Jump to content
  • Advertisement
Sign in to follow this  

model loading code I don't understand

This topic is 3958 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

I'm using this system to load models into an application ( this is using the cal3d format and library) and it keeps getting the wrong winding order. It will load for example 4 objects together, and one of them will get the wrong order but the others are right. I tracked down the part of the code that is probably causing it. It makes a check for something, but why does it come out with different results sometimes, why did they make it like this?, how can I just skip this check / test and make it always use the same method without messing it up? full file --http://svn.gna.org/viewcvs/cal3d/trunk/cal3d/src/cal3d/loader.cpp?rev=489&view=auto
// check if left-handed coord system is used by the object
		// can be done only once since the object has one system for all faces
		if (justOnce==0)
			// get vertexes of first face
			std::vector<CalCoreSubmesh::Vertex>& vectorVertex = pCoreSubmesh->getVectorVertex();
			CalCoreSubmesh::Vertex& v1 = vectorVertex[tmp[0]];
			CalCoreSubmesh::Vertex& v2 = vectorVertex[tmp[1]];
			CalCoreSubmesh::Vertex& v3 = vectorVertex[tmp[2]];

			CalVector point1 = CalVector(v1.position.x, v1.position.y, v1.position.z);
			CalVector point2 = CalVector(v2.position.x, v2.position.y, v2.position.z);
			CalVector point3 = CalVector(v3.position.x, v3.position.y, v3.position.z);

			// gets vectors (v1-v2) and (v3-v2)
			CalVector vect1 = point1 - point2;
			CalVector vect2 = point3 - point2;

			// calculates normal of face
			CalVector cross = vect1 % vect2;
			float	crossLength = cross.length();
			if (crossLength == 0.0f)
				CalError::setLastError(CalError::INVALID_FILE_FORMAT, __FILE__, __LINE__);
				return 0;
			CalVector faceNormal = cross / crossLength;

			// compare the calculated normal with the normal of a vertex
			CalVector maxNorm = v1.normal;

			// if the two vectors point to the same direction then the poly needs flipping
			// so if the dot product > 0 it needs flipping
			if (faceNormal*maxNorm>0)
				flipModel = true;

			// flip the winding order if the loading flags request it
			if (loadingMode & LOADER_FLIP_WINDING)
				flipModel = !flipModel;

			justOnce = 1;

		// flip if needed
		if (flipModel) 
			tmp[3] = face.vertexId[1];

Share this post

Link to post
Share on other sites
Are you sure that your model have correct vertex normals? Which 3d-content creation software you are using?
This code just compares triangle normal with vertex normal.

Share this post

Link to post
Share on other sites
The program is blender, the normals arn't wrong, but they may not be what the program is expecting. I guess basically The test assumes that you don't have a combination of smooth and flat normals? Which is not a good thing to assume, since almost every kind of model could have both.

Share this post

Link to post
Share on other sites
Sign in to follow this  

  • Advertisement

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

GameDev.net is your game development community. Create an account for your GameDev Portfolio and participate in the largest developer community in the games industry.

Sign me up!