【Unity】Standard Shader实现分析

记录Unity的标准着色器实现，基于Unity 2017.1版本的代码进行分析。

Standard Shader

文件位于DefaultResourcesExtraStandard.shader

// Unity built-in shader source. Copyright (c) 2016 Unity Technologies. MIT license (see license.txt)

Shader "Standard"
{
    Properties
    {
        _Color("Color", Color) = (1,1,1,1)
        _MainTex("Albedo", 2D) = "white" {}

        _Cutoff("Alpha Cutoff", Range(0.0, 1.0)) = 0.5

        _Glossiness("Smoothness", Range(0.0, 1.0)) = 0.5
        _GlossMapScale("Smoothness Scale", Range(0.0, 1.0)) = 1.0
        [Enum(Metallic Alpha,0,Albedo Alpha,1)] _SmoothnessTextureChannel ("Smoothness texture channel", Float) = 0

        [Gamma] _Metallic("Metallic", Range(0.0, 1.0)) = 0.0
        _MetallicGlossMap("Metallic", 2D) = "white" {}

        [ToggleOff] _SpecularHighlights("Specular Highlights", Float) = 1.0
        [ToggleOff] _GlossyReflections("Glossy Reflections", Float) = 1.0

        _BumpScale("Scale", Float) = 1.0
        _BumpMap("Normal Map", 2D) = "bump" {}

        _Parallax ("Height Scale", Range (0.005, 0.08)) = 0.02
        _ParallaxMap ("Height Map", 2D) = "black" {}

        _OcclusionStrength("Strength", Range(0.0, 1.0)) = 1.0
        _OcclusionMap("Occlusion", 2D) = "white" {}

        _EmissionColor("Color", Color) = (0,0,0)
        _EmissionMap("Emission", 2D) = "white" {}

        _DetailMask("Detail Mask", 2D) = "white" {}

        _DetailAlbedoMap("Detail Albedo x2", 2D) = "grey" {}
        _DetailNormalMapScale("Scale", Float) = 1.0
        _DetailNormalMap("Normal Map", 2D) = "bump" {}

        [Enum(UV0,0,UV1,1)] _UVSec ("UV Set for secondary textures", Float) = 0


        // Blending state
        [HideInInspector] _Mode ("__mode", Float) = 0.0
        [HideInInspector] _SrcBlend ("__src", Float) = 1.0
        [HideInInspector] _DstBlend ("__dst", Float) = 0.0
        [HideInInspector] _ZWrite ("__zw", Float) = 1.0
    }

    CGINCLUDE
        #define UNITY_SETUP_BRDF_INPUT MetallicSetup
    ENDCG

    SubShader
    {
        Tags { "RenderType"="Opaque" "PerformanceChecks"="False" }
        LOD 300


        // ------------------------------------------------------------------
        //  Base forward pass (directional light, emission, lightmaps, ...)
        Pass
        {
            Name "FORWARD"
            Tags { "LightMode" = "ForwardBase" }

            Blend [_SrcBlend] [_DstBlend]
            ZWrite [_ZWrite]

            CGPROGRAM
            #pragma target 3.0

            // -------------------------------------

            #pragma shader_feature _NORMALMAP
            #pragma shader_feature _ _ALPHATEST_ON _ALPHABLEND_ON _ALPHAPREMULTIPLY_ON
            #pragma shader_feature _EMISSION
            #pragma shader_feature _METALLICGLOSSMAP
            #pragma shader_feature ___ _DETAIL_MULX2
            #pragma shader_feature _ _SMOOTHNESS_TEXTURE_ALBEDO_CHANNEL_A
            #pragma shader_feature _ _SPECULARHIGHLIGHTS_OFF
            #pragma shader_feature _ _GLOSSYREFLECTIONS_OFF
            #pragma shader_feature _PARALLAXMAP

            #pragma multi_compile_fwdbase
            #pragma multi_compile_fog
            #pragma multi_compile_instancing
            // Uncomment the following line to enable dithering LOD crossfade. Note: there are more in the file to uncomment for other passes.
            //#pragma multi_compile _ LOD_FADE_CROSSFADE

            #pragma vertex vertBase
            #pragma fragment fragBase
            #include "UnityStandardCoreForward.cginc"

            ENDCG
        }
        // ------------------------------------------------------------------
        //  Additive forward pass (one light per pass)
        Pass
        {
            Name "FORWARD_DELTA"
            Tags { "LightMode" = "ForwardAdd" }
            Blend [_SrcBlend] One
            Fog { Color (0,0,0,0) } // in additive pass fog should be black
            ZWrite Off
            ZTest LEqual

            CGPROGRAM
            #pragma target 3.0

            // -------------------------------------


            #pragma shader_feature _NORMALMAP
            #pragma shader_feature _ _ALPHATEST_ON _ALPHABLEND_ON _ALPHAPREMULTIPLY_ON
            #pragma shader_feature _METALLICGLOSSMAP
            #pragma shader_feature _ _SMOOTHNESS_TEXTURE_ALBEDO_CHANNEL_A
            #pragma shader_feature _ _SPECULARHIGHLIGHTS_OFF
            #pragma shader_feature ___ _DETAIL_MULX2
            #pragma shader_feature _PARALLAXMAP

            #pragma multi_compile_fwdadd_fullshadows
            #pragma multi_compile_fog
            // Uncomment the following line to enable dithering LOD crossfade. Note: there are more in the file to uncomment for other passes.
            //#pragma multi_compile _ LOD_FADE_CROSSFADE

            #pragma vertex vertAdd
            #pragma fragment fragAdd
            #include "UnityStandardCoreForward.cginc"

            ENDCG
        }
        // ------------------------------------------------------------------
        //  Shadow rendering pass
        Pass {
            Name "ShadowCaster"
            Tags { "LightMode" = "ShadowCaster" }

            ZWrite On ZTest LEqual

            CGPROGRAM
            #pragma target 3.0

            // -------------------------------------


            #pragma shader_feature _ _ALPHATEST_ON _ALPHABLEND_ON _ALPHAPREMULTIPLY_ON
            #pragma shader_feature _METALLICGLOSSMAP
            #pragma shader_feature _PARALLAXMAP
            #pragma multi_compile_shadowcaster
            #pragma multi_compile_instancing
            // Uncomment the following line to enable dithering LOD crossfade. Note: there are more in the file to uncomment for other passes.
            //#pragma multi_compile _ LOD_FADE_CROSSFADE

            #pragma vertex vertShadowCaster
            #pragma fragment fragShadowCaster

            #include "UnityStandardShadow.cginc"

            ENDCG
        }
        // ------------------------------------------------------------------
        //  Deferred pass
        Pass
        {
            Name "DEFERRED"
            Tags { "LightMode" = "Deferred" }

