font used for the VS logo. There is another font (Andromeda) which is used for the "upon the coldest sea" text on top right, which I will commit later on.

[jackS]

I'll be puttering about in the artstyle guides a fair bit over the next couple of weeks, working off something akin to Deus Siddus's template of questions. Once I've got a set (plan is to start with the manufacturing groups, then the contracting groups, and then leave the end-user groups for last) of first drafts finished (and edited for grammatical clarity and removal of excessive parentheticals), I'll make an announcement, and hopefully I can get some feedback on how to improve the utility of said drafts and future additions.

[charlieg]

But, jackS, you forgot to give any direction or attention to the ship draft in this thread...

[jackS]

doh. My bad --

For higher-end personal craft there's an entity called the Luxury Travel Conglomerate that operates in Highborn space and does both its own civilian designs (cruise ships, yachts, etc.) and contract work for other groups (e.g. the Hunters contract with them for the production of the Vendetta and Duelist). They're the largest producers of such craft. However, there are plenty of smaller companies operating out of Purist space or one of the more prosperous LIHW systems, if you're thinking of rather smaller sized production runs.

One thing worth considering with such a ship is that the producing group won't be the only (or perhaps at all an) end user group. As an example, the LTC might be building these ships for sale to the Uln aristocracy, or CMT executives, or their own Highborn constituents, or all three. So, factional variants (in both "paint job" and accessories) may be desirable.

If you're thinking more Leer than Boeing, then you can be pretty free with the design -- small-run productions of civilian craft need not be particularly in line with the aesthetics of either the military fleets, nor the infrastructure workhorses. (i.e. unless you're hoping for given craft to be a mainstay of merchant or military fleets, it doesn't need to show up to work in a matching uniform, and the existing modeling direction should be fine). However, there are some things that you probably do want to take into account:
It's not a Rlaan ship, so it has clearly visible engines. The clearly visible engines currently appear to all be in the back -- you'll want some in the front/sides (textural, at the least, modeled is even nicer), or the explanation for how the ship comes to a stop becomes convoluted
If it's a civilian ship, keep it civil (we see a lot of proposed designs for "civilian" ships that end up with more guns than half the military fighters out there One imagines you're flying your Leer jet into a warzone, you're probably the weapons dealer, not a combatant. )
It'd be nice to see some more distinct radiator surfaces appearing more regularly on VS models, especially the civilian ones, which won't be as concerned about how radiator surfaces are very difficult to armor.
For a corporate-jet analog, don't worry too much about factional livery, but think about the corporate or personal logos/livery that the end user would have applied to their personal conveyance.
Even smaller spacecraft in VS are A) somewhat larger, compared to current in-atmosphere craft in analogous roles, and B) mostly filling that bulk with space devoted to things required for spaceflight (i.e. not walking-around-passenger space), and this should be taken into consideration both in terms of sizing the overall craft and the relative size of cockpits, windows, etc.

I don't know if that's what you were hoping for in terms of direction, but hopefully that will at least help you figure out which questions now need to be asked.

re: Intro video, we've actually got a script for one that some work has already been started on if you wanted to take a look in that direction.

Also, re: modeling and organization in general: right at this moment I think our priorities should be something along the lines of Concept art (2d and/or 3d) > Retexturing/remodeling in-game models > new full-quality models. However, until I get the Artstyle guides into at least a half-baked state, there's not going to be much to help highlight the concepts enough to have concept art, and attempts to rebuild the uglier ducklings in our existing model pool may not do much to move them closer to canon. Now, if there are artists who (like Nózmájner with the Unadorned) are interested in assisting me in putting together concept art sets for a given faction, such assistance would certainly be welcome (willingness to work over AIM/ICQ, a solid grasp of the English language (and a willingness to tolerate my perversions thereof), and a sleep schedule at least partially compatible with GMT-07 being among the things I'd be hoping for in such an artist. All except the first preferred but not required -- I've found that, even with a good working relationship and understanding of X being meant when Y is said, the lag time on e-mail based or forum based communications is a particular impediment to non-glacial forward progress :-/ ).

[rivalin]

I've got some basic ideas for the ship based on the above, they're just prelims but I'd like to get the ship finished quickly as it's turned into a bit of a botch job anyway so once I've got the go for these basic ideas I should have it finished (though untextured) sometime today.





