Python converter for .variant_weighted_mesh model

[Aradiel]

@Aradiel

I think most people don't handle lods at all.[...]Unless you're modding every unit, maybe the M2TW
"just don't worry about it" works.

Yea, you are probably right, i guess we lack of expirience with empire also.

[S-te-Fan]

There's a way in 3dsmax so you can merge models withouth having to re-rig everything.

[Aradiel]

I ran into a strange texture glitch which i dont understand:

This is the model with a basic texture i have made for mapping purposes,
as you can see the mapping is applied properly.



When i convert the ms3d file to mesh i get an error report,



although i would assume thats because i touched the heads like stefaneke did, indeed only head01
of the 4 shows up ingame.


The oddity is that partially the bodymesh mapping gets messed up after i converted it,
as seen here:



Most of it is right, just the shoulder and neck area gets messed up, as seen on this backconverted mesh.



edit: It seems that making a seperate mapping part for the collar caused the problem, as the collar mapping gets smashed onto
the chestmapping after the conversion, which obviously causes problems as seen. In order to avoid this behavior it is nessecary to make a joint mapping i guess.

[Killerbee]

Hey , just popping by , any progress ?

[Aradiel]

Hey , just popping by , any progress ?

Well this turns into my private diary it seems, but if noone else has or wants to report

I have fixed the diffuse texture issue with a workaround, still dont know how to create normal and gloss maps like ca does, but i bet some photoshop wizards have it in no time.



A more concerning issue is the lod management, i can tell for sure that having all lod1 is not an option, unless you run it on a highend system, which i cannot confirm as my 285 gtx is temporaly replaced by a 8800gts, just scrolling over my units causes real lag. So we need a way to create lower lod's as well. I have uploaded the testunit here if you want to see how it comes.

[joedreck]

Oh, yeah. I see some progress.

I don't know too how CA did the gloss and normal, so I do it manual. You can imagine how much work it is. Much more then textures for the unit.

Maybe someone know that how do do it right?

[S-te-Fan]

You can use the optimze modifier in 3dsmax to create LODs.

[Killerbee]

Well this turns into my private diary it seems, but if noone else has or wants to report

I have fixed the diffuse texture issue with a workaround, still dont know how to create normal and gloss maps like ca does, but i bet some photoshop wizards have it in no time.



A more concerning issue is the lod management, i can tell for sure that having all lod1 is not an option, unless you run it on a highend system, which i cannot confirm as my 285 gtx is temporaly replaced by a 8800gts, just scrolling over my units causes real lag. So we need a way to create lower lod's as well. I have uploaded the testunit here if you want to see how it comes.

Right , good job ! Keep up the good work

And nice knight stefaneke

[AmiR TimuR]

guys, what about horses, can we model them?

not with this tool for sure but maybe it is easier to write a conventer for them?

[S-te-Fan]

Well, if you import a model now it applies itself to the human skeleton, so if you want to edit horses you have to figure out the stuff from a different skeleton.

[Aradiel]

You can use the optimze modifier in 3dsmax to create LODs.

Indeed, good tip, i tested it and managed to get it down to a little above a vanilla lod3, replaced lod3 and 4s with it and noticed a considerable performance increase.

[KnightErrant]

@all

Sorry guys, I haven't been able to do a thing this past week or so, design review at work
going on. It should be a (hopefully) small thing to fix the tangent and bitangent problems
Stefaneke reported. I started on a rigid_model converter but got totally stuck on weird uv
coordinates coming out of the models. Some are correct as is, many are totally off, ranging
from -11 to +5.0, which certainly don't look like uvs to me. But not sure how to proceed
given that. Maybe just release the converter as is, and let people look at the uv problem,
maybe a piece of insight will present itself.

[Zigster]

Hey Aradiel, very nice looking model.

I downloaded it to try and test it, but I'm a noob so I don't really know what to do with it.

In the DBEditor tutorial, you assign a model to the swiss_line unit you make. I put euro_line_infantry in Model name and saved. Then in My_warscape_animated_lod_tables, I put the path as UnitModels/euro_line_infantry_lod1.variant_weighted_mesh. And filled out the rest.

