Page 46 of 81

Posted: Sun Apr 19, 2009 7:15 am
by Cmdr James
Killer Wolf wrote:
as a "normal" Coriolis, i think those textures look awful....but as part of an OXP about a lost (Aztec?!) civilisation w/ stone-carved stations, they'd be awesome! That might be a canny idea, have some ancient-looking ships, and planets/moon w/ Nazca-esque patterns.
Have you been reading Chariots of the Gods?

Posted: Sun Apr 19, 2009 7:20 am
by Simon B
The trouble here is that we do not know what a normal coriolis looks like - we only have this approximation to go by.

It's not amenable to the usual default neolite treatment where part of the model gets cut away and smoothed. The closest we get is some variation on a sphere.

The next best thing is to inset some panels and put a texture on them... I'm looking for the sort of thing that makes spacecraft substructures. What I've done just happens to be quick.

How about:
Image

So it's time for something constructive ... what to do instead:
Remember it has to be simple, fit with the neolite look, and be fitting as a default which may get replaced by station OXPs. Don't just sit there and shake your head - come up with something!

BTW: I can see the aztec reference - in the back of my mind is to convert a particular govt type to something themed - just for a few worlds in each system. Under dictatorships there are all sorts of Empire themes. An Aztec Empire (along the lines of Space Slaan say) would be pretty cool.

No idea how to do this - retheming a system shouldn't be too hard - but to be really worth it we want the empire to be expansionist... but I digress.

Posted: Sun Apr 19, 2009 7:27 am
by Thargoid
If it could be a bit redesigned, it could go from Aztec to Feudal (think crenelations and castles) and maybe of interest to Ramirez for his WIP OXP?

Posted: Sun Apr 19, 2009 9:21 am
by Micha
Since you're not completely redesigning the stations like you did with the ships, Simon, what about considering past station work for inclusion here, such as the Griff Coriolis or the paagrove stations?

Posted: Sun Apr 19, 2009 9:22 am
by Griff
@Simon re: odd lighting problem on the coriolis a few posts back
I think the problem you're seeing with the light appearing to come from different directions on different surfaces of the model is due to some of the the texture UV's being mirrored & flipped and stacked on top of each other to save texture space in the map, flipping texture map UV's is also flipping the normal map (as the 2 maps use the same co-ordinates) and giving you these 'different light direction effects'.
It looks like flipped UV's can be tested for and fixed in the shader though, there's a very complex looking thread on it here
http://www.gamedev.net/community/forums ... _id=347799
all that goes right over my head though!
I did make a very simple stab at coming up with a solution in the shaders outpost thread, which was to do this in the vertex shader:

Code: Select all

   // Build tangent basis
   vec3 n = normalize(gl_NormalMatrix * gl_Normal);
   vec3 t = normalize(gl_NormalMatrix *tangent);
   vec3 b = cross(n, t);
   mat3 TBN = mat3(t, b, n);
//these next 2 lines added by me - they test if the vertex x position is greated than 0 and if so flip the tangent component   
     if (gl_Vertex.x > 0.0)
     TBN = mat3(-t, b, n);   
   
   vec3 eyeVector = -vec3(gl_ModelViewMatrix * gl_Vertex);
   vEyeVector = eyeVector * TBN;
This won't work though for polygons that make up the top and bottom of the model though as they often cross the X=0 axis and the shader only tests for the vertex position and makes its 'shall i flip or not?' decision based on that - it should really test the position of all tiny pieces that make up the polygon (the fragments), this sound like it might take up a lot of computer power though, so maybe it's best to use more texture space and not layer mirrrored UV's on to of each other?

Posted: Sun Apr 19, 2009 10:27 am
by ovvldc
Micha wrote:
Since you're not completely redesigning the stations like you did with the ships, Simon, what about considering past station work for inclusion here, such as the Griff Coriolis or the paagrove stations?
Or just point people to the relevant OXPs?

Mind you, a few station OXPs can do with proper Wiki pages and screenshots in addition to download links..

best wishes,
Oscar

Posted: Sun Apr 19, 2009 11:13 am
by Simon B
Griff wrote:
@Simon re: odd lighting problem on the coriolis a few posts back
I think the problem you're seeing with the light appearing to come from different directions on different surfaces of the model is due to some of the the texture UV's being mirrored & flipped and stacked on top of each other to save texture space in the map, flipping texture map UV's is also flipping the normal map (as the 2 maps use the same co-ordinates) and giving you these 'different light direction effects'.
That is definately the problem with the coriolis - each outside square uses the same texture, but rotated 90 degrees.

The main issue there is not the different light directions but the way the direction flips as it rotates. The effect is that the textured squares seem to pop in and out of the faces as the station rotates.

You didn't seem to get this for the griff-coriolis - did you redo the uvmap for that?

I am deliberately trying to do something similar here - only with light metal rather than black, and not so fiddly. Probably I'll end up just replacing the aztec lines with domes, pipes and boxes.


I do have some issues with light directions on some models - where, interestingly, there is no panel flipping (like with the Adder). I can work around this by changing the sence of the normalmap for those sections.

