Jump to content
  • Advertisement
Sign in to follow this  
jswill100

OpenGL C++ STL Map for vertex normals. Help?

This topic is 4064 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

Hello all, Am having a small problem - right now I am exporting a large scene from 3ds into opengl using the .3ds format - and this does not contain any vertex normal information so i need to calculate it myself. Before you see some code you should know my vertices (and vert normals) are organised into vector<Vector3f> where Vector3f is my own structure for a vertex. The verts are organised so that there are repeated vertex, i.e. verts 1,2,3 are for the first triangle, and 3,4,5 are for the second, but none, 1 or 2 of these vertices may in reality be the same as 1,2 or 3. I started out using an extremely extremely time consuming method as below
void calculateSharedVerts(vector<Vector3f>& verts, vector<Vector3f>& vertNormals)
{
Vector3f temp;


        //here I set the intial vertex normals to that of the face normal for the face they belong to. 
	for(int i=0; i< verts.size(); i+=3)
	{
		temp.calculateNormal_2(verts, verts[i+1], verts[i+2]);

		vertNormals.push_back(temp);
		vertNormals.push_back(temp);
		vertNormals.push_back(temp);
	}


        //Now I loop through and whenever we find a shared vertex, we add the face normals from both
	for(int i=0; i< verts.size(); i++)
	{

		for(int j=0; j< verts.size(); j++)
		{
			if(i!=j)
			{
				if(verts==verts[j])
				{
					vertNormals+= vertNormals[j];
				}
			}
			
		}
	}

        //normalise the results
	for(int i=0; i< vertNormals.size(); i++)
	{
		vertNormals.normalise();
	}
}

The above worked entirely correctly, but because of the size of my scene, even just computing these normals once and writing to a file to be read later takes hours on end and so I searched for a better way of doing it. I realised I could use the stl Map so that whenever I was addressing a shared vertex, I only needed to use that one value so no looping to find shared verts is required. Because the other definition I use for Vector3f "less than" does not meet the requirements of "strict weak ordering" I define a new comparison function vect3fCompare to use. Basically the normals are coming out far from correct and was hoping someone might be able to give me some insight into what I might be doing wrong here.
	bool vect3fCompare(Vector3f p1, Vector3f p2) {

	Vector3f zero(0,0,0);

	//the below SHOULD satisfy strict weak ordering, which I read was is nessecery for this to work, defined by
		//1) a<a is false 
		//2) equality given by (!(a<b) && !(b<a)) - i.e. less than can be used to determine equality
		//3) If a<b and b<c, then a<c must be true.
	if(p1.distance(zero) < p2.distance(zero))
		return true;



	return false;
 }	

void addNormals(vector<Vector3f>& verts, vector<Vector3f>& vertNormals) {
	
		map<Vector3f, Vector3f, bool(*)(Vector3f, Vector3f)> myMap(vect3fCompare);

		Vector3f temp;

		//calculate the face normal, and set each vert to that face normal.
		for(int i=0; i< verts.size(); i+=3)
		{
			temp.calculateNormal_2(verts, verts[i+1], verts[i+2]);

			vertNormals.push_back(temp);
			vertNormals.push_back(temp);
			vertNormals.push_back(temp);
		}

		Vector3f zero(0,0,0);

                //fill it up with 0s.
		for(int i=0; i< verts.size(); i++)
			myMap[verts] = zero;


		//for every shared vertex, myMap will now collect all the proper face normals together to make individual vert normals.
		for(int i=0; i< verts.size(); i++)
		{
			myMap[verts] += vertNormals;
		}

                //now we loop back through, setting each vertNormal to the value calculated in the map.
		for(int i=0; i< verts.size(); i++)
			vertNormals = myMap[verts]; //addition of all verts same as verts
			
                //and finally normalise
		for(int i=0; i< vertNormals.size(); i++)
		{
			vertNormals.normalise();
		}
}



Share this post


Link to post
Share on other sites
Advertisement

This topic is 4064 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.

Guest
This topic is now closed to further replies.
Sign in to follow this  

  • Advertisement
×

Important Information

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

We are the game development community.

Whether you are an indie, hobbyist, AAA developer, or just trying to learn, GameDev.net is the place for you to learn, share, and connect with the games industry. Learn more About Us or sign up!

Sign me up!