Jump to content

  • Log In with Google      Sign In   
  • Create Account


Member Since 31 Aug 2012
Offline Last Active Oct 24 2012 06:34 AM

Posts I've Made

In Topic: Trace a ray through a 3D grid of cells?

11 October 2012 - 01:26 AM

You should really just look up the DDA algorithm as previously suggested, it is 'far' simpler implementation than what you have described.

My description was very poor. The algorithm I wrote is really simple, but it requires that your ray is inside the cell that you're finding the neighbor of.

Here's a simple version with a single value:

int sign = Math.Sign(direction)
float distance;

if(sign == 1)
distance = 1 - value;
else if(sign == -1)
distance = -value;
distance = infinity;

After you get the distance to the next face for each component (X, Y, and Z), you figure out which face you will get to first by dividing the distance by the direction.

Once you know which face you will encounter first, you know which cell is the next cell. The problem I'm having is that I rely on the position of the ray to trace through the cells.

Sorry that I'm being very vague, I have ADHD, and it's really hard for me to articulate what's going through my mind. All I see is a visualization of my algorithm, and I could give you the code, but it wouldn't make sense since there are a few things that you'll likely not know what's happening.

In Topic: Trace a ray through a 3D grid of cells?

11 October 2012 - 12:56 AM

I decided to just go ahead and write the hard, complicated code to find the next cell from the camera, and it works, but I'm having trouble tracing through the cells. My problem is a little simple. All I need to do is figure out how to properly get the position that the ray touched the next cell. I don't need the complicated math to do this, it's something I THOUGHT I would be able to do. The ray is positioned inside a cell, then I find the face that's closest, and I move into that cell, and move the ray along itself until it is against the edge of that cell. I figured that I could do this:

ray.Position = ray.Position + ray.Direction * distanceNeeded, but I now realize why that won't work.

if the ray is a simple ray with a position of (0.5,0.5,0.5) and a direction of (1.0,0.0,0.0), then the next position should be (1.0,0.5,0.5). The distance between the X component of the wall and the X component of the ray's position is 0.5f, it works just fine when you have those values, but when the direction has values for all three components, and it's normalized, you won't be able to get the same value by multiplication.

So, really, what my new question is: How can I find the position that a ray collides with a plane? I just want to move the ray along itself until it is inside the next cell, but that's not working for some reason.

In Topic: Trace a ray through a 3D grid of cells?

10 October 2012 - 01:59 AM

For simplicity, since my voxels are 1x1x1 in size, I decided I could just move the ray by 1 unit, then get the cell that the ray is in, then check collision within that cell. Then I just keep incrementing it until I either reach my max length, or I find collision. Now I just need to remember how to collide a 3D ray with a 3D quad.