I was trying to implement the unified height fog described at https://www.terathon.com/lengyel/Lengyel-UnifiedFog.pdf. This particular fog had already been implemented in a old unity asset as an ImageEffect and I wanted to port it to the PostProcessing stack to use volume blending, etc…
The formula I’m using is the exact same, still the fog behaves in a weird way :
The height fog volume doesn’t cover the whole ground as it should. Varying the fog height makes the fog “horizontally move” further away or closer to the camera. I’ve read the whole paper and the implementation makes sense to me. I played around with the dot products and such but in no way does the height fog behave as expected. Help appreciated : )
Shader code
Shader "Hidden/Evan/Global Fog Post Process"
{
Subshader
{
ZTest Always Cull Off ZWrite Off Fog{ Mode Off }
// the bible (´˘ -˘ 人)
// https://docs.unity3d.com/6000.0/Documentation/Manual/SL-UnityShaderVariables.html
// distance and height fog pass
Pass
{
HLSLPROGRAM
#pragma vertex VertDefault
#pragma fragment Frag
#include "Packages/com.unity.postprocessing/PostProcessing/Shaders/StdLib.hlsl"
TEXTURE2D_SAMPLER2D(_MainTex, sampler_MainTex);
TEXTURE2D_SAMPLER2D(_CameraDepthTexture, sampler_CameraDepthTexture);
float4 unity_FogColor;
float4 unity_FogParams;
float4x4 unity_CameraInvProjection;
float4x4 unity_CameraToWorld;
/* HEIGHT FOG VARIABLES SET FROM THE POSTPROCESSGLOBALFOGRENDERER */
float3 fogPlaneNormal;
float fogPlaneHeight;
float heightFogDensity;
float fogNormalDotCamera;
float k;
/* PARAMETERS SET FROM THE POSTPROCESSGLOBALFOGRENDERER */
float4 globalFogTint;
float energyLoss;
float globalFogMaxDensity;
float4 Frag(VaryingsDefault i) : SV_Target
{
float4 color = SAMPLE_TEXTURE2D(_MainTex, sampler_MainTex, i.texcoord);
float depth = SAMPLE_DEPTH_TEXTURE(_CameraDepthTexture, sampler_CameraDepthTexture, i.texcoord);
float linearDepth = Linear01Depth(depth);
// direction from camera towards this frag on screen
/*
/- - -
C -> | - - -
\- - -
*/
float distanceFromCamera = (linearDepth * _ProjectionParams.z /* (camera's far plane) */) - _ProjectionParams.y /* (camera's near plane) */;
/* DISTANCE FOG */
float density = unity_FogParams.y; // in UnityShaderVariables.cginc : y = density / ln(2)
half fogFactor = density * (distanceFromCamera);
fogFactor = exp2(-fogFactor);
/* HEIGHT FOG */
// calculate world position
float2 ndc = i.texcoord * 2 - 1; // normalized device coordinates ([0 -> 1] to [-1 -> 1])
float3 fragViewRay = mul(unity_CameraInvProjection, float4(ndc, 1.0, 1.0) * _ProjectionParams.z); // ray passing through UVs and far plane
float3 fragViewPosition = fragViewRay * linearDepth; // multiplied by depth to retrieve the view-space position
float3 fragWorldPosition = mul(unity_CameraToWorld, float4(fragViewPosition * float3(1.0,1.0,-1.0), 1.0)).xyz; // transform by the cameraToWorld matrix
// good. now we can calculate half-space fog thanks to https://www.terathon.com/lengyel/Lengyel-UnifiedFog.pdf
float3 cameraWorldPosition = _WorldSpaceCameraPos;
float fogNormalDotFragWorld = fragWorldPosition.y - fogPlaneHeight; // evaluated foreach frag
float fogNormalDotFragV = dot(fogPlaneNormal, cameraWorldPosition - fogNormalDotFragWorld);
float heightFogFactor = -(heightFogDensity * distanceFromCamera) * (k * (fogNormalDotFragWorld + fogNormalDotCamera) - ( pow(min( (1 - 2 * k) * fogNormalDotFragWorld , 0), 2) / ( abs(fogNormalDotFragV + 0.00001) ) ));
fogFactor -= heightFogFactor;
fogFactor = saturate(fogFactor);
//return heightFogFactor;
half4 colorDark = color * pow(fogFactor, clamp(energyLoss, 0.001, 100));
return lerp(unity_FogColor * half4(globalFogTint.rgb, 1), lerp(color, colorDark, globalFogMaxDensity), clamp(fogFactor, 1 - globalFogMaxDensity, 1));
//return lerp(unity_FogColor, color, fogFactor);
}
ENDHLSL
}
}
}
This is my first time writing shaders so sorry for the bad practices. Also the source is available at GitHub - evan-ciao/unityssmsfog: Screen-space multi scatter and global fog implementation for Unity's Post-processing Stack v2.. Thanks again.