modeling for pioneer

[TonySpike]

s20dan wrote:
Well done, glad you traced that down thanks. So it was because the pilot was not called with LOD ? Ok good to know. @Spike/ /for the moment, Im pretty sure you cant have ship to ship docking of any kind. So the only way that you could dock to your big ship, is if the game thinks its a station. Which means you cant buy or fly it... Although I have never tried to give a model the flags of both ship and station, but I dont think that will work as its not designed to (yet).So you can set the ship up as a station with a docking port and it will just be a station that looks like a ship. Later we should be able to do more with it once there is ship to ship docking and the like.Or the other thing that you can do is just load it up with the station as an object next to the station, like the old LRC used to. It wont be dockable, it would just sit there.It could even be made to look like it was docked to the station, via an external docking port.
aah but it must be do able because someones done it as an oolite oxp i dont know if the scource code is any differant but the theory is there

[Marcel]

I've noticed that the posting has become a little sparse the last few days. I assume that everyone is busy redoing their models due to tomm's latest alpha. Anyway, I thought I'd post an update on my progress with the Ladybird textures. They're finally done, including the lod textures! I couldn't have done it without potsmoke (the guy, I mean). I still have to clean up the code, removing my personal notes, and I'm thinking of adding the registration numbers. I split out the model from the Taipan and changed the cockpit size to reflect the actual size of the ship. Potsmoke suggested that I scale the model to .5 which seems about right. I didn't consider that the ship weighs 60 tons. I was thinking a starfighter would be much smaller than that. It's actually more like a small freighter, a bit bigger than the Adder. I finally figured out how to check the scale by landing it next to one of the building1 chateaus (duh!). My previous windshield was about two meters tall! I also want to put the textures in their own folder in the ships folder to keep things tidy. I'll probably be ready to post it in a week or so, then I'll do the Taipan. It should go much quicker 'cause it's the same model and most of the hard work is done. I'll try to give it a different snake-like look so nobody will notice the similarities. When I think about how much time and effort has gone into this relatively simple project, It really makes me appreciate the efforts of tomm and potsmoke66 and s20dan and unclebob and all. Thanks! You guys are great!

[s2odan]

Quote:
I couldn't have done it without potsmoke (the guy, I mean).

Hehe that made me laugh. In my case it would be couldn't have done it without potsmoke and potsmoke66 ;)60 tons isn't an awful lot when you think about it. An empty F-14 tomcat weighs 20tons,33 fully loaded. And they are much smaller than the average ship in Pioneer. Probably much closer to the Talon fighter, maybe a little bigger.Yes it has been a little quiet here. It was my birthday so Ive been slacking somewhat but I have been working on another model.. hmm I suppose I should really finalise the other ones But the final details do feel like work sometimes, Ill get round to it.Heres the new one Ive been working on:Its a mix of styles so I'm not sure what to think about it, I suppose I like it but not sure how the front mixes with the back-end.When you look at the pics imagine it as a big ship, as that is how I envisioned it. The bridge/front retracts so that it can fit into the landing area on the ground.And well done Marcel on finally finishing your model. I would recommend adding some ship doodads using Tom's new Arguments as they seem to add that little bit of extra detail and perhaps a pilot too if you haven't already done that.You should have no problem with that as you already have lod and landing gear, and the process is quite similar, but just give me a shout if any problems crop up.

[Marcel]

Happy birthday! I'm liking what you've done thus far. It has a utilitarian heavy freighter look to it. If I may suggest a name for it, I'd call it a Hammerhead. I want to point out that the Ladybird is tomm's model, not mine. I don't have the skillset to make models, yet. I'm just texturing the default bare-naked models so that they look better in the shipyard because it's something that I can do, and I'm hoping that tomm will like them enough to put them into the game. Then I'll feel like I've really contributed something. I'm not up to working on it tonight.I have a confession to make. I have become unstuck in time. I keep falling through a weird space-time warp and landing back in the early 21st century where I have to perform strange tasks before I can return to the present. It's mostly rather boring stuff, but sometimes it can be rather exciting. Uh oh, here I go again! Be seeing you!

