Welcome, starbright!
The first feedback you'd need would be in regards to UTCS canon, and I'm not qualified to provide it.
Just wanted to give some other types of feedback as early as possible, to save you time.
I'm officially the Art Technical Director, or some such title, can't remember now.
In the past, this project has been very lenient about some things; but won't be in the future, primarily
with respect to thrusters. Ship models need to show thrusters for up/down, sideways, and rotational
maneuvering; NOT just thrusters pointing towards the back. At the very least, there should be visible
retro-thrusters, for braking. Here I added them to the current Llama:
http://wcjunction.com/temp_images/llama/shot22.jpg
It should be obvious from looking at the ship where fuel and cargo are contained. There should be
obvious hatches for cargo and for crew.
By "fuel" I mean not just Helium3 for fusion, but more importantly, for propellant: The stuff that is
thrown out "the back" of the ship to produce forward thrust. The mass ratio of fuel + propellant
to the rest of the mass of the ship should be at least 1:1, and up to 4:1.
There should also be large areas of external surface devoted to radiators to get rid of excess heat.
In other words, ships shouldn't just "look pretty", but make some sort of sense. This is NOT currently
the case, though, --most of our ships are fanciful crap; but going forward we want more sense.
Your ship looks aerodynamic. Does that mean that it can fly down an atmosphere and land? If so, it
would need to be MORE aerodynamic. Otherwise it might be best to make it LESS aerodynamic, so
as to not step on the doggy doo of cheap sci-fi; --i.e. space-ships that look aerodynamic for NO
reason.
Additionally, we want details that help people tell the scale (size) of a ship. You can't put trees,
cows, or houses on them, so a good way is to have windows, EVA hatches, hand-holds in strategic
locations, etceteras.
http://wcpedia.com/dw/lib/exe/fetch.php ... e%3Acutter
http://wcpedia.com/dw/lib/exe/fetch.php ... e%3Acutter
http://wcpedia.com/dw/lib/exe/fetch.php ... e%3Acutter
http://wcpedia.com/dw/lib/exe/fetch.php ... e%3Acutter
Number two: What tool are you using? I hope NOT Wings3D... If so, abandon it immediately. That's
a dead end. Using Wings damages the minds of modelers almost permanently; they rarely recover.
Getting good at 3D modeling takes years, and you sound like you're pretty new at it. I can tell from
the fact that your mesh is set to flat-shading. First thing you want to do is set to smooth shading,
and only mark sharp edges that are meant to look sharp. Takes a few minutes, so your saying it's
an early stage of design is no excuse. Secondly, it looks like the model is about 300 polygons or so.
This is TOO early in design to even appreciate; and one picture is about 4 too few to tell even the
general shape of the thing.
Your mesh seems to show triangles. If you were an experienced modeler you'd know NOT to show
triangles. Meshes should be made of quads; --exclusively, if at all possible. Triangles are problems.
It is practically impossible to work with a mesh made of triangles. they are also not friendly to
sub-surface, and tend to cause shading problems with the bakes. Avoid triangles like the plague.
This mesh of a work in progress for a carrier is almost all quads, for example:
http://wcjunction.com/temp_images/lexington/shot13.jpg
http://wcjunction.com/phpBB2/viewtopic. ... 0858#20858
Also, coming up with a ship model can take a day or two, but adding details (greebles) can take a
week, UV-unwrapping has to be done manually, and may take well over a week of work, and then
texturing may take as much time as all the previous stages, including all the bakes, such as normal-
-map, ambient occlusion, plus material definitions, etceteras. Yep; that's what it takes to make
ships look realistic, as opposed to having them wear colorful pajamas, like many of them do now.
And after all that there's the work of adding LOD's (progressively simplified meshes), converting to
game mesh format (bfxm), and setting up all kinds of data in units.csv, such as mass, moment of
inertia, pilot location, locations of thrusters and weapon mounts, etceteras. TONS of work.
