fun with .pso

More
18 years 10 months ago #14543 by davel
Replied by davel on topic fun with .pso

I noticed on your site that you only cover the lwo2 format. What about the original lwo (pre-lightwave 6.0) format.

I'll look into finding documentation for that, I don't think I'll need it for my purposes, but it would be good to have.

I believe Jasper has considerable documentation for the old lwo format. It might be worthwhile combining your info with his. Then we'll get more info in one place. Just a thought.

I would very much like to do exactly that.
I'm hoping Jasper can help complete the .pso format doc currently on my site.

Jasper?

Please Log in or Create an account to join the conversation.

More
18 years 10 months ago #14544 by Second Chance
Replied by Second Chance on topic fun with .pso
Great news! I was wondering if mapping was included with the imported meshes, since I couldn't get it to work. I just figured I wasn't doing it right. Now I see what I did wrong. Very clever about including the map number in the mesh name. I'm really looking forward to posting some nice textured model pics. Great job!

No, there does not appear to be anything in the .pso that indicates that FLUX uses it or could support it either. I'm pretty sure they get stripped out in the conversion. I could be wrong.

DirectX, and therefore EoC definitely do use non-triangular polygons. I used to get arguments about this, but it is quite clearly the case. If non-tri polys were not supported, concave polygons would not cause display problems (see my tutorial on managing odd-shaped polys, way down in this thread www.i-war2.com/forum/topic.asp?TOPIC_ID=1584 ). But most obvious is simply the log from the converter, which specifies exactly how many polys are in the converted mesh and which vertices they are associated with. The original PSOs contain all triangles because that's the way they were constructed. If you would like some PSOs with non-triangular polys to work with I'll be glad to send you some.

Keep up the great work. I look forward to the next update.

mailto:second_chance@cox.net
The Ultimate Guide To Modding: I-War 2 - Edge Of Chaos
.

Please Log in or Create an account to join the conversation.

More
18 years 10 months ago #14545 by davel
Replied by davel on topic fun with .pso
Second Chance, glad to of helped.

Yeah, send me some .pso examples of non trianglular polys. I would say that if concave polys don't work, then it's not supported and that everything gets converted to tri faces. The converter log may just be reporting what it scanned through and not what it wrote to the file. My importer should yeild some strange results when trying to load one of those non triangular files because the face index should jump 2 indices ahead if it encounters a quad.

Please Log in or Create an account to join the conversation.

More
18 years 10 months ago #14546 by Second Chance
Replied by Second Chance on topic fun with .pso
Ok, I'll send you something as soon as you let me know where.

If you can look at the actual polygon count in the actual PSO file, then that will be the end of that. (Here's hoping I'm as right as I think I am.)

The lwo file I'll send you has 320 polys.
The converter counted and showed 320 after the conversion.
If the PSO file has more than that, then I'll definitely have to concede being wrong.

btw - If it's any indication, the PSO imported into Max without any problems with your script.

mailto:second_chance@cox.net
The Ultimate Guide To Modding: I-War 2 - Edge Of Chaos
.

Please Log in or Create an account to join the conversation.

More
18 years 10 months ago #14547 by Joco
Replied by Joco on topic fun with .pso

Originally posted by davel


---snip---
I would say that if concave polys don't work, then it's not supported and that everything gets converted to tri faces. The converter log may just be reporting what it scanned through and not what it wrote to the file.
---snip---


I agree on the first statement, not so sure about where the conersion is happening. But the converter is the most likely point.

The fact that things get "converted" into a .pso file hides some potential sins. BUT the fact that the game rendering engine does not deal with concave poly's indicates that they are not in fact dealt with correctly and might well be getting converted to triangles at some point under the hood, hence said surfaces not rendering as expected.

Further, if all meshes coming out of the original game assets are all trimeshes further indicates that trimeshes are the correct form of mesh.

Given the game engines age and that games of a similar age are mostly (if not all) trimesh based based, I would conclude EoC expects trimeshes. The fact we are able to use n-point polygons during modeling is just good luck, not design. IF it was designed/expected I would be expecting n-polygon meshes to be coming out of original game .pso files.

All up it is great to see the progress being made to dealing with this area of EoC custom content. Keep charging forward. :D

Cheers,
Joco.

Please Log in or Create an account to join the conversation.

More
18 years 10 months ago #14549 by davel
Replied by davel on topic fun with .pso
Second Chance: I'm looking through your concave poly tutorial and thought of something that you could add. Polys are somewhat arbitrary, basically they fill a polygon boundary with solid surface. You define the boundary with edges. Now, in our 3D programs, there's handlers that look for cutbacks in the boundaries and adapt the fill faces to them. However, it would also be legal for the fill faces to cross over the cutback edges because you're still filling the boundary and you're still using every vertex. Exporters probably aren't so careful! Look at the struts under the radar dish. The boundaries on the ouside define a 4 vert solid trapezoid, the cutback is just 4 more verts to connect to. There's two solutions, define all the edges (change to trimesh). Or add edges to create new boundaries that possible edges cannot break out of. See pic.

Hope it helps.

Please Log in or Create an account to join the conversation.