CSG faking with shaders: nearly there?

I've been trolling forum posts trying to find a current example of how to fake csg with shaders, my ultimate goal being that I would like to render nothing except for the location that an object is overlapped. I believe that I'm close to achieving this effect, although there seems to be something I'm missing.

My goal at this point is to have the small circle inside the sphere render, and nothing else.


At this point, the cube renders everywhere except for where the sphere is ahead of it on zdepth. Removing the cube from render, only rendering the circle, and not rendering the sphere is what I'd like to achieve


The cube is using the "Regular" shader.

The sphere that masks is using the "Special" shader

Cube;

Code (CSharp):

1. Shader "Custom/Regular" {
2.     Properties {
3.         _Color ("Color", Color) = (1,1,1,1)
4.         _MainTex ("Albedo (RGB)", 2D) = "white" {}
5.         _Glossiness ("Smoothness", Range(0,1)) = 0.5
6.         _Metallic ("Metallic", Range(0,1)) = 0.0
7.     }
8.     SubShader {
9.         Tags { "RenderType"="Opaque" "Queue"="Geometry+2"}
10.         LOD 200
11.        
12.         CGPROGRAM
13.         #pragma surface surf Lambert
14.         float4 _Color;
15.         struct Input {
16.             float4 color : COLOR;
17.         };
18.         void surf (Input IN, inout SurfaceOutput o) {
19.             o.Albedo = _Color.rgb;
20.             o.Normal = half3(0,0,-1);
21.             o.Alpha = 1;
22.         }
23.         ENDCG
24.     }
25.     FallBack "Diffuse"
26. }
27.  

Sphere "Mask";

Code (CSharp):

1. Shader "Custom/Special" {
2.     Properties {
3.         _Color ("Color", Color) = (1,1,1,1)
4.         _MainTex ("Albedo (RGB)", 2D) = "white" {}
5.         _Glossiness ("Smoothness", Range(0,1)) = 0.5
6.         _Metallic ("Metallic", Range(0,1)) = 0.0
7.     }
8.     SubShader {
9.         Tags { "RenderType"="Opaque" "Queue"="Geometry+1"}
10.         cull back
11.         Colormask 0
12.        
13.         CGINCLUDE
14.             struct appdata {
15.                 float4 vertex : POSITION;
16.             };
17.             struct v2f {
18.                 float4 pos : SV_POSITION;
19.             };
20.             v2f vert(appdata v) {
21.                 v2f o;
22.                 o.pos = mul(UNITY_MATRIX_MVP, v.vertex);
23.                 return o;
24.             }
25.             half4 frag(v2f i) : SV_Target {
26.                 return half4(1,1,0,1);
27.             }
28.         ENDCG
29.         Pass
30.         {
31.             Tags {"Queue"="Geometry+3" }
32.             Cull Front
33.             ColorMask 0
34.            
35.             CGPROGRAM
36.             #pragma vertex vert
37.             #pragma fragment frag
38.             ENDCG
39.        
40.         }
41.         Pass
42.         {
43.             Tags {"Queue"="Geometry+4" }
44.             cull back
45.             ztest Equal
46.             ColorMask rgba
47.            
48.             CGPROGRAM
49.             #pragma vertex vert
50.             #pragma fragment frag
51.             ENDCG
52.        
53.         }
54.     }
55.     FallBack "Diffuse"
56. }
57.  

Any help would be appreciated, I'm at a bit of a standstill atm

Thanks everyone,
-Gudge

 

Bump

 

I'm attempting to get the effect demonstrated in this forum post: forum . The section I'm trying to get is the red overlap area.

 

Try with ZTest NotEqual

You might find real accuracy issues doing this though, as the depth buffer is only approximate depending on your target device.

 

Unfortunately it didn't help, changing ztest equal to notequal made the sphere render yellow D:

 

I would have thought you'd be rendering the sphere with depth only?

 

Although...that won't really help from all views as you won't get depth for the backface!

 

Thinking about this, I'm not 100% sure you could do this without using rendertextures, so you'd go something like this:

1) Render Sphere Backface depth to RT1
2) Render Sphere Frontface depth to RT2
3) Render Cube and use Clip command if pixel depth equals RT1 or RT2.

I think thats the old way, then you can use a threshold to change the width of the cut too.

 

With ztest equal and colormask 0 on special, i get close to the effect that I want but the issue would be with regular i think. the rest of the cube needs to not be displayed.

I'm attempting to use this to replace roughly 2000 area lights, so any practical use of this would be awesome. That said, there will be roughly 2000 objects with the regular shader interacting with walls in my "arena".


If I'm understanding your render texture idea, it would only require two textures, one for the sphere (no matter how many there are), and one for the walls using the main camera? If that's the case, then it will almost certainly be an improvement.

 

Well, two both would be for the sphere no matter how many, then the cubes no matter how many would use both of the textures, you could probably just use 1 and use the depth texture instead of the second one (2 above) and probably depth test it, but I'm not 100% sure what you're going to be doing with this, so tricky to say...also I'm in the middle of something complex myself thats throwing my thoughts! :S

 

I'll fiddle around with rendertextures to see how that goes and i'll report back once that's done

 

The one great thing with the RT way, easier to debug, you can just output the RT and see what is going on!

 

Just to be clear so I know exactly what to try, are you suggesting the Unity Render texture which uses separate cameras, or render to a texture with a shader.

 

Quickest way is use a new camera as a child of you existing one and put it rendering to a rendertexture, then set that as a global shader texture for your other objects to pick up on.

 

From what I am getting at, you're trying to do an overlay of some sort, of which such overlay comes from the cut-out bits of a mesh, correct?