EDIT: Thought I'd clarify: Most newbie modellers that come here; specially Wings users; who have
no idea about modeling for games, they want to model a ship in one day, and then have their
tools vomit some automatic UV unwrap (good for nothing), on top of it they have multiple little
meshes, each with its own texture; and after all that, they expect someone else to do the texture
work and integration... UV unwrap is tedious work, has to be done by hand, and there are many
rules to follow for it to work. The entire ship must use a single UV map; --single texture set. And
you are expected to do your own texturing and integration work. Contributing just a 3D model is
wasting your time;
nobody will finish the work for you. Heck, we have experienced modelers here
that don't even WANT to learn to model and unwrap ships properly; who insist on having multiple
texture sets, and using projection textures, like when modeling for gallery art. Game modeling
is very demanding, and trying to get artists to do things right is like trying to push cable through
a pipe... Dozens upon dozens of their models are simply sitting there, because they won't learn
the integration work, and when I try to integrate them, I find horrible meshes and horrible UV
unwraps, and when I try to fix their crap, I can't, because the meshes are triangulated, and the
unwraps are automatic, and lack proper seams, and lack proper marking of sharp edges...
Huge waste of time.
So...
Typically you're talking about a month of work for a small fighter; --two for a carrier or station.
Working fast, that is.
3D modeling is not a casual pass-time; it's a dedicated passion or obsession; and you either got it
our you don't; and most people don't, and they come here and post some pictures for a while, and
then they go away and never finish what they started. Save yourself the trouble, if it rings true
for you.
I'd be willing to tutor you at the beginning, if you're serious; but I use Blender3D, and it's the only
tool I know well enough to help you with. It's free, and it's the ONLY free tool for 3D work that
one can recommend. So, unless you have 3D Studio or Maya or some expensive, professional tool
you're attached to, I'd recommend you go get Blender right now and start going through the
tutorials.
And, needless to say, I'd hate to tutor you only to see you move on once you've learned enough.
Such thing has happened before, and it's not funny.
We'd typically work together in a model at the beginning, as sometimes it's easier to just fix a
problem with a mesh than to explain how to fix it.
Or, if you prefer you could start working with existing models, fixing them up, re-unwrapping
them and re-texturing, to become familiar with the process. (And we NEED fixing of existing
art MUCH MORE than we need new models; so this would be the best course for you AND the
best course for this project.)
As for your question on how many polygons:
For a small fighter, 5,000 quads would be really good.
For a corvette-size ship, 15,000 quads
For a carrier, say 50,000 quads
For a space station, up to 150,000 quads
(Note: A while ago, after having heard this, some newbie modeler with a big ego and Wings3D
simply applied Smooth to his ridiculously simple mesh, as to push up the number of polygons...
When I say 5000 quads I mean 5000 painstakingly modeled quads; --i.e.
Real detail.
In fact, we expect meshes to be well optimized: There should be a minimum of interpenetrating
geometry, NO hidden, invisible polygons AT ALL, and polygon density in curved surfaces should
only increase in the vecinity of curved folds; not tile flat areas for no reason. So, the figures
above are guidelines for poly count AFTER optimization of the mesh. You'd probably start with
about double the poly count, for a fully tesselated mesh, and eventually get rid of about half
the polies by manual simplification (merging vertexes).)
Tesselation optimized example (about 6000 quads):
http://wcjunction.com/phpBB2/viewtopic. ... 0319#20319
Unwrap work:
http://wcjunction.com/phpBB2/viewtopic. ... 0000#20000
Bumpmap work in progress:
http://wcjunction.com/phpBB2/viewtopic. ... 0557#20557
Deciding materials:
http://wcjunction.com/phpBB2/viewtopic. ... 0747#20747
(This ship is for a WC-based mod that uses the vegastrike engine; not for UTCS.)
If the numbers seem high to you, it won't surprise me: Most people cling on to obsolete notions
that geometry is expensive, but textures are cheap. The situation has changed completely over
the past 15 years. Hardware geometry pipes can pump polygons at blindingly fast rates, whereas
textures remain up against the wall of memory bandwidth.
For a small fighter, texture size would preferably be 512 x 512. For a carrier, 2048 x 2048. And
space stations may use several submeshes with separate texture sets.
Normalmaps can be twice the size of the other textures, though, which increases the perception
of detail for all the other textures at a low over-all cost.
Finally, our engine supports "detail textures", which are small tiling textures that add "noise" or
other patterns to your textures, at sub-texel resolution, which helps hide pixelation and also
increases the perception of detail, but are yet to be used by any models.
Your ship reminds me a bit of the Hammer, which is not in game yet, but will be, soon.
http://wcjunction.com/temp_images/toad/toad_08.jpg
Though, it's not a stellar example of a ship making a whole lot of sense; but at least it has retro
thrusters, and some scale reference details.