            CGPROGRAM
            #pragma target 3.0
            #pragma exclude_renderers nomrt


            // -------------------------------------

            #pragma shader_feature _NORMALMAP
            #pragma shader_feature _ _ALPHATEST_ON _ALPHABLEND_ON _ALPHAPREMULTIPLY_ON
            #pragma shader_feature _EMISSION
            #pragma shader_feature _METALLICGLOSSMAP
            #pragma shader_feature _ _SMOOTHNESS_TEXTURE_ALBEDO_CHANNEL_A
            #pragma shader_feature _ _SPECULARHIGHLIGHTS_OFF
            #pragma shader_feature ___ _DETAIL_MULX2
            #pragma shader_feature _PARALLAXMAP

            #pragma multi_compile_prepassfinal
            #pragma multi_compile_instancing
            // Uncomment the following line to enable dithering LOD crossfade. Note: there are more in the file to uncomment for other passes.
            //#pragma multi_compile _ LOD_FADE_CROSSFADE

            #pragma vertex vertDeferred
            #pragma fragment fragDeferred

            #include "UnityStandardCore.cginc"

            ENDCG
        }

        // ------------------------------------------------------------------
        // Extracts information for lightmapping, GI (emission, albedo, ...)
        // This pass it not used during regular rendering.
        Pass
        {
            Name "META"
            Tags { "LightMode"="Meta" }

            Cull Off

            CGPROGRAM
            #pragma vertex vert_meta
            #pragma fragment frag_meta

            #pragma shader_feature _EMISSION
            #pragma shader_feature _METALLICGLOSSMAP
            #pragma shader_feature _ _SMOOTHNESS_TEXTURE_ALBEDO_CHANNEL_A
            #pragma shader_feature ___ _DETAIL_MULX2
            #pragma shader_feature EDITOR_VISUALIZATION

            #include "UnityStandardMeta.cginc"
            ENDCG
        }
    }

    SubShader
    {
        Tags { "RenderType"="Opaque" "PerformanceChecks"="False" }
        LOD 150

        // ------------------------------------------------------------------
        //  Base forward pass (directional light, emission, lightmaps, ...)
        Pass
        {
            Name "FORWARD"
            Tags { "LightMode" = "ForwardBase" }

            Blend [_SrcBlend] [_DstBlend]
            ZWrite [_ZWrite]

            CGPROGRAM
            #pragma target 2.0

            #pragma shader_feature _NORMALMAP
            #pragma shader_feature _ _ALPHATEST_ON _ALPHABLEND_ON _ALPHAPREMULTIPLY_ON
            #pragma shader_feature _EMISSION
            #pragma shader_feature _METALLICGLOSSMAP
            #pragma shader_feature _ _SMOOTHNESS_TEXTURE_ALBEDO_CHANNEL_A
            #pragma shader_feature _ _SPECULARHIGHLIGHTS_OFF
            #pragma shader_feature _ _GLOSSYREFLECTIONS_OFF
            // SM2.0: NOT SUPPORTED shader_feature ___ _DETAIL_MULX2
            // SM2.0: NOT SUPPORTED shader_feature _PARALLAXMAP

            #pragma skip_variants SHADOWS_SOFT DIRLIGHTMAP_COMBINED

            #pragma multi_compile_fwdbase
            #pragma multi_compile_fog

            #pragma vertex vertBase
            #pragma fragment fragBase
            #include "UnityStandardCoreForward.cginc"

            ENDCG
        }
        // ------------------------------------------------------------------
        //  Additive forward pass (one light per pass)
        Pass
        {
            Name "FORWARD_DELTA"
            Tags { "LightMode" = "ForwardAdd" }
            Blend [_SrcBlend] One
            Fog { Color (0,0,0,0) } // in additive pass fog should be black
            ZWrite Off
            ZTest LEqual

            CGPROGRAM
            #pragma target 2.0

            #pragma shader_feature _NORMALMAP
            #pragma shader_feature _ _ALPHATEST_ON _ALPHABLEND_ON _ALPHAPREMULTIPLY_ON
            #pragma shader_feature _METALLICGLOSSMAP
            #pragma shader_feature _ _SMOOTHNESS_TEXTURE_ALBEDO_CHANNEL_A
            #pragma shader_feature _ _SPECULARHIGHLIGHTS_OFF
            #pragma shader_feature ___ _DETAIL_MULX2
            // SM2.0: NOT SUPPORTED shader_feature _PARALLAXMAP
            #pragma skip_variants SHADOWS_SOFT

            #pragma multi_compile_fwdadd_fullshadows
            #pragma multi_compile_fog

            #pragma vertex vertAdd
            #pragma fragment fragAdd
            #include "UnityStandardCoreForward.cginc"

            ENDCG
        }
        // ------------------------------------------------------------------
        //  Shadow rendering pass
        Pass {
            Name "ShadowCaster"
            Tags { "LightMode" = "ShadowCaster" }

            ZWrite On ZTest LEqual

            CGPROGRAM
            #pragma target 2.0

            #pragma shader_feature _ _ALPHATEST_ON _ALPHABLEND_ON _ALPHAPREMULTIPLY_ON
            #pragma shader_feature _METALLICGLOSSMAP
            #pragma skip_variants SHADOWS_SOFT
            #pragma multi_compile_shadowcaster

            #pragma vertex vertShadowCaster
            #pragma fragment fragShadowCaster

            #include "UnityStandardShadow.cginc"

            ENDCG
        }

        // ------------------------------------------------------------------
        // Extracts information for lightmapping, GI (emission, albedo, ...)
        // This pass it not used during regular rendering.
        Pass
        {
            Name "META"
            Tags { "LightMode"="Meta" }

            Cull Off

            CGPROGRAM
            #pragma vertex vert_meta
            #pragma fragment frag_meta

            #pragma shader_feature _EMISSION
            #pragma shader_feature _METALLICGLOSSMAP
            #pragma shader_feature _ _SMOOTHNESS_TEXTURE_ALBEDO_CHANNEL_A
            #pragma shader_feature ___ _DETAIL_MULX2
            #pragma shader_feature EDITOR_VISUALIZATION

            #include "UnityStandardMeta.cginc"
            ENDCG
        }
    }


    FallBack "VertexLit"
    CustomEditor "StandardShaderGUI"
}

Standard Shader中主要有三个分支，一个是SM3.0的Forward渲染实现，一个是Deferred渲染实现，一个是针对SM2.0的Forward渲染实现。

在SM3.0下，Unity实现Forward渲染有两个Pass。第一个是Pass是针对主光源的ForwardBase，第二个Pass是针对其他光源的ForwardAdd。实现两个Pass的顶点着色器和片段着色器函数名称也已经给出，包含在"UnityStandardCoreForward.cginc"文件中。

