SUBSIM Radio Room Forums

SUBSIM Radio Room Forums (https://www.subsim.com/radioroom/index.php)
-   SHIII Mods Workshop (https://www.subsim.com/radioroom/forumdisplay.php?f=195)
-   -   Problem with 3D model of Wildcat (https://www.subsim.com/radioroom/showthread.php?t=229104)

Kendras 01-04-17 03:44 PM

Problem with 3D model of Wildcat
 
Hi,

I have a problem with aspect of a 3D model in game. It was good before I work on it. What I've done : I exported the 3D model of the Wildcat airplane as .obj (+ .mtl file). Then I opened it in wings 3D, I separated it and removed the pilot, then I reassembled it and exported as new .obj. But now, in game, I have no more a smooth surface as before, as you can see :

http://i.imgur.com/bW4u4mk.png

http://i.imgur.com/MDd10zV.png

Why is there this problem ??? How can I solve it ? :06:

gap 01-04-17 03:56 PM

Can you upload the two versions of the dat file (before and after your modifications) together with your working obj file(s), and post the link here?

Kendras 01-04-17 04:43 PM

Here are the 2 .dat files : http://www.mediafire.com/file/7jc20x...cat_models.zip

:hmmm:

HanSolo78 01-04-17 05:37 PM

Hm... :hmmm:

Looks like you saved the new files in wings 3d with more materials like before.

the_frog 01-04-17 06:18 PM

I think there are several problems.

One is that the removal of the pilot reduced the number of material associated with the model. That menas before re-importing the model into the *.dat remove the respective material/texture from the material list in "linked 3D model" window.

Also, I guess the model has isolated triangle faces around the wheels, partly wrongly inversed. (one may call this effect of excessive edge splitting :-) ) Repairing can be achieved using Meshlab or something similar; Wings3D cannot solve that issue.

Cheers

Wise 01-05-17 12:21 AM

The original model has topology errors.
The reworked model has the partial correction of these errors. But illiterate. Possibly, accidental. In reworked model there are excess (duplicated) polygons.
Cleaning of model is necessary. Delete the old (broken) polygons. Merge new sections together. "Welding".

gap 01-05-17 09:44 AM

Quote:

Originally Posted by the_frog (Post 2456085)
I think there are several problems.

One is that the removal of the pilot reduced the number of material associated with the model. That menas before re-importing the model into the *.dat remove the respective material/texture from the material list in "linked 3D model" window.

Also, I guess the model has isolated triangle faces around the wheels, partly wrongly inversed. (one may call this effect of excessive edge splitting :-) ) Repairing can be achieved using Meshlab or something similar; Wings3D cannot solve that issue.

Cheers

Quote:

Originally Posted by Wise (Post 2456103)
The original model has topology errors.
The reworked model has the partial correction of these errors. But illiterate. Possibly, accidental. In reworked model there are excess (duplicated) polygons.
Cleaning of model is necessary. Delete the old (broken) polygons. Merge new sections together. "Welding".

Ditto.
I had a look into the model myself, and I saw many of the problems stated by the_frog and by Wise :yep:

The sharp edges on the top of the fuselage are caused by isolated faces. Fixing the problem is a matter of welding them in MeshLab, or with any other program capable of advanced mesh manipulation. I use Mod Tool for that.

Sharp edges + dark triangles on the bottom of the fuselage near the landing gear, in the radiator, and in some of the canopy beams, is caused again by non-contiguous edges, and by those triangles being "twofold", i.e. being composed by two coincident faces welded together, each pointing to opposite directions. The cure is removing manually the faces with wrong normals, and welding the remaining triangles with each other.

Last, I see some redundant meshes, i.e. meshes which are entirely enclosed within the fuselage and cant be seen when the model is assembled. They don't do any harm, but they are consuming computer resources without a valid reason. Removing them is an easy task.

All in all, I think the model is the result of a quick conversion from a format not compatible with the Wavefront (obj) file format. Fixing it is possible, but it will require some tweaks and a general clean-up. If no one else is volunteering for that job, I will put it in my to-do list :up:

Wise 01-05-17 10:37 AM

@gap
Everything that you have listed, I completely agree with you.:yep:
Manual cleaning of the duplicated welded sides turned inside out, is a nightmare.)):88) :D
https://yadi.sk/d/TQ6ZYJQ3379BKd
Take. I started, and you continue.:)
I have eliminated mistakes, which on screens. Also, has eliminated errors of a lattice of a cabin of the pilot. For elimination of all mistakes long boring fuss is necessary. It is simpler to make new model.:/\\x:

Kendras 01-05-17 01:01 PM

Quote:

Originally Posted by Wise (Post 2456146)
For elimination of all mistakes long boring fuss is necessary. It is simpler to make new model.:/\\x:

Wow, really ? :eek:

Wise 01-05-17 01:52 PM

Quote:

Originally Posted by Kendras (Post 2456171)
Wow, really ? :eek:

Certainly, I exaggerate.:)
But slightly.

the_frog 01-05-17 03:15 PM

Hello folks,

tried to clean up the model by using Meshlab's "removal of duplicated face", "removal of isolated faces" and some more of those nice filters (there's one that allows for completely removal of edge splitting: "merge adjacent vertices" :03:). Well, the results were mixed because if a mixture of defects. In the end, I noticed much manual clean-up is still required ... as Wise already pointed out, the model's state "is a nightmare".

