Shadowmapping...

[serial]

Was thinking about this today.  Havn't had time to test the idea out. but was just curious if this would work for shadowmapping.

First you get the distance from the lightsource to the point on the landscape.  Then you cast a ray from the light source to the same point.  You then compare the length of the ray compared to the distance.  If the length of the ray is shorter then that means that part of the landscape is shadowed. 

Does that sound about right?  I spent just a little bit of time on this today but wanted to have some idea of what I'm doing before I get there.

[Zaknafein]

Hahah my friend, you've just invented shadowmapping.
That's exactly what shadowmapping is about, and how most (or all) of the implementations work. That's called depth comparison.

[zwiglm]

First you get the distance from the lightsource to the point on the landscape.  Then you cast a ray from the light source to the same point.  You then compare the length of the ray compared to the distance.

Forgive me for my bad english, but I don't get. Isn't that the same in your explanation?
1. distance form light to landscape
2. ray from light to landscape.

Where/what's the difference?

[serial]

Okay imagine it this way.  Lets say there is a brick wall about 200 meters away.  In front of that brick wall is a chair.  If you shine a laser pointer at every point on the brick wall the length of the laser would be 200 meters except where the chair is.
So for everypoint where distance <> laser length that point on the wall would be shadowed.

Getting the distance does not take in to count collisions.  You are check for collisions with the ray that you are cast.  If you get a collision and the distance returned is shorter you have a shadow. 

Another way you could do is check to see if the normal of a surface is pointing towards a light if it is then the surface is probably receiving light.

[WEst]

Getting the distance does not take in to count collisions.  You are check for collisions with the ray that you are cast.  If you get a collision and the distance returned is shorter you have a shadow. 

But normally you use not raycasting, but render the scene from the lights point of view using a depth shader, that returns the depth of the scene seen from the lights position. And compare this depth- or shadowmap to the normal scene rendering in a shader.

[Shadowsong]

Is there any tutorial out there that explains FROM SCRATCH how to create simple directional shadowmapping? Where from scratch means I am able to make a shader but I have no idea about shadowmapping, depht mapping etc...

[WEst]

I would recommend this article it's for DX9/XNA but you can easily convert it to TV3D:

http://www.riemers.net/eng/Tutorials/XNA/Csharp/Series3/Shadow_map.php

BUT, I also would advise that you look a bit into shadowmapping and the methods behind it (it's explained at the beginning of the article), because you will get very weird and strange problems in the beginning, and it's difficult to find them if you don't know about how it works.

[Shadowsong]

Thanks I'll read it carefully.