I'm back, so Bengal first, spirit?

Discuss the Wing Commander Series and find the latest information on the Wing Commander Universe privateer mod as well as the standalone mod Wasteland Incident project.
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

Tried that. Same problem.
Sorry about the outburst; I though you knew what the problem was, but were making me beg for the answer.
No answers at elysiun either.
My nerves are wrecked; too many hours spent on this crap. Can't take this anymore; I'm gonna take a couple of days off.
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

Got a reply at Elysiun:
EnV wrote:Yep, this happened to me too, the solution was that the script wasn't updated for the new cvs version of Blender I was using. Don't know if this is your case too. Somewhere, JMS posted an updated version of the script (on his site, probably) to work with last cvs builts.
Looks like a Halloween Build-specific problem. Must be a difference in file format, because I tried with version 2.40 Alpha 1 and it still gives me the same error. I even tried creating a new .blend in 2.40 and appending the file, and same problem.
What I haven't tried, come to think of it, is exporting the file to wavefront obj, and then importing it into 2.40...
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

Yep. The last idea worked: exported to an obj, reimported it into a new .blend file in 2.40a.

Image

And what that did also, for a refreshing change, was cleanup the project outline. It was an incomprehensible ratsnest before; now it's pretty clean.
Here's the file, for anyone interested...
http://deeplayer.com/dan_w/WCUweap/trip ... ooUV.blend
Half the size it was before, somehow...

EDIT:
Here's a new texture baker script, supposed to work with the later cvs releases:
http://www.blender.org/forum/viewtopic. ... b272#55942
It doesn't work for me, tho... (But then again, what does?)

EDIT2:
Speaking of which, as soon as I make any changes to the file, even in 2.40, I can't bake anymore and have to export to obj and re-import...
Managed to improve the UV layout, tho...

Image
DualJoe
ISO Party Member
ISO Party Member
Posts: 387
Joined: Tue Mar 21, 2006 2:37 pm
Location: Netherlands
Contact:

Post by DualJoe »

IThat's okay I'm stressed as well, with the problems I've got with my eye. Just got back from the hospital and they're pretty sure I won't go blind in that eye, but they can't make any promises on how well it will heal.
Dammit, I used to have more than 150% vision and now I've only got one good eye.

About your troubles with Blender, it could be a problem with the Halloween-cvs-release.
Because I have newer libraries, I have to compile the bf-blender-cvs by hand and I didn't encounter your problems with that version. 2.42a also works for me.
That said I could only get texturebaker to render the horizon in the first blend you posted. (Can't figure that one out either, but that could be because I don't know anything about the new features and options like nodes and renderpasses) Considering that an official release is about a month off and they are currently implementing a lot of new stuff some problems can be expected.

BTW, Klauss pointed out something pretty important about LOD's in my thread. Apparently the high-poly-version has to fit inside the LOD in order for parallax-mapping to work.
I now have to rework the Hornet, because I didn't do that and I'm getting some nasty problems with plain normalmapping as well.
klauss
Elite
Elite
Posts: 7243
Joined: Mon Apr 18, 2005 2:40 pm
Location: LS87, Buenos Aires, República Argentina

Post by klauss »

DualJoe wrote:BTW, Klauss pointed out something pretty important about LOD's in my thread. Apparently the high-poly-version has to fit inside the LOD in order for parallax-mapping to work.
I now have to rework the Hornet, because I didn't do that and I'm getting some nasty problems with plain normalmapping as well.
Sorry for that, it didn't occur to me that modellers would have to be careful with LODs in that respect. I mean, after you mentioned your problem, it was pretty clear to me, but not before.
I should have figured it out earlier (and mentioned it), sorry for making you retrace your steps, that sucks.

BTW: I won't be of much help with blender issues, and I'll try but I'm not sure I'll be able to bake GI on this thing, since I'm still using the older blender. I'll let you know how the GI thing looks when I do it (after getting the newer blender, the one with turbo and an apple slicer for free).
;)
Oíd mortales, el grito sagrado...
Call me "Menes, lord of Cats"
Wing Commander Universe
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

Thanks Dual, yeah, I always read your thread; I'd seen that. Sorry again; I have an extreme tendency to take things personally and perceive abuse where there's none. And I know it, and when it's happening I try to think about it and give at least the benefit of the doubt; but nothing works; it's like a flood of rage that almost invaribably proves to be unwarranted but when it happens I can't seem to stop it.

Klauss, don't worry; I'm working on a PRT setup that's going to make it all easier.