But I just get an APPCRASH when I try and load the game.

Could anyone tell me what I'm doing wrong?

Edit: Nevermind, I was overcomplicating things I didn't need to create a new Unit to get them into the game because they just replace the Line Infantry. They look great! I guess they're for the WWI Mod?

[Aradiel]

Hey Aradiel, very nice looking model.
They look great! I guess they're for the WWI Mod?

Thanks, and no they are for the civilwar mod.

[Ahiga]

Running into a new error. If it helps, this is on my Vista 32 bit Laptop. All I did was rename two different hats and I tried to convert it:

Reading ms3d file: C:/Video Games/Steam/steamapps/common/empire total war/Moddin
g Tools/east_armed_citizenry_india_lod4.ms3d
Read ms3d vertex section
Read ms3d triangle section
Read ms3d group section
Read ms3d material section
Read ms3d keyframer data
Read ms3d joint section
Read ms3d first subversion number
Read ms3d group comment section
Read ms3d material comment section, number of comments = 0
Read ms3d joint comment section
hasmodelcomment == 1
modelcomment is 13
light_scale  +1.00000000
offsetu0  +0.00000000
offsetv0  +0.00000000
offsetu1  +0.00000000
offsetv1  +0.00000000
bumpfactor  +1.00000000
specpower  +2.00000000
specbrightness  +1.00000000
specularfresnelpower  +0.01000000
glossfactor  +2.00000000
fresnelpower  +2.00000000
reflect_factor  +0.50000000
ambientfactor  +0.60000002
2
colourmapfactor  +0.66666669  +0.30196080  +0.30196080  +0.00000000
specfactor  +1.00000000  +1.00000000  +1.00000000  +1.00000000

Read ms3d model comment
Read ms3d second subversion number
Read ms3d weight section

13 %% light_scale  +1.00000000 %% offsetu0  +0.00000000 %% offsetv0  +0.00000000
 %% offsetu1  +0.00000000 %% offsetv1  +0.00000000 %% bumpfactor  +1.00000000 %%
 specpower  +2.00000000 %% specbrightness  +1.00000000 %% specularfresnelpower
+0.01000000 %% glossfactor  +2.00000000 %% fresnelpower  +2.00000000 %% reflect_
factor  +0.50000000 %% ambientfactor  +0.60000002 %% 2 %% colourmapfactor  +0.66
666669  +0.30196080  +0.30196080  +0.00000000 %% specfactor  +1.00000000  +1.000
00000  +1.00000000  +1.00000000
13
 light_scale  +1.00000000
 offsetu0  +0.00000000
 offsetv0  +0.00000000
 offsetu1  +0.00000000
 offsetv1  +0.00000000
 bumpfactor  +1.00000000
 specpower  +2.00000000
 specbrightness  +1.00000000
 specularfresnelpower  +0.01000000
 glossfactor  +2.00000000
 fresnelpower  +2.00000000
 reflect_factor  +0.50000000
 ambientfactor  +0.60000002
 2
 colourmapfactor  +0.66666669  +0.30196080  +0.30196080  +0.00000000
 specfactor  +1.00000000  +1.00000000  +1.00000000  +1.00000000
ipart =  0, (min,max) = 0, 69
ipart =  1, (min,max) = 70, 139
ipart =  2, (min,max) = 140, 163
ipart =  3, (min,max) = 164, 233
ipart =  4, (min,max) = 234, 303
ipart =  5, (min,max) = 304, 400
ipart =  6, (min,max) = 401, 497
ipart =  7, (min,max) = 498, 572
ipart =  8, (min,max) = 573, 621
ipart =  9, (min,max) = 622, 687
Exception in Tkinter callback
Traceback (most recent call last):
  File "C:\Program Files\Python\lib\lib-tk\Tkinter.py", line 1410, in __call__
    return self.func(*args)
  File "C:\Video Games\Steam\steamapps\common\empire total war\Modding Tools\ETW
variant_weighted_meshConverter_v1_0.py", line 6058, in ms3d2meshConverterButton
    backVWMobj.partsection     = ms3dobj.makeVWMPartSection( bonepalette )
  File "C:\Video Games\Steam\steamapps\common\empire total war\Modding Tools\ETW
variant_weighted_meshConverter_v1_0.py", line 4617, in makeVWMPartSection
    idet       = 1.0 / ( s1 * t2 - s2 * t1 )
ZeroDivisionError: float division

Any idea what the problem is?