1. Engines, should they be like in this picture with an actual exhaust type protrusion, or can they just be luminous Star Trek style things?

2.I'm thinking that taking into account the the distribution of mass towards the back and the engines, the rear wings can mount three thrusters each which should allow full flight control (eg newtonian flips to bring the ship to a halt) if there absolutely has to be retro thrusters I can add those to the larger side pods.

3. It was going to be an external fuel line, but I was thinking that the two external pipe esque things linking the middle and rear side pods could be the radiators? If hull based radiators are needed, would they look like the grill type things greebled onto most sci fi ships, or look more like simple flat surfaces slightly extuded from the hull (eg like those solar panels people have on their roofs?

[chuck_starchaser]

Nice to see a new modeler around who already understands smooth groups.

Radiators can be just almost perfectly black surfaces, like 050505 gray in diffuse and specular, except for a very tiny amount of red (almost black) in the glow texture. Something even as low as hex 080000 in the glow texture is enough to convey the idea that it's hot. But yeah, if your question was about the shape, fins don't do much for you in space; just a flat shape is fine; unless you feel artistically compelled to put more detail on them; in which case it can still be justified. If you look at the rads on the ISS, they do have structure, because they have tubes carrying the coolant. Then again, you could represent the tubing with the bump/normal map, and/or by varying heat distribution in the glow texture.

[jackS]

1. Engines, should they be like in this picture with an actual exhaust type protrusion, or can they just be luminous Star Trek style things?

2.I'm thinking that taking into account the the distribution of mass towards the back and the engines, the rear wings can mount three thrusters each which should allow full flight control (eg newtonian flips to bring the ship to a halt) if there absolutely has to be retro thrusters I can add those to the larger side pods.

As I see it, the important thing with engines and retro thrusters and such is that by suggesting their existence visually, we acknowledge their necessity and presence. The particular details are less important: if we have obvious multi-directional thrusters somewhere on the craft, I wouldn't worry too much about exactly where, because by that point we've already achieved a differentiation from "yet another airplane in space" and its not as if we're going to be able to accurately calculate the center of mass of the ship anyway

Similarly with radiators, one could make some assumptions as to how hot they could run, calculate the rate at which they could therefore radiate energy into vacuum, guess at the efficiencies of all of the internal ship components, choose a particular coolant (I've always been partial to gallium alloys) and attempt to calculate the percentage of the ships surface that would then be covered by radiators.... or one could just say "we're going to have to have discernible radiator surfaces" and by doing so already be paying more attention to what a spacecraft might really need than most models currently do

[chuck_starchaser]

Apparently there'd be no way to make radiators truly realistic. A few years ago we had two physicists in the forum; Zeog and someone else, I forgot the name, who tackled the issue of engine and weapons energy budgets and efficiency, and came up with a minimum figure of excess energy that was so high that if you made the entire surface of a ship do double duty as radiator, it would have to glow at like 1500 degrees. Failing on absolute realism, we agreed at that time, best second choice would be to try for "lip service realism", by giving all craft at least some half-decent radiators. They were still too shocked at their own figures and what lip service could serve. So then I suggested a new idea: An active material that could radiate heat at greater than the "black body radiation" natural limit. I called this "MetaBlack", and skipping technical details, it would boild down to its suppressing its own radiation at the longer waves and using non-linear optics handwaivium to enhance radiation at shorter wave-lengths. I figured a nice and intuitive way of representing its nature would be to make it reflect like a mirror in red and green (like gold), and be emissive in blue. Finally, I gave it a honeycomb tileable texture to make it look like it has some structure, striving for distribution efficiency.

Here's a simulation in Gimp (with sky reflections) of what Metablack should look like in space:



If you like it, here's the diffuse texture:



And the specular:



For glow, it could use just a flat blue color, brightness to taste.

The above textures are tileable.

EDIT:
To use the textures above in a ship texture, they need to be tiled and/or cut and pasted to fit the area of the rad in the UV map. Two important things to keep in mind, tho: (1) The diffuse and specular must register perfectly for those seams and dirt streaks to look right. (2) The direction of the dirt streaks should be aligned with the front-to-back direction of the ship.

[rivalin]

Thanks for the directions, they've helped me to get most of the needed details on to the ship.