[s2odan]

Cheers. Hammerhead was actually one of the names I was considering, so now that you also came up with that name I'll go with that one. The other name was Albatross, but I think Hammerhead is more apt.Personally I find texturing to be the hardest part of modelling, so if you have that nailed your well on your way.Yes I know what you mean, those pesky time-warps, they seem to be everywhere these days.

[Pinback]

Excellent design there s20dan.

[s2odan]

** Pulls himself free of the timewarp **Thanks man 8-)Here's some more pics just for fun:The Bridge:Inside:Outside with dirty textures:

[HawkerT]

Hey s20dan!Really cool idea showing the full crew of three in the cockpit (cool little workstations in front of them also). This gives the ship perspective and binds what the game shows with what the player must do (hire two additional crew members (in addition to your self) to be able to fly that baby (right?)). I haven't taken the time yet to commend you on that big space-station-dome fly-into proof-of-concept you did in that other thread. Really gives basis for execution of a lot of the ideas put forth in that thread as well. Thank you for all the hard work s20dan. Much appreciated.

[s2odan]

HawkerT wrote:
Really cool idea showing the full crew of three in the cockpit (cool little workstations in front of them also). This gives the ship perspective and binds what the game shows with what the player must do (hire two additional crew members (in addition to your self) to be able to fly that baby (right?)).
Cheers, Yes your right you will eventually need multiple crew for the ship, but probably a lot more than 3, to simulate crew in the engine room ect. If I had it my way, you would still be able to fly a ship with limited crew, but it would not function as well, perhaps the easiest way to simulate that would be to lower thrust and shield regen values and hyperspace range (although that is rather arbitrary and there would be better ways of showing decreased efficency), I would also like to have it so that when you hire an extra crew member, they appear on the bridge, or somewhere on the ship visibly.EG, you buy the ship, in the bridge there is only one pilot (you). Hire more guys, and more guys pop up around the ship, maybe you see one in the hangar bay or some more in the bridge.Ideally I would have added more crew members to the bridge and some different workstations for them also but at the moment the model is right on the limit. But I should be able to fix that once I have added a full LOD setup for the ship.Glad you liked the stations idea. Tom has added some really neat functions that will allow this to happen Like internal lighting and inside out models .So at the moment we could have stations up to 200km radius (I think) with internal areas with their own objects and lighting.While thats plenty of space for most spacestations, it still wont let us have the mega-stations. Like ring worlds and sphere-worlds. I'm looking forward to when we can add these, as I have run some tests and the game can easily handle real moon sized objects like ships or stations. Its just the clipping issue, when you are more than 200km from the center and look to the side, the object disappears.(But the inside-out function might fix this )Also there are small problems like when you haev a moon sized station orbiting a moon the same size. If it were real, they would have similar mass, which means they would orbit each other(I forget the name for this). But its a small problem whcih could be ignored.

[Marcel]

I'm back from the 21st century, at least for the moment. It's been like a weird adventure game. Things are rather intense right now, but at least the graphics are good.
Quote:
Personally I find texturing to be the hardest part of modelling
Well you seem to have gotten that down. Those "dirty" textures really make it. It looks like a no-nonsense working machine. I'd almost expect an alien to be skulking around in there!

[Potsmoke66]