Image

I made a sphere and baked it with three radiant circles; then made the sphere vertex colors vcol-light, and pointed the vectors inwards. We'll just have to append it to a blend file and bake. I'll test it after work.
DualJoe
ISO Party Member
ISO Party Member
Posts: 387
Joined: Tue Mar 21, 2006 2:37 pm
Location: Netherlands
Contact:

Post by DualJoe »

Thanks for the link chuck it solved my problem with the see through textures.
Blender-cvs comes with a number of new bake scripts now, they are under the image button in the uv-editor. Including one for baking vertex-colors (for the radiosity-GI-baking).

Another tip is to encase the stuff you are going to put in a normalmap, floating pipes and such won't look right.
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

My PRT sphere is not working. It baked allright, but when I try to use it as light source for baking the turret, radiosity uses the diffuse color. I tried vcol_light and vcol_paint, and neither works. I posted a question at Elysiun
http://blenderartists.org/forum/showthr ... 1e&t=82298
but I fear what the answer will be: "Can't do it". Which makes sense because the raybaker would have to interpolate colors for subpatching, which is probably not coded in.

On another note, I shrank my model by displacing all polygons by -0.01 along the normals, prior to subsurfacing. Hopefully this will keep the high detail mesh inside the original. Will the normalmap baker see added floating pipes and stuff? Or do I have to modify the surface itself?
DualJoe
ISO Party Member
ISO Party Member
Posts: 387
Joined: Tue Mar 21, 2006 2:37 pm
Location: Netherlands
Contact:

Post by DualJoe »

Three tools I use (Braybaker, xNormal, ATINormalmapper) all let you set various options and rules for ray-depth and intersections. The best explanation I've read about the different techniques was in the pdf-manual that comes with xNormal.

Basically there are two methods for what you want.
1. Export the LOD-model plus a cage. This cage represents the outer- and innermost ranges that raytracing is done. Every vertex on the LOD therefore corresponds with two vertices on this "cage". Everything that's inside this cage on the high-poly-model will represented in the normalmap. I've never done it this way so you'll have to figure this method out yourself.

2. With the tools I use you can set how far in either direction the rays will travel. (Well in the case of BRaybaker the distance of the camera to the surface and clippingrange) Again everything inside this area will be in the normalmap. With ATINormalmapper you can additionally set which intersaction gets preference, like: keep the intersection closest to the LOD / keep the furthest intersection outside of the LOD, but the closest one on the inside and so on.

I'm in the middle of experimenting myself and haven't gotten the hang of it yet. It basically comes down to trial and error.
Normalmapping the engine area of the Hornet is really getting on my nerves.

EDIT
I'll post the blend I used for the GI-baking in a bit.

EDIT2
Damn those big harddisks I can't find the file anymore. Nevermind it's easy enough to setup.
I'd advice against using a sphere or any complete enclosure for that matter. It works nice for AO, but it won't wont work for this. The light will bounce around in the sphere and that's NOT what you want.
I made a cube and assigned the three emissive materials for the three directions in their corresponding colors.
Next I separated the positive and negative parts and moved them to two separate layers.
Save this file and append the model you to your liking. Choosing one of the two layers for either the positive or the negative bakings.. I also used a subdivided version of the hornet for the baking, to increase the resolution of the solution.
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

Thanks, I just downloaded xNormal and been looking at the examples. Pretty funky. I'll read the docs next.
I'd advice against using a sphere or any complete enclosure for that matter. It works nice for AO, but it won't wont work for this. The light will bounce around in the sphere and that's NOT what you want.
By "this" you mean normal-mapping? The sphere was just for PRT baking for GI. I think I'll replace the sphere with 3 hemispheres, layered, with a different material for each concentric ring. You can make the material non-traceable and non-shadow-casting. That way light won't bounce off it and will go right through it.

For the engines, you might want to consider editing the full LOD mesh to make it easier for you. I'd bring up the back wall of the engine cavity forward, to make it almos flush, since you can't see lighting effects for the whole depth of the engine, precisely because it's in a recessed space, hidden from light sources. A shallow engines space should be good enough for the LOD. The latter could have a flat back plane, and the engines could be just normal-mapped onto it.
Kangaroo
Venturer
Venturer
Posts: 543
Joined: Sun Jan 22, 2006 9:55 am
Location: Baltic States
Contact:

Post by Kangaroo »

Sorry for interfering, but the Texture Baker script only works for me correct, if the active 3D window is in UV face select model.
There are no stupid people on Earth; they are only alternatively thinking.
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

Thanks, Kangaroo; no interference whatsoever. I just tried it, though; from inside UV image editor, and I get the exact same error message. I'm positive this has everything to do with its being the "Halloween Release", which is a cvs snapshot, unofficial release, put together by a third party, from sources that are out of synch with some of the python scripts. But anyhow, I kind of know what to do, now, to get a texture baked. A bit wound around, but basically, I just export a .obj, create a new file with good old version 2.40, import the .obj, and bake the texture from that.
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

Here's a color theme I'm sure Dual's gonna like... :D
A Mafia boss ship could use it, I'm sure...