I'm not used to modelling for games and my models don't normally have to be particularly low poly, but I've tried to be efficient. The edge loops aren't too bad along the longest axis, the smooth groups aren't irreversibly applied so someone should be able to adjust them in Blender if they aren't Vegastrike friendly (I normally model angular mechanical stuff so I have limited experience of bothering with them ) . To make UV'ing easier I've separated all the complex bits into sub objects in separate meshes (which should help with LODs too). However, I'm modelling in Modo, and it has a fairly new code base compared to most modelling programs, and I've only just transitioned to it, so it may create certain oddities when exporting to certain programs. I don't have any experience of Blender or the VS engine so I'd appreciate it if someone could look the mesh over when it's done and give me some feedback on how compatible it is (and which export format is preferable etc).

As for the design;

-radiators have been modelled as simple rectangles along the bottom of the hull.

-multi directional thrusters have been added to the rear pods

-for an idea of scale, the front protrusion is a two man cockpit, and the craft is made to carry around 10-20 people, so in comparison to a lear jet of similar capacity it's several times bigger.

-There are several flat areas around the surface of the hull where a different paint scheme could be applied and I might be able to make some type of "frame" esque area for displaying a logo or symbol of some kind.


Thanks for that radiator texture as well, that was almost exactly what I visualised the surface as looking like. I'm still hoping someone will be interested in texturing it as I'm fairly awful at UVing, and it should be a fairly simple job to do a basic texture + I could do an overpaint to provide some idea of what it should look like.

I would think the best place to put it in the canon would be that it's manufactured by a mid sized corporation operating out of the LIHW, designing and producing a number of small to medium sized high end transports. I'll come up with a small back story gloss and maybe if the Minister of Information likes it I can make a few other craft of a similar style, perhaps a 2 seater and a larger luxury corporate transport for a 100 or so passngers to round out the range. Anyway, just an idea



updated pictures:

[chuck_starchaser]

I'm not used to modelling for games and my models don't normally have to be particularly low poly, but I've tried to be efficient.

Please don't make assumptions about poly counts, and what is or isn't game-engine friendly. Poly count should be the last and the least of your worries. Modern gpu's can put millions of triangles on the screen at full frame-rate. The belief that you need to be sparing with poly counts has outlived its usefulness by more than a decade already. It is a popular misconception.
The biggest limiting factor in frame-rate is memory bandwidth and latency. This again divides into input and output bandwidth.
Input bandwidth is the ability to read data; and most of the data being read is texture data, rather than geometry. Textures 2k by 2k and larger are big input bandwidth hogs.
Output bandwidth is the the ability to write pixels to the screen, and the biggest problem in this area is overdraw. Overdraw occurs when an area of the screen has to be painted over multiple times because there are surfaces covering surfaces. This is the worst problem in current Vegastrike models. Many modelers are so misguided that, to make a raise square area on a surface, they float a rectangular box onto it, thinking that this would be better than making a few extra cuts to modify the surface itself, because "it takes less polys". It takes less polys at a terrible cost, because the original surface will be painted, pixel by pixel; and then the rectangular box that covers it will be painted over, pixel by pixel. Well, I'm simplifying the description because the box will probably be drawn first, and the surface underneath will not be drawn, as it fails alpha test; but in any case, the alpha test has to be done pixel by pixel, so it takes a lot of bandwidth, even if not as much as a painting over. Thus, one of the most critical considerations in modeling is that there be no --or as little as posible of-- hidden surfaces. Ideally, if you move the camera inside the ship you'd see a totally empty space, rather than all kinds of planes and geometry crisscrossing the inside space.

Check out this post where I put links to an excellent tutorial on mesh optimization.
http://wcjunction.com/phpBB2/viewtopic. ... 833f8ce5f4

[charlieg]

Looks like a cool ship, kinda Battlestar Gallactica meets Wing Commander.

(That could be a poor description, I'm not the most sci-fi literate amongst us.)

[Deus Siddis]

Output bandwidth is the the ability to write pixels to the screen, and the biggest problem in this area is overdraw. Overdraw occurs when an area of the screen has to be painted over multiple times because there are surfaces covering surfaces. This is the worst problem in current Vegastrike models. Many modelers are so misguided that, to make a raise square area on a surface, they float a rectangular box onto it, thinking that this would be better than making a few extra cuts to modify the surface itself, because "it takes less polys".