Quote:
Hehe that made me laugh. In my case it would be couldn't have done it without potsmoke and potsmoke66
gosh, and they asked me @ sci-fi meshes if i never get banned for my nickname even when i worked sometimes so hard on it because of lack of it.i had to help me out with coffee & tobacco then.thanks for your rewards guysmarcel, scripting models is much harder as build one with the help of cad, i respect that.maybe i fix up once a guide to thatsomething i have allready for you is a way to make a simple controlled quarter sphere.i did that in spring, when the sphere - sphere_slice primitive didn't exists.i guess that has made tomm laughing, but it works and you can fix up that way relatively quick a bezier quad shape which you can bend on all axes. it's to imagine to create shapes from ovals to something like a almond shape. i had some examples made then, but lost them when i lost the drive in may.the extensions can be controlled so it could be easely animated which quickly leads to the idea of a ship like MAX (flight of the navigator).i hope i still find the example script, i'm shure i have seen that not so long ago......i found one
Code:
function sophlos (pos, seg, rad, x1, y1, y2, z1, z2, float)cubic_bezier_quad(seg, rad, pos+v(0,0,z1*0.5)*float, pos+v(0,y1*-0.7,0.5)*float, pos+v(0,y1*-0.7,-0.5)*float, pos+v(0,0,z2*-0.5)*float,pos+v(0,0,z1*0.5)*float, pos+v(0.35,y1*-0.7,0.5)*float, pos+v(0.35,y1*-0.7,-0.5)*float, pos+v(0,0,z2*-0.5)*float, pos+v(0,0,z1*0.5)*float, pos+v(x1*0.7,-0.35,0.5)*float, pos+v(x1*0.7,-0.35,-0.5)*float, pos+v(0,0,z2*-0.5)*float,pos+v(0,0,z1*0.5)*float, pos+v(x1*0.7,0,0.5)*float, pos+v(x1*0.7,0,-0.5)*float, pos+v(0,0,z2*-0.5)*float)cubic_bezier_quad(seg, rad, pos+v(0,0,z1*0.5)*float, pos+v(x1*0.7,0,0.5)*float, pos+v(x1*0.7,0,-0.5)*float, pos+v(0,0,z2*-0.5)*float,pos+v(0,0,z1*0.5)*float, pos+v(x1*0.7,0.35,0.5)*float, pos+v(x1*0.7,0.35,-0.5)*float, pos+v(0,0,z2*-0.5)*float, pos+v(0,0,z1*0.5)*float, pos+v(0.35,y2*0.7,0.5)*float, pos+v(0.35,y2*0.7,-0.5)*float, pos+v(0,0,z2*-0.5)*float, pos+v(0,0,z1*0.5)*float, pos+v(0,y2*0.7,0.5)*float, pos+v(0,y2*0.7,-0.5)*float, pos+v(0,0,z2*-0.5)*float) end
it will have a bit a strange shading on the "tips" due to the fact that this bends the quad on the poles to tris, but if you alter that and space them 0.001 from each, it's enough to evade that problem i found out in my version for the Merlin.(i added a quadric bezier quad to the wings front to make it more round and you might remember merlins wings goes to the very tip of the model and all lines meet at point 0)oh what, the merlin? yes but iced up due to some other projects and some not satisfying problems with the flat for the wings, it's near to original but not quiet right. mine is basically a quadric bezier quad while frontiers is from two cubic bezier quads and the lines to connect them. i really get close to it but allways one part is turned around wrong or has wrong normals, how ever i tried. after a while i felt disapointed and stopped working on it.i would have liked to surprise you with the finished model, but ok, now it's to late...i added some wingtips to it to get a little closer to the original shape but, no i'm still not satisfied.guess i would need some of the "modelers helper" to get in the right mood to pick up the model and have the right patience to work on that again.(don't look at the wheel, i know completely wrong positioned, but i had change the setup for the wings once to make them animated so they are now completely wrong, but also ot important, first the wings have to be like i wanted them, then the uc can be fixed to it. it was allready once in a state the uc was working but like i said i couldn't get then the wings animated right and the "fin" had to be made from one of the wings to. so i reworked the wings and took the fin as base model, which is more regular as the wings itself as i found out).that's the script for the right wingi stripped everything from the script to show only the important lines
Code:
local v0 = v(0,-27,0) local v2 = v(0,27.35,0) local v4 = v(0,27.278,2.799)local v6 = v(27.829,30.405,1.479) local v8 = v(29.375,16.495,1.121) -- ex 10 local v10 = v(30.037,10.534,0.967) --ex 12local v12 = v(1.767,29.522,1.456) -- 1.503-- wing frontquadric_bezier_quad(8*lod,2*lod,v0+v(0,0,.001), v2, v8+v(0,0,.001),v0, v2+v(4,-4,1.4), v8,v0-v(0,0,.001), v4, v8-v(0,0,.001))-- winguse_material('top') flat(8*lod,v(.1,0,-1), {v2},{v6},{v8},{v2,v0}) --{v2},{v0},{v2,v8},{v6})use_material('bot')flat(8*lod,v(.1,0,1), {v4},{v0},{v4,v8},{v6}) --{v4},{v6},{v8},{v4,v0})
you can see the tip looks like all lines meet at vector 0, but in fact they are spaced 0.001, but you can never get that close in the game to see the little gap. if i wouldn't have spaced them 0.001 it would have a black shade on the wings front bezier quad (the strange divisions of the vectors are because i reworked the shape in blender and wrote back the resulting vectors from it, but it was important to have a exactly regular shaped wing, which border fits exactly to the angle of the body, this was the main problem on my first model. of course i could have used the bodys vectors for the wing like in the original, but texture would be streched then to if you texture the model and i guess that would look bad. so i build up a single wing, extracted it with GLXtractor and positioned it exactly to the body line, after i wrote down the resulting vectors. the gap that appears then i would have filled with a flexible tri, that would look somewhat more realistic rather a streched/bended texture).if you check the vectors you can see the wing is build with the front tip facing down, so that the rotational axis is y.the top right flat shape is build like this, divs, normals(face up), line starts at v2(back) goes to v6(wingtip), then v8(wingtip) and the bezier leads from last vector (v8) over v2(marks zenith) to v0(front tip) and the shape will be filled to v2(start) again.it took me a while allready to find that out and by the way i found the way how to turn the vectors around to create the same on the top left wing (vectors as comment behind the command), since you will have allways to respect that the poly's are build clockwise. when i converted the eagle i wasn't shure if that is possible at all and i chose xref_flat then for top right and bottom left (same rule as top right).one could guess now, hey why don't you just take a different vector for the zenith of the curve and pull the curve more to the back of the model. good thinking, but the bezier shape and the triangle will overlap then.it looks easy and logical now, but each time i start such a flat i get all kind of shapes but not the one i like to get.something i like much about the bezier stuff, it has allways absolute harmonic curves, where ever you lead to with the basic line, total harmony. it's that absolute that it allready looks artificial. if i draw a curve by hand i have to rework it often to find such a harmony.the original shape would look something like thatbtw, i checked the wheels anim in the modelviewer right now and i'm surprised how exactly the piston is animated, really surprised, usually it gives me headakes because rotation and transpose didn't work in the same speed, but here the piston stays on one point and transposing gets faster as more as the wheel fleds from position 0. i wonder how i did that then, endless experiments (even a look at my own script didn't help me to understand what i did)?