Image

I've got some manual fixes to do on the subsurfed mesh, and then I'll start greebling it.

EDIT:
Now, how's this?

Image

EDIT2:

Bad news: Subsurf wrecked the UV map. Man, Blender really sucks when it comes to subsurf and smooth subdiv and smoothing and all that. It just has some really stupid smooth function that a) pulls the vertices at the edges, so you can't have split smoothing groups; you're forced to rem doubles before subsurfing. And b) does all kinds of other weird things, like where the squarish parts meet the ovoid parts, I had a nice bevel there; but subsurf pulled those bevels in, at the corners, and I was having a terrible time of manually pulling them back out.

Image

Then I looked at the areas near the corners in the UV map, and what Blender did was, instead of interpolating the new UV vertices by distance along the surface, it did something else, I don't know what, that the added vertices fold the UV surface and draw figures all over the place. Plus, it destroyed my seams, and it's practically impossible to find out where the seams should fall. This is trash, now. Luckily I save under incrementing revision numbers, so I can go back. Unluckily, I don't know what better to do.

EDIT3:
Sorry, no, the seams were gone before subsurfing, somehow; I just reloaded the pre-sub-surf version and I don't see any of my seams there.
I'm going to put my seams back in, and try and bevel my bevels some more, before sub-surfing, see if that fixes the prob.
Hmmm... putting the seams back in... I wish it were as easy as said. How did they go away, I wonder?
DualJoe
ISO Party Member
ISO Party Member
Posts: 387
Joined: Tue Mar 21, 2006 2:37 pm
Location: Netherlands
Contact:

Post by DualJoe »

Remove doubles can remove seams and I agree about that being a very irritating feature.
Blender uses weighted-edges instead of smoothgroups and that's why the latter have very little support featurewise.
I haven't run into any problems with UV and subdividing . I can even let radiosity increase vertexcount and retain the unwrap.
Are these all features of the Halloween-Edition? I'd better steer clear of that release then.

I especially like the green one. Could I have one of those when you're finished.
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

You can have two, but where on the Hornet is it going to fit? :D

Yeah, don't use Ctrl-E Mark Sharp either, found out the hard way. Not only does it crash as soon as you press Tab, but if you save before going to Object or UV modes, you can't open the file ever again...

I'm almost done with the second, non-GI, overlapping UV-mapping. Looks to be ridiculously efficient. Might be able to use a 256x256 texture. I'll subsurf both versions after that.
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

@ Zeog and dandandaman:

We were talking about photon mass... I was just reading an article about spinning mass force in the new issue of New Scientist, and it mentions in passing that photons acquire mass as they enter a superconductor.
Stuart Clark wrote:Argyris's idea piqued de Matos and Tajmar's interest because of the parallel with the normally massless photon, which inside a superconductor develops a mass when the temperature drops below the critical temperature and the substance becomes superconducting... [sic]
(Wonder if New Scientist has any proofreaders on staff, tho...)
dandandaman
Artisan
Artisan
Posts: 1270
Joined: Fri Jan 03, 2003 3:27 am
Location: Perth, Western Australia
Contact:

Post by dandandaman »

That's actually really interesting! Man, now I want to read more :-)

Dan
"Computers are useless. They can only give you answers."
-- Pablo Picasso
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