Mostly this is not a problem if I keep the differential small.

Posted: Sun Apr 19, 2009 11:36 am
by Griff
I did re-do they UV map for the griff_coriolis.
I can't think that i did anything clever to avoid lighting problems though, clever solutions to problems doesn't sound like me at all so it's probably just the laying out of the UV's that's the important step, checking that they are all the correct way around and not mirrored will avoid a lot of lighting problems i think.
this is kind of how i wished the stations looked, if i have another go at doing one, i'm going to rip this off totally!
Image
This the new tweakable ambient lighting code that's hopefully coming in the next test release we could really get these lovely colour contrasts

Posted: Sun Apr 19, 2009 11:45 am
by Simon B
Micha wrote:
Since you're not completely redesigning the stations like you did with the ships, Simon, what about considering past station work for inclusion here, such as the Griff Coriolis or the paagrove stations?
That occurred to me - the OXP stations are pretty special: I'm looking for something to use as a default, in-keeping with the neolite style (so it does not clash in the previews etc say) simple, but won't clash with the OXPs - gritty coriolis etc.

I figured a toned down griff coriolis as my model for style, but keep the silvery metal.

That said - I have tried dark and rusty colormaps for the dodo and isoc too.. Just for the change. Still, I think continuing the style across the range is better for the defaults.
Image
... dark colors can work - just add lights for windows - it makes the station hard to see from a distance: something that affects the early gameplay strongly. Hence, why I'd prefer to stay away from this sort of thing.

The fibery texture is another experiment.

I could just leave the inset part of the squares plain-flat? Use a more sparse set of lines?

Someone want to draw a square and fill it with greyscale shapes?

Here's a template:
Image

Hunting down alternate textures - discovered this site ... seems helpful. How about this for a station texture?
Image

Posted: Sun Apr 19, 2009 12:14 pm
by Simon B
Just to keep up thu dizzying pace - here's an escape pod proposal. The model is untextured at the mo. I also have a more pyramedal edition.
Image

Posted: Sun Apr 19, 2009 12:24 pm
by another_commander
Simon, I see in previous shots that you are using 1.72. I strongly recommend testing normal maps with a trunk build. Version 1.72 does not support tangent space normal mapping.
Edit: To demonstrate the point, see below two images of the same Griff Coriolis. To the left, we have the station as it appears on 1.72.2, and to the right as it appears (correctly) on 1.73, with tangent space normals. The light is coming from the same direction in both pictures. Note the issues with the relief of the first pic. It is very likely that the problem you had earlier is related to this.
Image Image

Posted: Sun Apr 19, 2009 7:21 pm
by DaddyHoggy
Cmdr James wrote:
Killer Wolf wrote:
as a "normal" Coriolis, i think those textures look awful....but as part of an OXP about a lost (Aztec?!) civilisation w/ stone-carved stations, they'd be awesome! That might be a canny idea, have some ancient-looking ships, and planets/moon w/ Nazca-esque patterns.
Have you been reading Chariots of the Gods?
Now there's a book I haven't read for a long time...

Cobra mkII X

Posted: Mon Apr 20, 2009 5:41 am
by Simon B
When I feel in a rut - I design a new ship.

Image
... from the x-ships oxp

The other craft are devients of existing models.

python X is the std python with an expanded engine section, the asp-high is the above treatment given the std asp. So I'll start from the existing models for those.

There are two barrels in that oxp - do you guys like cylindrical barrels?

Note - move the mussurana into x-ships and we can erase aegidean specials.
That would be a nice tidy-up of the historic oxp.

cmII is supposed to be a fighter/escort design IIRC: but the aegidean model was smaller than the gecko - I don't know how thought-out that was. As a player ship, it's witchspace capable - so I figured it should be bigger than that and scaled the model to be consistent with the aspII.

That should bring cmII into the same marked, and make the cheaper mamba (no witchdrive) believable as the winning entry.

... kept the classic profile and honoured Aegideans aft-mod. Lost is the classic cobra triangle aft-section, (big A left only the suggestion anyway) and thus some of the link with cmIII ... so I'll do a vanilla cmII which keeps it. The X is a mod after all.

I've exaggerated the cut in the nose, and tried to link the underside with the neo-cobra III version in it's styling.

If I match the vanilla tail section to the neo-cmIIs, the link should be preserved. Whatcha think?

Posted: Mon Apr 20, 2009 6:50 am
by ovvldc
Cobra MkIIx and the pod look nice (maybe a little big, almost like a small ship), and so does the final texture you found. But the other station attempts don't really work for me.

Best wishes,
Oscar

Posted: Mon Apr 20, 2009 10:24 am
by Disembodied
I think you're definitely getting somewhere with the station textures, Simon. The darker models work better, and I think the texture on the big sphere and cube could work OK. What I think has been the problem (for me) has been the lack on contrast, especially on the light-coloured stations. It makes the bumps and pimples seem small and low. With a sharper contrast between the base material and the detailing, though, the greebles get harder edges and start to look much more like very large details seen from a long way away.