在UnityStandardCoreForward.cginc文件中出现了分支，一个是Simple实现一个是标准的实现。从学习的目的来讲，主要看Unity的标准实现。

根据上述代码，我们在UnityStandardCore.cginc中找到了顶点着色器和片段着色器的具体实现。为了减轻工作量，先研究Forward渲染的代码。顶点着色器为vertForwardBase/vertForwardAdd，片段着色器为fragForwardBase/fragForwardAdd。ForwardAdd实现和ForwardBase实现类似，只有少量区别。所以主要分析ForwardBase，ForwardAdd会在之后简单介绍与ForwardBase的差异。

vertForwardBase函数

作为一个顶点着色器，vertForwardBase的实现很常规，主要是一系列相关的坐标变换工作。在分析vertForwardBase函数之前，需要先分析一下顶点着色器输出到片段着色器的结构体VertexOutputForwardBase。这部分内容不重要，只会简单的说明一下。

struct VertexOutputForwardBase
{
    UNITY_POSITION(pos);
    float4 tex                          : TEXCOORD0;
    half3 eyeVec                        : TEXCOORD1;
    half4 tangentToWorldAndPackedData[3]    : TEXCOORD2;    // [3x3:tangentToWorld | 1x3:viewDirForParallax or worldPos]
    half4 ambientOrLightmapUV           : TEXCOORD5;    // SH or Lightmap UV
    UNITY_SHADOW_COORDS(6)
    UNITY_FOG_COORDS(7)

    // next ones would not fit into SM2.0 limits, but they are always for SM3.0+
    #if UNITY_REQUIRE_FRAG_WORLDPOS && !UNITY_PACK_WORLDPOS_WITH_TANGENT
        float3 posWorld                 : TEXCOORD8;
    #endif

    UNITY_VERTEX_INPUT_INSTANCE_ID
    UNITY_VERTEX_OUTPUT_STEREO
};

总体来说做了以下的工作：

1.定义变量：顶点坐标，纹理坐标，视线向量。UNITY_POSITION(pos)宏定义位于HLSLSupport.cginc，不赘述。

2.tangentToWorldAndPackedData[3]：大小为3x4，其中3x3矩阵为切线空间变换到世界空间矩阵（xyz分量），1x3为视差视线向量或世界坐标（w分量）。

3.定义变量：环境或光照贴图的坐标，阴影坐标，雾坐标。

4.针对SM3.0，定义变量：顶点世界坐标。

5.UNITY_VERTEX_INPUT_INSTANCE_ID为顶点实例化一个ID，UNITY_VERTEX_OUTPUT_STEREO来声明该顶点是否位于视线域中，来判断这个顶点是否输出到片段着色器。两个宏定义位于UnityInstancing.cginc中，GPU Instancing所需，这里不赘述。

顶点函数vertForwardBase用于填充VertexOutputForwardBase结构体。

VertexOutputForwardBase vertForwardBase (VertexInput v)
{
    UNITY_SETUP_INSTANCE_ID(v);
    VertexOutputForwardBase o;
    UNITY_INITIALIZE_OUTPUT(VertexOutputForwardBase, o);
    UNITY_TRANSFER_INSTANCE_ID(v, o);
    UNITY_INITIALIZE_VERTEX_OUTPUT_STEREO(o);

    float4 posWorld = mul(unity_ObjectToWorld, v.vertex);
    #if UNITY_REQUIRE_FRAG_WORLDPOS
        #if UNITY_PACK_WORLDPOS_WITH_TANGENT
            o.tangentToWorldAndPackedData[0].w = posWorld.x;
            o.tangentToWorldAndPackedData[1].w = posWorld.y;
            o.tangentToWorldAndPackedData[2].w = posWorld.z;
        #else
            o.posWorld = posWorld.xyz;
        #endif
    #endif
    o.pos = UnityObjectToClipPos(v.vertex);

    o.tex = TexCoords(v);
    o.eyeVec = NormalizePerVertexNormal(posWorld.xyz - _WorldSpaceCameraPos);
    float3 normalWorld = UnityObjectToWorldNormal(v.normal);
    #ifdef _TANGENT_TO_WORLD
        float4 tangentWorld = float4(UnityObjectToWorldDir(v.tangent.xyz), v.tangent.w);

        float3x3 tangentToWorld = CreateTangentToWorldPerVertex(normalWorld, tangentWorld.xyz, tangentWorld.w);
        o.tangentToWorldAndPackedData[0].xyz = tangentToWorld[0];
        o.tangentToWorldAndPackedData[1].xyz = tangentToWorld[1];
        o.tangentToWorldAndPackedData[2].xyz = tangentToWorld[2];
    #else
        o.tangentToWorldAndPackedData[0].xyz = 0;
        o.tangentToWorldAndPackedData[1].xyz = 0;
        o.tangentToWorldAndPackedData[2].xyz = normalWorld;
    #endif

    //We need this for shadow receving
    UNITY_TRANSFER_SHADOW(o, v.uv1);

    o.ambientOrLightmapUV = VertexGIForward(v, posWorld, normalWorld);

    #ifdef _PARALLAXMAP
        TANGENT_SPACE_ROTATION;
        half3 viewDirForParallax = mul (rotation, ObjSpaceViewDir(v.vertex));
        o.tangentToWorldAndPackedData[0].w = viewDirForParallax.x;
        o.tangentToWorldAndPackedData[1].w = viewDirForParallax.y;
        o.tangentToWorldAndPackedData[2].w = viewDirForParallax.z;
    #endif

    UNITY_TRANSFER_FOG(o,o.pos);
    return o;
}

输入为VertexInput结构体，分析见后文。

总体来说做了以下的工作：

1.初始化顶点信息。这部分是GPU Instancing的相关宏定义，位于UnityInstancing.cginc中。

2.顶点世界坐标计算，并根据Shader Mode的不同来将其存储在posWorld（SM3.0）或tangentToWorldAndPackedData[3]的w分量（SM2.0）中。在SM3.0下，tangentToWorldAndPackedData[3]的w分量用来存储视差视线。

3.计算裁剪空间顶点坐标，纹理坐标，世界空间视线以及法线。TexCoords函数实现在UnityStandardInput.cginc，UnityObjectToClipPosInstanced在UnityInstancing.cginc，NormalizePerVertexNormal在UnityStandardCore.cginc，不赘述。

4.计算切线空间变换到世界空间矩阵。CreateTangentToWorldPerVertex位于UnityStandardUtils.cginc。

5.阴影坐标转换，雾坐标转换。UNITY_TRANSFER_SHADOW位于AutoLight.cginc，UNITY_TRANSFER_FOG位于UnityCG.cginc。雾的计算会根据SM不同，选择逐顶点或逐像素的计算。