Well, the New Scientist article is not about photons, but about this mysterious attractive force generated by spinning superconductors, which it says was measured to be about a trillion times stronger than predicted by Relativity. Recommended reading anyhow. Back in the 1980's there was sort of like a popular movement towards DIY Science, and I jumped into that bandwagon for a while. Most of the stuff was garbage, but there was one article that I can never forget. Well, I can definitely forget names and that kind of stuff... The author claimed to have found a new fundamental force he called "kinemassic". His theory was that what keeps neutrons and protons paired was their spins, and he built a machine to prove it. It was table-sized but very heavy. Basically a massive square loop of very pure copper with two cylindrical holes, one on each side. Each cylindrical hole housed a cylinder of copper that could rotate. And each of these copper cylinders had a missing gap in the middle which housed a flat, circular cake of copper that was spun at high RPM's by an air jet hitting it. One of the two cylinders was allowed to turn freely, while the other was moved very slowly by a motor. The freely turning cylinder recorded oscillations that kept phase with the moved cylinder.
To further remove the possibility that it could be magnetism getting around through impurities in the copper, the author put an iron bar in the middle of the loop, which had no effect. Instead, he said, a bar of copper in the middle kills the interaction. And one of the things he observed was that this field progressed very slowly through the copper core; --took about an hour to get from one side to the other and start having an effect on the detector gyroscope.
So, reading this article in New Scientist I wonder if maybe that "kinemassic field" paper author wasn't so much a crackpot but actually had something.

Alright, this was the non-overlapping, PRT UV:

Image

Now this is the overlapping texture UV:

Image

Next I'm gonna try to subsurf these thingies. Wish me luck...

(Damn! I have to subsurf and greeble both meshes in parallel, to have normal mapping and GI bakings... Why can't Blender support dual UV's?)
DualJoe
ISO Party Member
ISO Party Member
Posts: 387
Joined: Tue Mar 21, 2006 2:37 pm
Location: Netherlands
Contact:

Post by DualJoe »

about the current CVS-release Ton Roosendaal wrote:...
- multi-level UV coordinates editing
...
If that means what I think it means, then I think we'll be needing the updated CVS-version of Blender.
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

Wow! Just in time development! :D
I guess I'll have to figure out how to compile from cvs, after all.
Thanks!
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

Allright, this is the highly detailed version,

Image

Image

I was going to greeble it some more, but it started to corrupt my UV layout, so I figure I'll just add some bumpmap stuff manually. Can't procrastinate it anymore; next is baking the normal map... :roll:

So far I get poltergheist out of xNormal. On the other hand, I have a lot of overlapping areas. So, basically, I need to remove all redundancy. When I do, this is what I end up with:

Image

Image

Which makes me wonder, how are the normals at the edges where stuff was removed going to behave? Starting to get the feeling that normal baking will have to move to the non-overlapping, GI-related UV map...

EDIT:
My first attempt at normalmap baking looks abnormal:

Image

For one thing, it doesn't look tangent space to me. For another, there are more artifacts than pixels. "Sub-pixel INaccuracy". But we'll get there soon...
DualJoe
ISO Party Member
ISO Party Member
Posts: 387
Joined: Tue Mar 21, 2006 2:37 pm
Location: Netherlands
Contact:

Post by DualJoe »

Don't expect any of the tools to create a perfect normalmap without artifacts. It requires a lot of tweaking to get proper results out of any of them and even then you'll probably have to adjust it manually. I get some good results by compositing the solutions of BRaybaker, ATINormalmapper and xNormal. I find that xNormal doesn't give very detailed normalmaps.

Since you've got floating geometry you'll have to take great care with your raydistance values and uncheck “Closest if ray failsâ€
chuck_starchaser
Elite
Elite
Posts: 8014
Joined: Fri Sep 05, 2003 4:03 am
Location: Montreal
Contact:

Post by chuck_starchaser »

Yep...

Image

... enough fooling around. Time to read the manual... :D
loki1950
The Shepherd
Posts: 5841
Joined: Fri May 13, 2005 8:37 pm
Location: Ottawa
Contact:

Post by loki1950 »

re that modview app ut2k4 shipped with a simalar app Upiant they only worked with each games mesh files not universal.

Enjoy the Choice :)
my box::HP Envy i5-6400 @2Q70GHzx4 8 Gb ram/1 Tb(Win10 64)/3 Tb Mint 19.2/GTX745 4Gb acer S243HL K222HQL
Q8200/Asus P5QDLX/8 Gb ram/WD 2Tb 2-500 G HD/GF GT640 2Gb Mint 17.3 64 bit Win 10 32 bit acer and Lenovo ideapad 320-15ARB Win 10/Mint 19.2
spiritplumber
Developer
Developer
Posts: 1831
Joined: Mon Mar 07, 2005 10:33 pm
Contact:

Post by spiritplumber »

That's really neat... I'm going to talk to a physics prof over here who has some liquid nitrogen and hi-temp superconductors. Thanks!
My Moral Code:
- The only sin is to treat people as if they were things.
- Rules were made for people, not the other way around.
- Don't deceive. Real life is complicated enough.
- If all else fails, smash stuff.
Post Reply