Well a box might be one thing, but when you have alot of places where two complicated surfaces are intersecting, making them all one mesh can become troublesome too, with a good deal more geometry being added, shading artifacts, complicated UV unwrapping, more development time/work overall, etc.

[chuck_starchaser]

Output bandwidth is the the ability to write pixels to the screen, and the biggest problem in this area is overdraw. Overdraw occurs when an area of the screen has to be painted over multiple times because there are surfaces covering surfaces. This is the worst problem in current Vegastrike models. Many modelers are so misguided that, to make a raise square area on a surface, they float a rectangular box onto it, thinking that this would be better than making a few extra cuts to modify the surface itself, because "it takes less polys".

Well a box might be one thing, but when you have alot of places where two complicated surfaces are intersecting, making them all one mesh can become troublesome too, with a good deal more geometry being added, shading artifacts, complicated UV unwrapping, more development time/work overall, etc.

Well, if it's small boxes or pieces of pipes you're putting onto a surface, you might as well float those; but if it's a big surface covering another, then you're much better off welding them together. You can use a boolean union function, to do it for you; but there's a lot of work involved cleaning up the huge mess of triangles boolean unions produce. The worst situation is if you have like a large, flat, 32-sided cylinder covering a surface, because when you try to weld that it produces gazillions of triangles on the surface it covered. When I'm faced with such a situation, I float the flat cylinder, rather than weld it, but then I cut the surface under it to make an octogonal hole under the cylinder, to minimize the amount of redraw.

But forget about making a ship out of big, intersecting meshes. That's when the game slows down to a crawl. You absolutely HAVE TO weld big meshes together. Interpenetrated meshes also look ridiculous.
http://vegastrike.sourceforge.net/wiki/HowTo:Weld

[klauss]

...but then I cut the surface under it to make an octogonal hole under the cylinder, to minimize the amount of redraw.

And z-fighting!
One of the biggest benefits of doing what chuck says is that enormously ameliorates z-fighting artifacts.

[chuck_starchaser]

I was being lazy to mention it because then I figured I'd have to come up with pictures to explain it, or prove it, but yes...

When you float things onto a surface, you'll find that they may look okay up close, but from further away they seem to disappear. That's due to precision problems with depth representation (z-buffer).

[rivalin]

Looks like a cool ship, kinda Battlestar Gallactica meets Wing Commander.

(That could be a poor description, I'm not the most sci-fi literate amongst us.)

thanks, actually I was going for a little bit of the Battlestar Galactica retro-futurism syle with the design.





Thanks for the article, interesting stuff. With regard to this mesh, the main body is completely hollow, with no surfaces covered by other ones (unless I've missed one or two faces somewhere when I wasn't paying attention. The small sub objects were separated out mainly to, as Deus Siddis said, make it easier for someone doing a UV, and for LODS. The sub objects all intersect with the outer surface of the main mesh, they could be reattached fairly easily, but that would make UVing considerably harder.

Anyway there are different workflows for different engines, UE3 apparently autowelds sub objects back into meshes, some engines prefer surfaces of different meshes interpenetrate at maximum angles rather than sharing coplanar surfaces etc etc, so I wan't really sure what to go with, and as I said, game modelling isn't my area of expertise, apologies if the model isn't entirely up to scratch.







These shots might give more of an idea of what I've done with the model, the second shot shows how the mesh exports as an obj to wings, the sub objects all have the hole material applied and remain separate meshes

[chuck_starchaser]

The types of mesh optimization considerations I was discussing are NOT engine-dependent. They are graphics hardware related.

I don't know what you mean by separate meshes; probably and hopefully just that they are not welded; but just in case you mean separate mesh objects, let me warn you that, as a general rule, for games you need each ship or station to be a single mesh object, with a single UV-layout, and a single set of textures. Well, large stations can use submeshes, long story; but for small ships, single mesh, single texture layout. The only exception allowed and mandated is windows (glass and transparent materias), which have to be a separate .obj export because (a) they needs to have a different blend mode specified (ONE INVSRCALPHA, instead of the ONE ZERO used for opaque materials); as well as a different shader technique specification (details pending, this is the new way, actually not yet committed to trunk). In any case, if you have any windows or glass objects, export them all together to a separate .obj; the rest should be a single mesh object with a common uv unwrap and textures.