6.视差视线计算，ObjSpaceViewDir和rotation都位于UnityCG.cginc

7.环境或光照贴图纹理坐标的计算，VertexGIForward的实现位于UnityStandardCore.cginc。

VertexInput结构体位于UnityStandardInput.cginc，具体实现如下。

struct VertexInput
{
    float4 vertex   : POSITION;
    half3 normal    : NORMAL;
    float2 uv0      : TEXCOORD0;
    float2 uv1      : TEXCOORD1;
#if defined(DYNAMICLIGHTMAP_ON) || defined(UNITY_PASS_META)
    float2 uv2      : TEXCOORD2;
#endif
#ifdef _TANGENT_TO_WORLD
    half4 tangent   : TANGENT;
#endif
    UNITY_VERTEX_INPUT_INSTANCE_ID
};

VertexInput结构体包含了Unity提供给顶点着色器的模型的各项信息，包括模型空间的顶点坐标，纹理坐标，顶点法线和切线。纹理坐标有三个，第一个是贴图的纹理坐标，第二个是静态光照UV（Bake GI），第三个是动态光照UV（Precompute Realtime GI）。

VertexGIForward主要进行顶点的GI计算，具体的实现分析如下。

inline half4 VertexGIForward(VertexInput v, float3 posWorld, half3 normalWorld)
{
    half4 ambientOrLightmapUV = 0;
    // Static lightmaps
    #ifdef LIGHTMAP_ON
        ambientOrLightmapUV.xy = v.uv1.xy * unity_LightmapST.xy + unity_LightmapST.zw;
        ambientOrLightmapUV.zw = 0;
    // Sample light probe for Dynamic objects only (no static or dynamic lightmaps)
    #elif UNITY_SHOULD_SAMPLE_SH
        #ifdef VERTEXLIGHT_ON
            // Approximated illumination from non-important point lights
            ambientOrLightmapUV.rgb = Shade4PointLights (
                unity_4LightPosX0, unity_4LightPosY0, unity_4LightPosZ0,
                unity_LightColor[0].rgb, unity_LightColor[1].rgb, unity_LightColor[2].rgb, unity_LightColor[3].rgb,
                unity_4LightAtten0, posWorld, normalWorld);
        #endif

        ambientOrLightmapUV.rgb = ShadeSHPerVertex (normalWorld, ambientOrLightmapUV.rgb);
    #endif

    #ifdef DYNAMICLIGHTMAP_ON
        ambientOrLightmapUV.zw = v.uv2.xy * unity_DynamicLightmapST.xy + unity_DynamicLightmapST.zw;
    #endif

    return ambientOrLightmapUV;
}

输入：VertexInput，顶点世界坐标，世界法线。

Unity的GI实现有两种方式，一种是烘焙GI（Bake GI），一种是预计算实时GI（PRGI，Precompute Realtime GI）。对于烘焙GI来说，lightmap是静态的；对于预计算实时GI来说，lightmap是动态的。预计算实时GI在Unity官方中文论坛有比较详细的介绍，不赘述。

然而，在VertexGIForward的计算中，根据GI的实现方式有三个分支。首先是烘焙GI的实现，unity_LightmapST的xy记录了lightmap的scale值，zw记录了lightmap的offset值。第二个分支是SH(球谐函数)的计算，SH的计算在存在GI的情况下是不进行计算的，因为lightmap中已经包含了漫反射间接环境光照。所以在没有lightmap的情况下，进行SH计算。追求效率，用于SH计算的点光源被设置为了4个，同时unity在QualitySetting中的pixel light count也是设置为4。第三个分支是预计算实时GI的实现，unity_DynamicLightmapST和unity_LightmapST类似。

ambientOrLightmapUV在启用光照贴图的情况下，其xyzw分量用来存储光照贴图的UV。在不启用光照贴图的情况下，其rgb(xyz)分量用来保存SH计算的颜色。

Shade4PointLights，ShadeSHPerVertex函数实现暂时不赘述，Shade4PointLights计算四个点光源的方式比较巧妙，有时间会讲。

fragForwardBaseInternal函数

half4 fragForwardBaseInternal (VertexOutputForwardBase i)
{
    UNITY_APPLY_DITHER_CROSSFADE(i.pos.xy);

    FRAGMENT_SETUP(s)

    UNITY_SETUP_INSTANCE_ID(i);
    UNITY_SETUP_STEREO_EYE_INDEX_POST_VERTEX(i);

    UnityLight mainLight = MainLight ();
    UNITY_LIGHT_ATTENUATION(atten, i, s.posWorld);

    half occlusion = Occlusion(i.tex.xy);
    UnityGI gi = FragmentGI (s, occlusion, i.ambientOrLightmapUV, atten, mainLight);

    half4 c = UNITY_BRDF_PBS (s.diffColor, s.specColor, s.oneMinusReflectivity, s.smoothness, s.normalWorld, -s.eyeVec, gi.light, gi.indirect);
    c.rgb += Emission(i.tex.xy);

    UNITY_APPLY_FOG(i.fogCoord, c.rgb);
    return OutputForward (c, s.alpha);
}

总体来说做了以下的工作：

1.初始化片段设置，暂时不赘述。

2.获取设置的主光源信息。UnityLight结构体记录了灯光的方向，颜色。MainLight函数则把主光源的信息填充到结构体中。

UnityLight结构体定义在UnityLightingCommon.cginc中。

struct UnityLight
{
    half3 color;
    half3 dir;
    half  ndotl; // Deprecated: Ndotl is now calculated on the fly and is no longer stored. Do not used it.
};

MainLight函数定义在UnityStandardCore.cginc中。

UnityLight MainLight ()
{
    UnityLight l;

    l.color = _LightColor0.rgb;
    l.dir = _WorldSpaceLightPos0.xyz;
    return l;
}

3.计算灯光的衰减信息。

4.计算遮罩，遮罩的意义是利用Occlusion Map以及Occlusion Strength（SM3.0）来控制物体表面接收间接光照的强度。

Occlusion函数——UnityStandardInput.cginc中。

half Occlusion(float2 uv)
{
#if (SHADER_TARGET < 30)
    // SM20: instruction count limitation
    // SM20: simpler occlusion
    return tex2D(_OcclusionMap, uv).g;
#else
    half occ = tex2D(_OcclusionMap, uv).g;
    return LerpOneTo (occ, _OcclusionStrength);
#endif
}

实现原理比较简单：在SM2.0下，由于指令数限制，直接从遮罩图中采样返回green通道即可；在SM3.0下，采样后，需要多做一步计算。Occlusion Map一般是一张灰度图，Unity这里只使用了它的Green通道。_OcclusionMap和_OcclusionStrength都是Standard Shader暴露给编辑器的变量。

