Sphere generation problem

Graphics and GPU Programming Programming

Started by LordFallout March 20, 2008 09:19 AM

1 comment, last by LordFallout 16 years, 1 month ago

132

Author

March 20, 2008 09:19 AM

I'm attempting to generate Sphere Vertex data based on the sphere games Sky Domes pdf. found at http://www.spheregames.com/files/SkyDomesPDF.zip if i implement the code in the pdf for generating a dome it works, however the dome is on its side. This is easy to fix. The second problem is when i try to extend the Dome into a full Sphere You can see the problem here

Basically instead of extending the sphere it goes back over the Dome. I use the following method to determine the vertices for the dome


	m_pVerts = new sVertexXYZCUV[m_iNumberOfVertices];

	int n = 0;
	for(int iPhi = -90; iPhi < 90; iPhi += m_idPhi)
	{
		for(int iTheta = 0; iTheta < 360; iTheta += m_idTheta)
		{
			m_pVerts[n].fX = m_fRadius * sinf(iPhi * DTOR) * cosf(DTOR * iTheta);
			m_pVerts[n].fY = m_fRadius * sinf(iPhi * DTOR) * sinf(DTOR * iTheta);
			m_pVerts[n].fZ = m_fRadius * cosf(iPhi * DTOR);
			n++;

			m_pVerts[n].fX = m_fRadius * sinf((iPhi + m_idPhi) * DTOR) * cosf(iTheta * DTOR);
			m_pVerts[n].fY = m_fRadius * sinf((iPhi + m_idPhi) * DTOR) * sinf(iTheta * DTOR);
			m_pVerts[n].fZ = m_fRadius * cosf((iPhi + m_idPhi) * DTOR);
			n++;
			
			m_pVerts[n].fX = m_fRadius * sinf(DTOR * iPhi) * cosf(DTOR * (iTheta + m_idTheta));
			m_pVerts[n].fY = m_fRadius * sinf(DTOR * iPhi) * sinf(DTOR * (iTheta + m_idTheta));
			m_pVerts[n].fZ = m_fRadius * cosf(DTOR * iPhi);
			n++;
			
			if(iPhi > -90 && iPhi < 90)
			{
				m_pVerts[n].fX = m_fRadius * sinf((iPhi + m_idPhi) * DTOR) * cosf(DTOR * (iTheta + m_idTheta));
				m_pVerts[n].fY = m_fRadius * sinf((iPhi + m_idPhi) * DTOR) * sinf(DTOR * (iTheta + m_idTheta));
				m_pVerts[n].fZ = m_fRadius * cosf((iPhi + m_idPhi) * DTOR);
				n++;
			}
		}
	}

DTOR is defined as D3DX_PI / 180.0f This is me attempting to implement the latitude / longitude method :S [Edited by - LordFallout on March 20, 2008 9:42:11 AM]

LordFallout

132

Author

March 20, 2008 10:38 AM

I managed to sort the orientation problem by switching the Y and Z co-ordinate calculations.

Also, i managed to make a whole sphere, there still exists a problem thought

I'm not sure if its clear or not from this photo but the vertices are mixed up towards the poles.

This was the semi solution

	m_pVerts = new sVertexXYZCUV[m_iNumberOfVertices];	int n = 0;	for(float fPhi = -90; fPhi < 90; fPhi += m_fdPhi)	{		for(float fTheta = 0; fTheta < 360; fTheta += m_fdTheta)		{			m_pVerts[n].fX = m_fRadius * sinf(fPhi * DTOR) * cosf(DTOR * fTheta);			m_pVerts[n].fY = m_fRadius * cosf(fPhi * DTOR);			if(fPhi < 0)				m_pVerts[n].fY = -m_pVerts[n].fY;			m_pVerts[n].fZ = m_fRadius * sinf(fPhi * DTOR) * sinf(DTOR * fTheta);			n++;			m_pVerts[n].fX = m_fRadius * sinf((fPhi + m_fdPhi) * DTOR) * cosf(fTheta * DTOR);			m_pVerts[n].fY = m_fRadius * cosf((fPhi + m_fdPhi) * DTOR);			if(fPhi < 0)				m_pVerts[n].fY = -m_pVerts[n].fY;			m_pVerts[n].fZ = m_fRadius * sinf((fPhi + m_fdPhi) * DTOR) * sinf(fTheta * DTOR);			n++;						m_pVerts[n].fX = m_fRadius * sinf(fPhi * DTOR) * cosf(DTOR * (fTheta + m_fdTheta));			m_pVerts[n].fY = m_fRadius * cosf(fPhi * DTOR);			if(fPhi < 0)				m_pVerts[n].fY = -m_pVerts[n].fY;			m_pVerts[n].fZ = m_fRadius * sinf(fPhi * DTOR) * sinf(DTOR * (fTheta + m_fdTheta));			n++;						m_pVerts[n].fX = m_fRadius * sinf((fPhi + m_fdPhi) * DTOR) * cosf(DTOR * (fTheta + m_fdTheta));			m_pVerts[n].fY = m_fRadius * cosf((fPhi + m_fdPhi) * DTOR);			if(fPhi < 0)				m_pVerts[n].fY = -m_pVerts[n].fY;			m_pVerts[n].fZ = m_fRadius * sinf((fPhi + m_fdPhi) * DTOR) * sinf(DTOR * (fTheta + m_fdTheta));			n++;		}	}

The initial problem arose because the cos of a number would be the same no matter if it was positive or negative, hence the solution to check if the number is negative or not, i think there might have been a better solution, if anyone notices it, please post for me. I'm going to continue my search.

LordFallout

132

Author

March 20, 2008 10:49 AM

wow, i've spent the past five hours doing nothing, :) the fix was to

	for(float fPhi = 0; fPhi < 180; fPhi += m_fdPhi)	{

instead of

	for(float fPhi = -90; fPhi < 90; fPhi += m_fdPhi)	{

Maybe that'll help someone else out in the future :)

Now onto the colouring.

Sphere generation problem

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Sphere generation problem

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Reticulating splines