Also, and this IS an engine-dependent ideosyncracy, when you do the UV unwrap, you want the front-to-back direction in the mesh to align vertically in the textures, with the forward direction of the ship pointing UP (North) in the texture. Also, you don't want to fold the layout to exploit symmetries, or overlap similar UV islands. The only islands you can overlap are ones for black materials, as shading for them is irrelevant. You can, in fact, overap all islands for black material in one neat pile, then scale them down to a few texels 'size, and put them in a corner of the texture. Anything else should be layed out without overlaps, and with a bit of space between islands.

[rivalin]

The types of mesh optimization considerations I was discussing are NOT engine-dependent. They are graphics hardware related.

Don't worry I got it, I meant those were my initial considerations/thoughts before reading the article/s

I don't know what you mean by separate meshes; probably and hopefully just that they are not welded; but just in case you mean separate mesh objects,

The standard Modo workflow is to have separate "meshes" inside "items", the files export as a single obj or 3DS though. I've taken a look at some models made for commercial games while checking out workflows and they're all in single files, but often parts of the mesh inside the single file are often not joined together. I don't really know any blender terminology, but yes as you said it, they're just not welded, and that can be fixed by a merge vertices command in Modo that will join the objects to the main part of the mesh (without increasing poly counts or anything else)

Also did some messing around with paint jobs.

[chuck_starchaser]

Good!

That's looking very nice and clean, by the way; nice work.

[rivalin]

Good!

That's looking very nice and clean, by the way; nice work.

Thanks, you might want to see the model up close before giving it any more complements though


I'm going to resist the urge to continue refining the model and adding detail, and basically count it as finished. I get the impression the untextured models simply disappear into the ether around here so I will make an effort to improve my non existent texture skills and do some basic temporary textures after I've submitted the model, but I may take a while to get round to it, so it's probably going to stay an open texture project for a while.

I'll upload two versions later today, one with the sub meshes joined into a single mesh, and one with the sub meshes not joined to the main mesh, which should make it very easy for someone to LOD it as desired and update the LODs when necessary. I've seen threads on here with people being told that "we don't all have 8800 GTXs" for suggesting that the VS poly counts are a bit conservative, and other people saying that poly counts are generally insignificant so I'll leave it to someone else to decide what sort of LOD's are appropriate, or if my modelling style is of acceptable quality or not.



shots of the completed model;

[chuck_starchaser]

Starting off with a set of temporary textures is a great idea. Often, texturing takes a lot more work than modeling, and such delay is not strictly necessary, as even the simplest temporary textures can make a ship usable in-game. Those renders look excellent, and there's no reason why you couldn't get a similar quality of visuals from simply blending a few bakes.

Here's an example of a ship we've got in-game, in PU, and it has no "textures" per se, just material bakes modulated by an ambient occlusion bake:
http://wcjunction.com/phpBB2/viewtopic. ... 1606#11606

As for LOD's, again, you could come up with a "temporary lod", a bare minimum LOD: A box. Far from perfect, but a whole lot better than nothing. The worst problem, when a ship has NO lod's is that even if the ship is very far away and you don't even see it, it adds its entire poly count to the scene. A simple shoebox solves that problem. Eventually you can add a couple of real lod's, like 1/4 and 1/16th of the poly count of the top lod.

I'd say the ships that really evaporate are those submitted as a mesh file, without LOD's, without a UV layout, and without textures. If you post a whole package --bfxm with at least one LOD and a UV layout, and very basic textures, plus a units.csv entry--, I think you'll see your ship in-game in like no time at all.

[Fendorin]

OH very nice ship very nice design too

and very good rendering picture (i like lighty thing)

it can be a aera boarding craft : the POLYDORUS no???

it look like renforced ship for find way under fire, delivery aera marine special assault warrior and return to big unit or stay near the front-line(if it s possible to have a front line in space)

i thougt corvette size no??
it don't look like a fregate or fighter size but more like an armed and agile corvette !

plus i thought Boarding unit could be a first target of new mission ;title :
defend ship/planet/station to the marine troop invasion!!!