LerpOneTo函数——UnityStandardUtils.cginc

half LerpOneTo(half b, half t)
{
    half oneMinusT = 1 - t;
    return oneMinusT + b * t;
}

LerpOneTo的实现比较简单，类似Lerp函数的计算，返回值为1-_OcclusionStrength+occ*_OcclusionStrength。其实也就等价于Lerp(1,occ,_OcclusionStrength)。

5.计算片段GI。

在分析Fragment函数之前，先简单过一下Fragment函数用到的几个结构体。

FragmentCommonData：包括了片段函数需要的一些常规的数据，包括漫反射颜色，镜面反射颜色，一减反射率，平滑度，世界空间法线，视线，世界空间坐标，alpha。以及如果是simple模式下，定义反射uvw和切线空间法线。

struct FragmentCommonData
{
    half3 diffColor, specColor;
    // Note: smoothness & oneMinusReflectivity for optimization purposes, mostly for DX9 SM2.0 level.
    // Most of the math is being done on these (1-x) values, and that saves a few precious ALU slots.
    half oneMinusReflectivity, smoothness;
    half3 normalWorld, eyeVec, posWorld;
    half alpha;

#if UNITY_STANDARD_SIMPLE
    half3 reflUVW;
#endif

#if UNITY_STANDARD_SIMPLE
    half3 tangentSpaceNormal;
#endif
};

UnityGI：记录了一个记录灯光信息的UnityLight对象和一个记录间接光照信息的UnityIndirect对象。

struct UnityGI
{
    UnityLight light;
    UnityIndirect indirect;
};

UnityIndirect结构体的两个变量分别是漫反射颜色和镜面反射颜色。

struct UnityIndirect
{
    half3 diffuse;
    half3 specular;
};

UnityGIInput——-UnityLightingCommon.cginc

struct UnityGIInput
{
    UnityLight light; // pixel light, sent from the engine

    float3 worldPos;
    half3 worldViewDir;
    half atten;
    half3 ambient;

    // interpolated lightmap UVs are passed as full float precision data to fragment shaders
    // so lightmapUV (which is used as a tmp inside of lightmap fragment shaders) should
    // also be full float precision to avoid data loss before sampling a texture.
    float4 lightmapUV; // .xy = static lightmap UV, .zw = dynamic lightmap UV

    #if defined(UNITY_SPECCUBE_BLENDING) || defined(UNITY_SPECCUBE_BOX_PROJECTION)
    float4 boxMin[2];
    #endif
    #ifdef UNITY_SPECCUBE_BOX_PROJECTION
    float4 boxMax[2];
    float4 probePosition[2];
    #endif
    // HDR cubemap properties, use to decompress HDR texture
    float4 probeHDR[2];
};

包括一个UnityLight对象，以及片段的世界空间坐标，世界空间视线，灯光的衰减，环境色。光照贴图的UV，出于精度考虑使用float来避免光照贴图采样精度丢失。xy分量是静态光照贴图UV，zw分量是动态光照贴图UV。之后是用于反射探针盒投影，反射探针混合，以及HDR天空的变量，在FragmentGI函数中会用到。

UNITY_SPECCUBE_BOX_PROJECTION&UNITY_SPECCUBE_BLENDING——UnityStandardConfig.cginc。

// "platform caps" defines: they are controlled from TierSettings (Editor will determine values and pass them to compiler)
// UNITY_SPECCUBE_BOX_PROJECTION:                   TierSettings.reflectionProbeBoxProjection
// UNITY_SPECCUBE_BLENDING:                         TierSettings.reflectionProbeBlending
// UNITY_ENABLE_DETAIL_NORMALMAP:                   TierSettings.detailNormalMap
// UNITY_USE_DITHER_MASK_FOR_ALPHABLENDED_SHADOWS:  TierSettings.semitransparentShadows

TierSettings来控制的，当设置为启用时，会自动生成相关的宏定义。具体使用查看Unity Scripting API。

UNITY_SPECCUBE_BOX_PROJECTION: TierSettings.reflectionProbeBoxProjection——指定反射探针盒投影是否启用。
UNITY_SPECCUBE_BLENDING: TierSettings.reflectionProbeBlending——指定反射探针混合是否启用。

FragmentGI函数——UnityStandardCore.cginc

inline UnityGI FragmentGI (FragmentCommonData s, half occlusion, half4 i_ambientOrLightmapUV, half atten, UnityLight light, bool reflections)
{
    UnityGIInput d;
    d.light = light;
    d.worldPos = s.posWorld;
    d.worldViewDir = -s.eyeVec;
    d.atten = atten;
    #if defined(LIGHTMAP_ON) || defined(DYNAMICLIGHTMAP_ON)
        d.ambient = 0;
        d.lightmapUV = i_ambientOrLightmapUV;
    #else
        d.ambient = i_ambientOrLightmapUV.rgb;
        d.lightmapUV = 0;
    #endif

    d.probeHDR[0] = unity_SpecCube0_HDR;
    d.probeHDR[1] = unity_SpecCube1_HDR;
    #if defined(UNITY_SPECCUBE_BLENDING) || defined(UNITY_SPECCUBE_BOX_PROJECTION)
      d.boxMin[0] = unity_SpecCube0_BoxMin; // .w holds lerp value for blending
    #endif
    #ifdef UNITY_SPECCUBE_BOX_PROJECTION
      d.boxMax[0] = unity_SpecCube0_BoxMax;
      d.probePosition[0] = unity_SpecCube0_ProbePosition;
      d.boxMax[1] = unity_SpecCube1_BoxMax;
      d.boxMin[1] = unity_SpecCube1_BoxMin;
      d.probePosition[1] = unity_SpecCube1_ProbePosition;
    #endif

    if(reflections)
    {
        Unity_GlossyEnvironmentData g = UnityGlossyEnvironmentSetup(s.smoothness, -s.eyeVec, s.normalWorld, s.specColor);
        // Replace the reflUVW if it has been compute in Vertex shader. Note: the compiler will optimize the calcul in UnityGlossyEnvironmentSetup itself
        #if UNITY_STANDARD_SIMPLE
            g.reflUVW = s.reflUVW;
        #endif

        return UnityGlobalIllumination (d, occlusion, s.normalWorld, g);
    }
    else
    {
        return UnityGlobalIllumination (d, occlusion, s.normalWorld);
    }
}

FragmentGI函数主要可以分为两个部分，一个是填充UnityGIInput结构体，一个计算反射调用UnityGlobalIllumination函数的过程。

