Forza Motorsport Resource Extraction (.carbin)

[terzer]

is it possible now to view the xds textures. i just only want to see them, to make screenshot and use them. i saw in the previous posts its possible. thanks!

[TomWin]

ok problem found!
i hardly ripped model from Forza Studio by 3dprintscreen (too bad I can only view a model there), but what I can see, mesh is perfect, no bumps.

so now is clear that .obj exporter is damaged or so


what about export to this free program http://zmodeler2.com ? there is sdk
http://zmodeler2.com/?mode=dl&ID=18

[Simon]

Zmodeler 2 sucks!

If you Export the Models as 3ds you get only Damaged Normals... (Export from Zm2)

I would prefer Open Collada and then you can Import in Maya, Max without any Errors.

Zmodeler 2 is only great for exporting to game files like GTA IV, NFS Shift. But Export to 3D Apps sucks because of damaged Normals...

[TomWin]

That's why I never export as .3ds from zmodeler LOL. .ase or .x much better way of export from it

if you say .dae is that good, is worth try, because this .obj fails. I hope is not a problem make .dae exporter. I mentioned 3ds because i though it might be easy to code.

[Simon]

Problem is Open Collada is written in #C, Forza Studio in C++

http://code.google.com/p/opencollada/so ... wse/trunk/

But I'm sure Ernegien is able to write a C Wrapper and use this

[ALYX]

Problem is Open Collada is written in #C, Forza Studio in C++

It's easy to translate C# and C++ are similar but with a few differences.

[Veegie]

Rendering is definitely improved.

[ALYX]

Please Post new build of Forza Studio.
And also in FM2 car 3d models stored in carbin formats. Maybe someone can make improvments to support both format types.

[Ernegien]

I'm not so sure it's an obj exportation issue. All I export is vertices, normals, and faces and I've tried unwinding the faces a few different ways, none with different results. I still feel this may be a precision issue, or something entirely else. The only reason why you can't see these imperfections in my renderer is because it's not using any reflective surface, just simple shading...

[Varel]

I have tried the 3D print screen on the forzastudio as tomwin did and I can confirm that the shading is perfect. Then I would also assume that the mesh reading in forzastudio is OK. When I then use xml to max conversion script by koichi I have the same shading problems as if I import the obj and very similar to what the people are refering in the pictures.

Good work btw and thanks for zour effort!

[Simon]

is it possible now to view the xds textures. i just only want to see them, to make screenshot and use them. i saw in the previous posts its possible. thanks!

It's possible with XDK but don't ask me how - no idea

[terzer]

i understand. please, can somebody make a screenshot of the brake light texture and interior of ferrari california? i'm making a mod, the mesh is complete, i only need textures. i can provide the files, if needed.
sorry for the offtopic.

[Veegie]

[TomWin]

i understand. please, can somebody make a screenshot of the brake light texture and interior of ferrari california? i'm making a mod, the mesh is complete, i only need textures. i can provide the files, if needed.
sorry for the offtopic.

and what more? why dont u make textures self? stop act like unpatient kid who want to do a mod quick. we try here solve more important problems than your impatience.

@ Varel: yes saddly, I tried now, and the problem appeared again. so must be something wrong with forza studio or other possibility Turn1 protected models somehow so is not easy import them correct.

@ Veegie: nice, saddly bumps arent gone.

[Ernegien]

Veegie and I have pretty much confirmed that the smoothing problem can be fixed by manually adding smoothing groups. I took another look at the vertex data and don't immediately see any ids that could be used as such, but there are also the possibility of blended vertex weights or UVW coordinates that wrap around different smoothing groups. I'm not exactly sure how these things operate quite yet, so instead I started writing a function to manually loop through the vertices and read the normals which would kind of botch together some rough smoothing groups. This algorithm has lots of problems in itself that I won't bother going into, but it should serve my purpose of testing for smoothing groups. I'm going to need to take a little break on this project and concentrate on school matters for a few days, but here's the code where I left off if anyone is interested...

Code: Select all

// assign smoothing groups (assume groups won't cross sections, so calculate per section)
int smoothingGroup = 1;
while (true)
{
    Vector3 baseNormal = Vector3.Zero;
    bool unassignedVertexFound = false;
    for (int i = 0; i < Vertices.Length; i++)
    {
        if (Vertices[i].smoothingGroup == 0)
        {
            if (!unassignedVertexFound)
            {
                // first one
                baseNormal = Vertices[i].normal;
                Vertices[i].smoothingGroup = smoothingGroup;
                smoothingGroup++;
            }
            else
            {
                // continue checking differences...
                float angleDifference = (float)(Math.Acos(Vector3.Dot(baseNormal, Vertices[i].normal)) * (180 / Math.PI));
                if (angleDifference < 30.0f)    // angle threshhold
                {
                    // part of same smoothing group
                    Vertices[i].smoothingGroup = smoothingGroup;
                }
            }
            unassignedVertexFound = true;
        }
    }
    if (!unassignedVertexFound) break;
}