[s2odan]

I quite like that Merlin already, the curves are very tasty

[Potsmoke66]

god do i really have to read all i missed?the bridge for the hammerhead looks cool (who did that pilot's? ). no, in fact a cool idea to put some crew in it, even if we haven't actually in pioneer.i guess they are all females, hm? (in the modelviewer at least).further i guess i relly should work on some cad builded pilots, because poly count is still extemely high for a detail like a pilot.and you took 3 of them that will be about 20k only for the crew!droids and alien members would be a idea and possible to place them randomly or linked to a argument.Tomm,if you read this here i would suggest a argument for the autopilot to.i checked if it's enough just to enter a new one to the list, that will work in the modelviewer but the game has (of course) no idea about.i had that weird idea to make the droid for the R-41 from xwa depending on the autopilot, not so important but a nice idea (nice ideas? any idea about that...? might would need a timewarp then)i guess crew could be simply part of equipment to, you purchase a laser and get the weapons officer for free and like i said it could be selected randomly, so you won't know if you get a human, a alien (a protopode i.e.) or a droid as member.

[s2odan]

Quote:
no, in fact a cool idea to put some crew in it, even if we haven't actually in pioneer.i guess they are all females, hm? (in the modelviewer at least).further i guess i relly should work on some cad builded pilots, because poly count is still extemely high for a detail like a pilot.and you took 3 of them that will be about 20k only for the crew!
Every pilot I have ever added appears to be a female, Im not sure if the selector works as it should (50/50)I did a litle test on the number of tris. Max LOD with pilots and consoles is 9137 tris, and then without pilots and consoles its 7379 tris. You have to zoom right in to see the pilots, as they are only on the max lod which only appears about 20m away from the ship.It goes up even higher once you start adding weapons and scanners too.Thats basically why I havent released it yet, I still have yet to add a proper LOD to the main model.I actualy mentioned this autopilot request to Tomm before, as well as more changes to appearing passengers crew ect, but he never responded... Here look:
Quote:
Another random idea: This might be pointless but I'll put it out there anyway. Since potsmoker and myself have started adding cockpits inside the ships we could also add an argument for Autopilot, and any other internal components that might be visible in the cockpit.When they are added, we could add in a small object that you can see inside the cockpit.What do you think?Another one i was thinking of was for certain passenger ships. I was going to make a large passenger ship and thought it would be cool to have passengers at the windows, but there could be an argument for passenger cabins, or even better for if you have a passenger on board.So when you have passengers, the windows fill up one by one, depending on how many you have

[Marcel]

I just spotted a reference to this offer on Boingboing. It looks like some serious modeling software for real cheap!http://projectmessiah.com/x6/shop.html

[Potsmoke66]

a friendly (looking) marketing strategy anyway.but for what i need Blender's far enough.to get a model running in a game you won't need many of the features, in fact far less for nubes.a simple program like Z-Modeler can get you faster to your goal as even Blender or MAX.not to much of the rendering, physics for animations and all you won't need to just build a model.it will look anyway complete different in the game (whichever).sadly free Z-Modeler 1.7 don't offers .obj export, else i would use it for pioneers models and would strongly recommend it for any newcomer*.it's nice to have a lot of features like max, lightwave, maya or blender offer, but to learn the basic stuffit's far better to have a basic program.or to tell in other words we have nowadays modelers that have no idea about polygon winding, or what a crease is and what i have to fear from or if can gain some profit of them.when i started out several years ago i used a very basic modeling program, carCAD for NFS.but if you start with such polywinding and creases are no aliens for you, you will know where the limits are even when they are far expanded today.might some feel my uv mapping is very basic, truely it is. i never use the possibilities of a automated unfolding or projection and do it in a rather old fashioned style.but after all it was no big riddle to me how uv projection