Gap, good luck ...

gap 01-05-17 11:50 PM

Quote:

Originally Posted by Wise (Post 2456146)
@gap
Everything that you have listed, I completely agree with you.:yep:
Manual cleaning of the duplicated welded sides turned inside out, is a nightmare.)):88) :D
https://yadi.sk/d/TQ6ZYJQ3379BKd
Take. I started, and you continue.:)

Thank you mate, looks like me you and the_frog started fixing the plane each one on his own, and then started wondering if it was worth all the work involved :haha:

Quote:

Originally Posted by Wise (Post 2456146)
I have eliminated mistakes, which on screens. Also, has eliminated errors of a lattice of a cabin of the pilot.

Do you mean the error in the canopy (i.e. the glass covering the pilot's cabin)? If so, I noticed it and fixed it too :up:

Quote:

Originally Posted by Wise (Post 2456146)
For elimination of all mistakes long boring fuss is necessary. It is simpler to make new model.:/\\x:

Quote:

Originally Posted by the_frog (Post 2456190)
as Wise already pointed out, the model's state "is a nightmare".

Gap, good luck ...

Thank you the_frog :)

Talking about the state of the model, I realize how poor/basic it is in general. Maybe we should take Wise's exageration literally, and start working on a better model rather than fixing the current one. :hmm2:

Remodelling the aircraft from scratch is a possibility, but these days free, good quality 3D models of most WWII-era aircraft are available on the web. For a start, I would like to have a look into the detailed model below, but unfortunately I don't have 3DS Max. Can someone download it and save it for me in obj or simple 3ds format?

http://www.cadnav.com/3d-models/model-10481.html

Wise 01-06-17 03:27 AM

Quote:

Originally Posted by gap (Post 2456236)
Do you mean the error in the canopy (i.e. the glass covering the pilot's cabin)?

No, glasses were normal initially. The lattice (a frame of glasses) had mistakes. Similarly duplicated welded sides with normals in both parties. I have removed this lattice and have made another.
Quote:

Originally Posted by gap (Post 2456236)
Can someone download it and save it for me in obj or simple 3ds format?

https://yadi.sk/d/qWd_7NUg37GngW
There is an original file 3ds, original textures and my OBJ+MTL.
Gap, this plane is good, but... Too it is necessary to work with him. To optimize in game. To simplify. In other words, high-poly to transfer to Low-poly.
His present state is no good. 68.000 polygons.

https://s20.postimg.org/fkw3bqirx/001.png


For the plane it is a lot of. Many vessels have much smaller "weight".

gap 01-06-17 05:48 AM

Quote:

Originally Posted by Wise (Post 2456244)
No, glasses were normal initially. The lattice (a frame of glasses) had mistakes. Similarly duplicated welded sides with normals in both parties. I have removed this lattice and have made another.https://yadi.sk/d/qWd_7NUg37GngW

Yes, I got you :up:

Quote:

Originally Posted by Wise (Post 2456244)
There is an original file 3ds, original textures and my OBJ+MTL.
Gap, this plane is good, but... Too it is necessary to work with him. To optimize in game. To simplify. In other words, high-poly to transfer to Low-poly.
His present state is no good. 68.000 polygons.

For the plane it is a lot of. Many vessels have much smaller "weight".

Thank you! Yes, when I read "highly detailed" in its description, I suspected it encompassed "high poly" too :-?
Some other free-to-use alternatives:

http://www.sharecg.com/v/60638/relat...an-F4F-Wildcat
https://3dwarehouse.sketchup.com/mod...2a5e4027d22588
https://3dwarehouse.sketchup.com/mod...a-f6eff7c8ea25
https://3dwarehouse.sketchup.com/mod...a-f6eff7c8ea25

The sharecg model is low poly, but it looks a bit cartoonish. On the other hand the two 3dWarehouse models look much more realistic, but with ca. 100k and 45k polygons each, they are equally unsuitable for the game :hmmm:

the_frog 01-06-17 06:18 AM

Hello Gap and Wise,

there are two models in the Pack 3D Models collection, from different games. One is a reasonably well looking low-poly version but offerend only in 3ds (Wise, can you take care of that? I do not use 3Dmax ...):
http://p3dm.ru/files/37_grumman-f4f-wildcat-.html

By the way, converting of Sketch-up models often results in numerous fragmented and duplicated triangles. Same problem is also with Flight Simulator models (if I remember right, Sergbuto's Wildcat was converted from Flight Simulator).

Due to these nightmarisch problems I gave up on expanding my personal SH3 world by converting models.

Cheers


All times are GMT -5. The time now is 02:51 PM.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2025, Jelsoft Enterprises Ltd.
Copyright © 1995- 2025 Subsim®
"Subsim" is a registered trademark, all rights reserved.