[KnightErrant]

@Ahiga

Hopefully version 1.3 will fix that.

@All

Here's a link to version 1.3 of the variant_weighted_mesh converter:

http://www.twcenter.net/forums/downl...o=file&id=2612

I've tried to put in traps for any zero divides when calculating the tangent
space basis for the back conversion. When all else fails I just stick in an
arbitrary tangent and bitangent and leave it at that.

I also included a beta converter for rigid_model. I think it does the geometry
ok but the uvs that come out of the rigid_model meshes are sometimes greater than
1.0 or less than 0.0. Some of the uvs look just right in Milkshape so I don't think it's
a common scaling or transforming problem. I've left in the code that writes out the file
as text if anyone wants to look at the uvs that way. In the vertex sections that look like:

  +3.5622740   +1.4599494   -3.6472118     -0.7400353   +0.5327236   -0.4105525     -0.7343975   +0.3912459     -0.7447522   +0.2723613   +0.6092319     -0.5261092   -0.8012646   -0.2849285     +1.0000000   +1.0000000   +1.0000000   +1.0000000    +54.8713150 -129.3623962   
  +3.9002080   +1.6108029   -4.0606055     -0.7400353   +0.5327236   -0.4105525     -0.5507886   +0.3912459     -0.7447522   +0.2723613   +0.6092319     -0.5261092   -0.8012646   -0.2849285     +1.0000000   +1.0000000   +1.0000000   +1.0000000    +54.8713226 -129.3623962   
  +3.7921476   +1.3065975   -4.2605696     -0.2850863   -0.8007205   -0.5268515     -7.0573659   +0.2836526     +0.4109596   -0.5335796   +0.7391921     +0.7440025   -0.2722855   -0.6101810     +1.0000000   +1.0000000   +1.0000000   +1.0000000    +54.8713417 -144.4552917   
  +4.9820123   +1.0369465   -4.4945993     -0.2850863   -0.8007205   -0.5268515     -6.7502384   +0.0098621     +0.4109596   -0.5335796   +0.7391921     +0.7440025   -0.2722855   -0.6101810     +1.0000000   +1.0000000   +1.0000000   +1.0000000    +91.4522247 -144.4552917   
  +4.2155566   +0.6947786   -3.5598257     -0.2850863   -0.8007205   -0.5268515     -6.7502384   +0.4258575     +0.4109596   -0.5335796   +0.7391921     +0.7440025   -0.2722855   -0.6101810     +1.0000000   +1.0000000   +1.0000000   +1.0000000    +91.4522247 -144.4552917   
  +4.2155566   +0.6947786   -3.5598257     -0.2839274   -0.8006893   -0.5275244     -6.7502384   +0.4258575     +0.4108024   -0.5329615   +0.7397253     +0.7436124   -0.2735850   -0.6100752     +1.0000000   +1.0000000   +1.0000000   +1.0000000    +91.4522247 -144.4552917

Columns 7 and 8 are supposed to be u and v, respectively, but, as you can see, some of the u's
are -6.7. If anyone has some ideas as to what might be going on, please post.

[danova]

yup! Thanks mate your works!
I have little moding orgasm...
.

Sorry! Where are the muskets and equipments? in wich file?
.

[KnightErrant]

@danova,

Sorry, don't know the answer to that, I don't own the game so I don't know
how it's organized. I make the converters from files people send me. But don't
worry, Stefaneke, Aradiel, Ahiga, etc., the real modellers for ETW, know that
stuff so they'll be able to answer your question.

KE

[danova]

ok, thanks your help!
...and more thanks for the converter
.

[joedreck]

Very good job as usual.