works in pioneer. it gave me some headakes in the beginning, but soon as i started to think different from what CAD programs teach you, i understood the system.for interested ones a short explanation.you just have to imagine that a texture projection in a script won't project the texture on the mesh (the word is allready misleading thoughts, because it's truely never). it's rather a projection of the selected mesh on the texture, so mesh projection would be the fitting term.if you look at it from this POV it makes clear why values get diveded if the mesh gets bigger then a relation of 1:1.in general any mesh will be projected on a texture sized allways 1:1, the rest is logical. bigger dimension in any of both projection vectors (U,V) will mean a dividing of the meshs size in the amount the mesh is bigger as 1:1.with this knowledge at hand you can calculate now the values needed for a proper mapping.since 1:1 means 1unit to 1unit or 1m x 1m you can simply divide 1 with the size of the mesh it has in the chosen vector of the projection plane. i.e. a model sized 3x5x8 will have at a x,y projection a uv of .333... to .2 and at a y,z projectionof .2 to .125.fortunately z hasn't to be changed (i don't know exactly why), but it's often enough to divide x or y vector to get a fitting mapping.with this at hand it's even possible to let a texture shrink, grow or move with the mesh (moving in the same amount as a mesh sizes up or down, can be very difficult i fund out)it's like any craftsmatter, learning from groundup helps better understanding and ends in better results.you can't go skiing just with a good equipment and to be a good skier it needs a lot of experience.and good skiers like i learned skiing on wooden skis with a steel core... had been walking themselfs up the hill, preparing the ski run by themself, all just for a 3 minute ride (but we never wore helmets, no). well if you ever had prepared the run yourself, you know all conditions snow can have, at least one when the snow starts to get sticky and at each step it builds up more and more under your skis.well to pick coolhands thoughts up and i was even thinking about such,a modeling software fitting to pioneers scripted models would be a nice thing to have, people could build and learn how to build scripted models in a easy as 1,2,3 way.i can see this allready as somekind of educational software, teaching the very basics of 3D modeling, like it was 20 years ago. could be used to teach geometry and mathematics in a joyful way, what scripting models does for shure to me."...all this i can do only with a expression?"really to myself this matter has become far less dry when you have some result you can see, instead of a bunch of expressions that won't tell nothing.---* try to build a model using carCAD, make some screen shots of your favorite spaceship and start to place the verticles on all 3 planes. that works really, my courier is the testamonial for that it will work.it's nice that we can take now a simple cube and form nearly any mesh of it, but by all the edge splitting gods, many basic stuff got forgotten then.a often seen mistake is the disregarding of creases or how the edge of the 2 tris that form a quad will break the plane.some example pics from my pilot, i guess you can't make a face of less then 49 trislet's say i made this one from a box, it comes out this wayeven when i set the mesh smooth, the creases are still to see and some are wrong because all edges leading from down left to up right. that will stay no matter how many divisions i use.in most cases, like for cars or spaceships it would be enough to simply cut the model at y axis and use the mirrored better looking half, but this low-poly face needs some tweaking of the edges.you can easely see all edges help now to form the shape, instead of some leading to a wrong side and form sunken in or stud out planes where i don't need them.a small difference with a big effect and like i said you can subdivide as much you like, the wrong look of the upper example will stay, and when smoothened the shading will show the wrong creases even more and they start to look like wrinkles.what can i say, the quads on the forehead are now the only ones that didn't fit and no matter how i turn the edge a small crease is allways shown.have phun...of course we have such nice things now like subsurf, but then you end up using 1000's of polys only for a face.apart from that, pressing buttons even apes can do...yes, subsurf is a cool tool, but i got now close to 700 faces (quads) and the basic creases are still to see!besides of that, i won't look much better now

