VolumetricLighting/Assets/Scenes/Materials/StandardAlphaBlended-VolumetricFog.shader
2016-11-06 00:21:11 +01:00

102 lines
3.4 KiB
Plaintext

Shader "Custom/StandardAlphaBlended-VolumetricFog"
{
Properties
{
_Color ("Color", Color) = (1,1,1,1)
_MainTex ("Albedo (RGB)", 2D) = "white" {}
_Glossiness ("Smoothness", Range(0,1)) = 0.5
_Metallic ("Metallic", Range(0,1)) = 0.0
}
SubShader
{
Tags {"Queue" = "Transparent" "RenderType"="Transparent" }
LOD 200
CGPROGRAM
#pragma surface surf Standard fullforwardshadows alpha finalcolor:ApplyFog
#pragma target 3.0
#pragma multi_compile _ VOLUMETRIC_FOG
#if VOLUMETRIC_FOG
#include "../../VolumetricFog/Shaders/VolumetricFog.cginc"
#endif
sampler2D _MainTex;
sampler2D _CameraDepthTexture;
struct Input {
float2 uv_MainTex;
float4 screenPos;
};
void ApplyFog(Input IN, SurfaceOutputStandard o, inout fixed4 color)
{
#if VOLUMETRIC_FOG
half3 uvscreen = IN.screenPos.xyz/IN.screenPos.w;
half linear01Depth = Linear01Depth(uvscreen.z);
fixed4 fog = Fog(linear01Depth, uvscreen.xy);
// Always apply fog attenuation - also in the forward add pass.
color.rgb *= fog.a;
// Alpha premultiply mode (used with alpha and Standard lighting function, or explicitly alpha:premul)
// uses source blend factor of One instead of SrcAlpha. `color` is compensated for it, so we need to compensate
// the amount of inscattering too. A note on why this works: below.
#if _ALPHAPREMULTIPLY_ON
fog.rgb *= o.Alpha;
#endif
// Add inscattering only once, so in forward base, but not forward add.
#ifndef UNITY_PASS_FORWARDADD
color.rgb += fog.rgb;
#endif
// So why does multiplying the inscattered light by alpha work?
// In other words: how did fog ever work, if opaque objects add all of the inscattered light
// between them and the camera, and then the transparencies add even more?
//
// This is our scene initially:
// scene |---is0---------------------------------------> camera
//
// And that's with the transparent object added in between the opaque stuff and the camera:
// scene |---is1---> transparent |---is2---------------> camera
//
// When rendering, we start with the opaque part of the scene and add all the light inscattered between that and the camera: is0.
// Then we add the transparent object. It does two things (let's consider the alpha premultiply version):
// - Dims whatever was behind it (including is0) by OneMinusSrcAlpha
// - Adds light inscattered in front of it (is2), multiplied by Alpha
//
// So all in all we end up with this much inscattered light:
// is0 * OneMinusSrcAlpha + is2 * Alpha
//
// Judging by the diagram, though, the correct amount should be:
// is1 * OneMinusSrcAlpha + is2
//
// Turns out the two expressions are equal - who would've thunk?
// is1 = is0 - is2
// (is0 - is2) * OneMinusSrcAlpha + is2
// is0 * OneMinusSrcAlpha - is2 * (1 - Alpha) + is2
// is0 * OneMinusSrcAlpha - is2 + is2 * Alpha + is2
// is0 * OneMinusSrcAlpha + is2 * Alpha
// I leave figuring out if the fog attenuation is correct as an exercise to the reader ;)
#endif
}
half _Glossiness;
half _Metallic;
fixed4 _Color;
void surf (Input IN, inout SurfaceOutputStandard o)
{
fixed4 c = tex2D (_MainTex, IN.uv_MainTex) * _Color;
o.Albedo = c.rgb;
o.Metallic = _Metallic;
o.Smoothness = _Glossiness;
o.Alpha = c.a;
}
ENDCG
}
FallBack "Standard"
}