Assimp issue

Are the number of aiMeshes (mNumMeshes) the same when loading .obj and .stl? It looks like the wheels/lower parts of the mesh aren't rendered for .obj files.

Have you tried other 3d models than this one?

Thanks for reply.

No they aren't, the number of mesh change because the model has been converted. This is my problem, my application don't display all meshes of model.

Have you tried to open both of these meshes with the Assimp Viewer?

Yes, both of these meshes with the Assimp Viewer are displayed correctly. I don't understand because my engine doesn't load all the meshes

I can check it in my application, just upload this .obj model if you can.

Thanks :D,

Direct link:

http://rs513p2.rapidshare.com/cgi-bin/rsapi.cgi?sub=download&fileid=1246163158&filename=Car.obj&cookie=D43D984F2E4AAF9B40F185DD2D28279357615E833B8E80DC2D92A141C94A16B373D0E30E4AE683200DF2331CE0130B0B&directstart=1

hm it displayed correctly in my engine, and i cant find error in your code o.0 only thing im not sure is flag "aiProcess_SplitLargeMeshes" maybe it just not working?

I think you can just get correct indices and vertices count for each part of mesh, and draw it via separate drawcall

Just suggesting, can you try to remove these and check the results:

aiProcess_SplitLargeMeshes
aiProcess_SortByPType

and

The problem should be the hierarchy of your object. In your code-snippet you don't traverse the node-objects. In every node a local matrix is stored which you can use to transform your vertices. In the stl-file the matrices should all be an identity-matrix so your modell is displayed correctly without transforming. In .obj there should be differences.

In my own model-loader I'm also struggling with transforming the vertices with the local matrices. There are all interpreted but I think the local positions and matrices don't match in my case. Or my hierarchy is wrong.

So when you figured out how to display it correctly, you could post your code here.

Update: Your car.obj works fine in my application

Ok Thanks for advise,

It doesn't work removing:

aiProcess_SplitLargeMeshes
aiProcess_SortByPType

But I am sure that work with separate drawcall, I'm trying tomorrow.

If I find an alternative solution I share here.

I know the car is a bit transparent, but the mesh you provided loaded correctly, so I believe that something is wrong with your rendering code as others suggested, prof:

Maybe you can show us your rendering, if you want?

I can relate to your problems, as when I tried to integrate Assimp . Didn't go so well the first time...

Rendering code:

void ModelClass::Render(ID3D11DeviceContext* deviceContext)
{
    unsigned int stride;
    unsigned int offset;


    // Set vertex buffer stride and offset.
    stride = sizeof(VertexType);
    offset = 0;
    
    // Set the vertex buffer to active in the input assembler so it can be rendered.
    deviceContext->IASetVertexBuffers(0, 1, &m_vertexBuffer, &stride, &offset);

    // Set the index buffer to active in the input assembler so it can be rendered.
    deviceContext->IASetIndexBuffer(m_indexBuffer, DXGI_FORMAT_R32_UINT, 0);

    // Set the type of primitive that should be rendered from this vertex buffer, in this case triangles.
    deviceContext->IASetPrimitiveTopology(D3D11_PRIMITIVE_TOPOLOGY_TRIANGLELIST);

    return;
}

void ShaderClass::Render(ID3D11DeviceContext* deviceContext, int indexCount)
{
    // Set the vertex input layout.
    deviceContext->IASetInputLayout(m_layout);

    // Set the vertex and pixel shaders that will be used to render this triangle.
    deviceContext->VSSetShader(m_vertexShader, NULL, 0);
    deviceContext->PSSetShader(m_pixelShader, NULL, 0);

    // Set the sampler state in the pixel shader.
    deviceContext->PSSetSamplers(0, 1, &m_sampleState);

    // Render the triangle.
    deviceContext->DrawIndexed(indexCount, 0, 0);

    return;
}

You want to pass this flag to your read function "aiProcess_PreTransformVertices" (http://assimp.sourceforge.net/lib_html/postprocess_8h.html#a64795260b95f5a4b3f3dc1be4f52e410) this will pretransform all of the sub nodes and all of the meshes will just be below the root node. No need to traverse the object any more if you don't care about the scene graph in the file.

Other flags of interest to a DX app is aiProcess_MakeLeftHanded which applies all the transforms needed for DX applications.

Also be careful with just assuming that an assimp mesh has normals by default they only guarantee that an aiMesh will have verts and faces.

Great now it work! Problem solved. Thanks .

So what was the problem? Just curious

I added the flag "aiProcess_PreTransformVertices" in my read function .