填充UnityGIInput的过程，灯光+世界空间顶点坐标+观察方向（视线的反方向）+衰减直接赋值即可。随后是光照贴图，在启用了静态光照贴图或者动态光照贴图的情况下，环境光为0，然后获得光照贴图的UV。否则的话，ambient直接使用VertexGIForward计算的rgb值。

然后是反射探针的相关计算，这部分计算需要涉及到一系列变量声明。

Reflection Probes——UnityShaderVariables.cginc

UNITY_DECLARE_TEXCUBE(unity_SpecCube0);
UNITY_DECLARE_TEXCUBE_NOSAMPLER(unity_SpecCube1);

CBUFFER_START(UnityReflectionProbes)
    float4 unity_SpecCube0_BoxMin;
    float4 unity_SpecCube0_ProbePosition;
    half4  unity_SpecCube0_HDR;

    float4 unity_SpecCube1_BoxMax;
    float4 unity_SpecCube1_BoxMin;
    float4 unity_SpecCube1_ProbePosition;
    half4  unity_SpecCube1_HDR;
CBUFFER_END

UNITY_DECLARE_TEXCUBE：声明了一个TextureCube类型的对象。
UNITY_DECLARE_TEXCUBE_NOSAMPLER：声明了一个TextureCube类型的对象（无Sampler）。

CBUFFER_START&CBUFFER_END：声明了一块常量缓冲区。

以上的宏定义在HLSLSupport.cginc文件中。

HDR探针将常量缓冲中的对应变量保存。在UNITY_SPECCUBE_BOX_PROJECTION或者UNITY_SPECCUBE_BLENDING被启用的情况下，保存相应的反射探针属性。

如果反射为真的情况下，计算反射，先计算反射的环境数据，包括镜面照明和天空等。然后调用UnityGlobalIllumination函数。

Unity_GlossyEnvironmentData结构体：延迟渲染只有一个cubemap，前向渲染的情况下可以有两个混合的cubemap，不常用应该会被弃用。另外，粗糙度这里是感性粗糙度，因为兼容性而使用了粗糙度的变量名。关于粗糙度和感性粗糙度之间的区别，见后文。

struct Unity_GlossyEnvironmentData
{
    // - Deferred case have one cubemap
    // - Forward case can have two blended cubemap (unusual should be deprecated).

    // Surface properties use for cubemap integration
    half    roughness; // CAUTION: This is perceptualRoughness but because of compatibility this name can't be change :(
    half3   reflUVW;
};

UnityGlossyEnvironmentSetup函数：计算感性粗糙度，计算反射，并返回对象。

Unity_GlossyEnvironmentData UnityGlossyEnvironmentSetup(half Smoothness, half3 worldViewDir, half3 Normal, half3 fresnel0)
{
    Unity_GlossyEnvironmentData g;

    g.roughness /* perceptualRoughness */   = SmoothnessToPerceptualRoughness(Smoothness);
    g.reflUVW   = reflect(-worldViewDir, Normal);

    return g;
}

UnityGlobalIllumination函数

inline UnityGI UnityGlobalIllumination (UnityGIInput data, half occlusion, half3 normalWorld)
{
    return UnityGI_Base(data, occlusion, normalWorld);
}

inline UnityGI UnityGlobalIllumination (UnityGIInput data, half occlusion, half3 normalWorld, Unity_GlossyEnvironmentData glossIn)
{
    UnityGI o_gi = UnityGI_Base(data, occlusion, normalWorld);
    o_gi.indirect.specular = UnityGI_IndirectSpecular(data, occlusion, glossIn);
    return o_gi;
}

UnityGlobalIllumination函数位于同名的cginc文件中，同名的函数有四个（形参不同）。有两个函数实现是旧版本的，并在注释上说明了只是为了旧版本兼容即将被移除，所以我们这里只看最新的函数实现。

由FragmentGI函数可知，两种实现分别对应无反射和有反射两种情况。

函数1（无反射）：直接返回UnityGI_Base函数的值，UnityGI_Base解释见下文。

函数2（有反射）：调用UnityGI_Base函数，得到返回值o_gi 然后调用UnityGI_IndirectSpecular修改o_gi.indirect.specular的值，UnityGI_IndirectSpecular解释见下文。

UnityGI_Base函数

inline UnityGI UnityGI_Base(UnityGIInput data, half occlusion, half3 normalWorld)
{
    UnityGI o_gi;
    ResetUnityGI(o_gi);

    // Base pass with Lightmap support is responsible for handling ShadowMask / blending here for performance reason
    #if defined(HANDLE_SHADOWS_BLENDING_IN_GI)
        half bakedAtten = UnitySampleBakedOcclusion(data.lightmapUV.xy, data.worldPos);
        float zDist = dot(_WorldSpaceCameraPos - data.worldPos, UNITY_MATRIX_V[2].xyz);
        float fadeDist = UnityComputeShadowFadeDistance(data.worldPos, zDist);
        data.atten = UnityMixRealtimeAndBakedShadows(data.atten, bakedAtten, UnityComputeShadowFade(fadeDist));
    #endif

    o_gi.light = data.light;
    o_gi.light.color *= data.atten;

    #if UNITY_SHOULD_SAMPLE_SH
        o_gi.indirect.diffuse = ShadeSHPerPixel (normalWorld, data.ambient, data.worldPos);
    #endif

    #if defined(LIGHTMAP_ON)
        // Baked lightmaps
        half4 bakedColorTex = UNITY_SAMPLE_TEX2D(unity_Lightmap, data.lightmapUV.xy);
        half3 bakedColor = DecodeLightmap(bakedColorTex);

        #ifdef DIRLIGHTMAP_COMBINED
            fixed4 bakedDirTex = UNITY_SAMPLE_TEX2D_SAMPLER (unity_LightmapInd, unity_Lightmap, data.lightmapUV.xy);
            o_gi.indirect.diffuse = DecodeDirectionalLightmap (bakedColor, bakedDirTex, normalWorld);

            #if defined(LIGHTMAP_SHADOW_MIXING) && !defined(SHADOWS_SHADOWMASK) && defined(SHADOWS_SCREEN)
                ResetUnityLight(o_gi.light);
                o_gi.indirect.diffuse = SubtractMainLightWithRealtimeAttenuationFromLightmap (o_gi.indirect.diffuse, data.atten, bakedColorTex, normalWorld);
            #endif

        #else // not directional lightmap
            o_gi.indirect.diffuse = bakedColor;

            #if defined(LIGHTMAP_SHADOW_MIXING) && !defined(SHADOWS_SHADOWMASK) && defined(SHADOWS_SCREEN)
                ResetUnityLight(o_gi.light);
                o_gi.indirect.diffuse = SubtractMainLightWithRealtimeAttenuationFromLightmap(o_gi.indirect.diffuse, data.atten, bakedColorTex, normalWorld);
            #endif