[Marcel]

This is a wonderful example of what I meant when I said that you have tutorials all over this forum. With some editing, this and some of your other posts would make excellent chapters in your tutorial thread.

[Potsmoke66]

i was asked for how to place the decals in a comfortable way.good news, i added two functions to the squad_sign sub-model included in the last compilationthey allow you to call only the texture for the squad sign, it is possible now to copy and separate a quad or array from a object and to apply then the texture/material to it in the models script. the uv mapping will be given by the new objects uv.don't forget to set a higher zbias for the overlayed object as the rest of your model has.[attachment=533:2011-02-11_212546.jpg]it's supposed to be called in the dynamic part of your model, because it's simply like yow would call the selector function, this requieres the reg. no from get_arg_string(0) which can only be requested dynamic.to call the function is very easyfed_sign()zbias(1,v(0,0,0),v(0,0,0))load_obj('models/ships/adc/adc_fedsign.obj')zbias(0)that's for the roundlet with the triangledon't forget to add the material 'squad' to the materials used in the models info of the calling model[attachment=534:2011-02-11_213405.jpg]the imperial sign texture is even more flexible and you can call it withimp_sign()instead of a texture, similar to above.you won''t have to add the 'squad' material to the info, but you will have to specify a material for it, that could be a opaque one, then the imperial sign will come "clean" filled, but if you use a .99 set alpha of the material it will become washed out.a function for the decals i will write soon, i guess i simply forgot to, because it's no big thing.still interested in how to place a decal? ok,when you have build a model with a CAD program, usually you can read out somewhere the coordinates of a poly, which in case for the decals a regular quad of 1x1m. v0 is unfortunately in the lower left corner of the quad, but important to know if you like to place one and read out later the transpose or rotation values....to be continued, i need a break

[s2odan]

Quote:
it is possible now to copy and separate a quad or array from a object and to apply then the texture/material to it in the models script. the uv mapping will be given by the new objects uv.
Thats great, it will make it so much simpler to align the decals.
Quote:
...to be continued, i need a break
I eagerly await the rest :)Thanks.

[Taquito]

quick question: Is it possible to import models from Wings 3d?