        #endif
    #endif

    #ifdef DYNAMICLIGHTMAP_ON
        // Dynamic lightmaps
        fixed4 realtimeColorTex = UNITY_SAMPLE_TEX2D(unity_DynamicLightmap, data.lightmapUV.zw);
        half3 realtimeColor = DecodeRealtimeLightmap (realtimeColorTex);

        #ifdef DIRLIGHTMAP_COMBINED
            half4 realtimeDirTex = UNITY_SAMPLE_TEX2D_SAMPLER(unity_DynamicDirectionality, unity_DynamicLightmap, data.lightmapUV.zw);
            o_gi.indirect.diffuse += DecodeDirectionalLightmap (realtimeColor, realtimeDirTex, normalWorld);
        #else
            o_gi.indirect.diffuse += realtimeColor;
        #endif
    #endif

    o_gi.indirect.diffuse *= occlusion;
    return o_gi;
}

函数依次实现的是阴影遮罩，SH计算（非静态GI非动态GI），静态GI，动态GI。

ShadowMask阴影遮罩是Unity5.6版本的新特性。

烘焙GI的实现：首先从烘焙的光照贴图中取得对应的像素值，其次根据编码对像素进行译码。如果定义了directional lightmap混合，会进行和烘焙光照贴图类似的过程并进行混合计算。否则的话，直接使用烘焙GI的颜色进行计算。

动态GI的实现：和静态GI的实现类似。

ResetUnityGI函数：将UnityGI结构体初始化，无需多说。

inline void ResetUnityGI(out UnityGI outGI)
{
    ResetUnityLight(outGI.light);
    outGI.indirect.diffuse = 0;
    outGI.indirect.specular = 0;
}

UnityGI_IndirectSpecular函数

inline half3 UnityGI_IndirectSpecular(UnityGIInput data, half occlusion, Unity_GlossyEnvironmentData glossIn)
{
    half3 specular;

    #ifdef UNITY_SPECCUBE_BOX_PROJECTION
        // we will tweak reflUVW in glossIn directly (as we pass it to Unity_GlossyEnvironment twice for probe0 and probe1), so keep original to pass into BoxProjectedCubemapDirection
        half3 originalReflUVW = glossIn.reflUVW;
        glossIn.reflUVW = BoxProjectedCubemapDirection (originalReflUVW, data.worldPos, data.probePosition[0], data.boxMin[0], data.boxMax[0]);
    #endif

    #ifdef _GLOSSYREFLECTIONS_OFF
        specular = unity_IndirectSpecColor.rgb;
    #else
        half3 env0 = Unity_GlossyEnvironment (UNITY_PASS_TEXCUBE(unity_SpecCube0), data.probeHDR[0], glossIn);
        #ifdef UNITY_SPECCUBE_BLENDING
            const float kBlendFactor = 0.99999;
            float blendLerp = data.boxMin[0].w;
            UNITY_BRANCH
            if (blendLerp < kBlendFactor)
            {
                #ifdef UNITY_SPECCUBE_BOX_PROJECTION
                    glossIn.reflUVW = BoxProjectedCubemapDirection (originalReflUVW, data.worldPos, data.probePosition[1], data.boxMin[1], data.boxMax[1]);
                #endif

                half3 env1 = Unity_GlossyEnvironment (UNITY_PASS_TEXCUBE_SAMPLER(unity_SpecCube1,unity_SpecCube0), data.probeHDR[1], glossIn);
                specular = lerp(env1, env0, blendLerp);
            }
            else
            {
                specular = env0;
            }
        #else
            specular = env0;
        #endif
    #endif

    return specular * occlusion;
}

// Deprecated old prototype but can't be move to Deprecated.cginc file due to order dependency
inline half3 UnityGI_IndirectSpecular(UnityGIInput data, half occlusion, half3 normalWorld, Unity_GlossyEnvironmentData glossIn)
{
    // normalWorld is not used
    return UnityGI_IndirectSpecular(data, occlusion, glossIn);
}

BoxProjectedCubemapDirection函数

inline half3 BoxProjectedCubemapDirection (half3 worldRefl, float3 worldPos, float4 cubemapCenter, float4 boxMin, float4 boxMax)
{
    // Do we have a valid reflection probe?
    UNITY_BRANCH
    if (cubemapCenter.w > 0.0)
    {
        half3 nrdir = normalize(worldRefl);

        #if 1
            half3 rbmax = (boxMax.xyz - worldPos) / nrdir;
            half3 rbmin = (boxMin.xyz - worldPos) / nrdir;

            half3 rbminmax = (nrdir > 0.0f) ? rbmax : rbmin;

        #else // Optimized version
            half3 rbmax = (boxMax.xyz - worldPos);
            half3 rbmin = (boxMin.xyz - worldPos);

            half3 select = step (half3(0,0,0), nrdir);
            half3 rbminmax = lerp (rbmax, rbmin, select);
            rbminmax /= nrdir;
        #endif

        half fa = min(min(rbminmax.x, rbminmax.y), rbminmax.z);

        worldPos -= cubemapCenter.xyz;
        worldRefl = worldPos + nrdir * fa;
    }
    return worldRefl;
}

 6.计算颜色，这部分为重点内容，后文详细讲解。

7.计算自发光。

half3 Emission(float2 uv)
{
#ifndef _EMISSION
    return 0;
#else
    return tex2D(_EmissionMap, uv).rgb * _EmissionColor.rgb;
#endif
}

8.应用雾

// ------------------------------------------------------------------
//  Fog helpers
//
//  multi_compile_fog Will compile fog variants.
//  UNITY_FOG_COORDS(texcoordindex) Declares the fog data interpolator.
//  UNITY_TRANSFER_FOG(outputStruct,clipspacePos) Outputs fog data from the vertex shader.
//  UNITY_APPLY_FOG(fogData,col) Applies fog to color "col". Automatically applies black fog when in forward-additive pass.
//  Can also use UNITY_APPLY_FOG_COLOR to supply your own fog color.

// In case someone by accident tries to compile fog code in one of the g-buffer or shadow passes:
// treat it as fog is off.
#if defined(UNITY_PASS_PREPASSBASE) || defined(UNITY_PASS_DEFERRED) || defined(UNITY_PASS_SHADOWCASTER)
#undef FOG_LINEAR
#undef FOG_EXP
#undef FOG_EXP2
#endif

#if defined(UNITY_REVERSED_Z)
    //D3d with reversed Z => z clip range is [near, 0] -> remapping to [0, far]
    //max is required to protect ourselves from near plane not being correct/meaningfull in case of oblique matrices.
    #define UNITY_Z_0_FAR_FROM_CLIPSPACE(coord) max(((1.0-(coord)/_ProjectionParams.y)*_ProjectionParams.z),0)
#elif UNITY_UV_STARTS_AT_TOP
    //D3d without reversed z => z clip range is [0, far] -> nothing to do
    #define UNITY_Z_0_FAR_FROM_CLIPSPACE(coord) (coord)
#else
    //Opengl => z clip range is [-near, far] -> should remap in theory but dont do it in practice to save some perf (range is close enought)
    #define UNITY_Z_0_FAR_FROM_CLIPSPACE(coord) (coord)
#endif

#if defined(FOG_LINEAR)
    // factor = (end-z)/(end-start) = z * (-1/(end-start)) + (end/(end-start))
    #define UNITY_CALC_FOG_FACTOR_RAW(coord) float unityFogFactor = (coord) * unity_FogParams.z + unity_FogParams.w
#elif defined(FOG_EXP)
    // factor = exp(-density*z)
    #define UNITY_CALC_FOG_FACTOR_RAW(coord) float unityFogFactor = unity_FogParams.y * (coord); unityFogFactor = exp2(-unityFogFactor)
#elif defined(FOG_EXP2)
    // factor = exp(-(density*z)^2)
    #define UNITY_CALC_FOG_FACTOR_RAW(coord) float unityFogFactor = unity_FogParams.x * (coord); unityFogFactor = exp2(-unityFogFactor*unityFogFactor)
#else
    #define UNITY_CALC_FOG_FACTOR_RAW(coord) float unityFogFactor = 0.0
#endif

#define UNITY_CALC_FOG_FACTOR(coord) UNITY_CALC_FOG_FACTOR_RAW(UNITY_Z_0_FAR_FROM_CLIPSPACE(coord))

#define UNITY_FOG_COORDS_PACKED(idx, vectype) vectype fogCoord : TEXCOORD##idx;

#if defined(FOG_LINEAR) || defined(FOG_EXP) || defined(FOG_EXP2)
    #define UNITY_FOG_COORDS(idx) UNITY_FOG_COORDS_PACKED(idx, float1)

    #if (SHADER_TARGET < 30) || defined(SHADER_API_MOBILE)
        // mobile or SM2.0: calculate fog factor per-vertex
        #define UNITY_TRANSFER_FOG(o,outpos) UNITY_CALC_FOG_FACTOR((outpos).z); o.fogCoord.x = unityFogFactor
    #else
        // SM3.0 and PC/console: calculate fog distance per-vertex, and fog factor per-pixel
        #define UNITY_TRANSFER_FOG(o,outpos) o.fogCoord.x = (outpos).z
    #endif
#else
    #define UNITY_FOG_COORDS(idx)
    #define UNITY_TRANSFER_FOG(o,outpos)
#endif

#define UNITY_FOG_LERP_COLOR(col,fogCol,fogFac) col.rgb = lerp((fogCol).rgb, (col).rgb, saturate(fogFac))


#if defined(FOG_LINEAR) || defined(FOG_EXP) || defined(FOG_EXP2)
    #if (SHADER_TARGET < 30) || defined(SHADER_API_MOBILE)
        // mobile or SM2.0: fog factor was already calculated per-vertex, so just lerp the color
        #define UNITY_APPLY_FOG_COLOR(coord,col,fogCol) UNITY_FOG_LERP_COLOR(col,fogCol,(coord).x)
    #else
        // SM3.0 and PC/console: calculate fog factor and lerp fog color
        #define UNITY_APPLY_FOG_COLOR(coord,col,fogCol) UNITY_CALC_FOG_FACTOR((coord).x); UNITY_FOG_LERP_COLOR(col,fogCol,unityFogFactor)
    #endif
#else
    #define UNITY_APPLY_FOG_COLOR(coord,col,fogCol)
#endif

#ifdef UNITY_PASS_FORWARDADD
    #define UNITY_APPLY_FOG(coord,col) UNITY_APPLY_FOG_COLOR(coord,col,fixed4(0,0,0,0))
#else
    #define UNITY_APPLY_FOG(coord,col) UNITY_APPLY_FOG_COLOR(coord,col,unity_FogColor)
#endif

9.返回颜色和Alpha值。

UNITY_BRDF_PBS函数

代码中有三个BRDF的实现函数，对应不同的BRDF模型实现。这里主要分析第一个BRDF函数，也就是BRDF1_Unity_PBS，使用的BRDF模型的公式可以参考：http://graphicrants.blogspot.com/2013/08/specular-brdf-reference.html

首先，我们要清楚BRDF1_Unity_PBS函数的依据，即根据Torrance-Sparrow 微表面模型的公式：
f(l,v)=D(h)F(v,h)G(l,v,h)/(4(n⋅l)(n⋅v))
以及BRDF公式：
BRDF = kD / pi + kS * (D * V * F) / 4
然后拆分公式，一项项的实现。

D——微表面分布项
V——遮挡可见性项
F——菲涅尔反射项
kD——漫反射系数
kS——镜面反射系数
Note：V(Visibility)项即G(l,v,h)/(4(n⋅l)(n⋅v))的集合。

菲涅尔反射F

Schlick菲涅尔反射的公式为：

F=F0+(1-F0)*(1-（H*V))^5
F0：光线垂直入射时的表面反射率
H：半角向量
V：视线
此处输入的参数cosA，为saturate(dot(H,V))的值，也可能是abs(dot(H,V))。

inline half3 FresnelTerm (half3 F0, half cosA)
{
    half t = Pow5 (1 - cosA);   // ala Schlick interpoliation
    return F0 + (1-F0) * t;
}

V&D分支

V项和D项，在代码中出现了分支，用来选择不同的模型：
如果UNITY_BRDF_GGX为真，V项和D项使用GGX的公式来实现。
否则，V项和D项使用Smith-Beckmann和Blinn-Phong公式实现。

GGX相比Blinn-Phong，它更接近粗糙表面上真实反射的外观。

分支选择代码如下：

#if UNITY_BRDF_GGX
    // GGX with roughtness to 0 would mean no specular at all, using max(roughness, 0.002) here to match HDrenderloop roughtness remapping.
    roughness = max(roughness, 0.002);
    half V = SmithJointGGXVisibilityTerm (nl, nv, roughness);
    float D = GGXTerm (nh, roughness);
#else
    // Legacy
    half V = SmithBeckmannVisibilityTerm (nl, nv, roughness);
    half D = NDFBlinnPhongNormalizedTerm (nh, PerceptualRoughnessToSpecPower(perceptualRoughness));
#endif

遮挡可见性项V

依次分析两个Visbility函数。首先是SmithJointGGXVisibilityTerm ，Unity使用了简化版的近似公式。

在Unity中V项即是公式中的G项，只是Unity为了简化运算，使得V=G/(4(n⋅l)(n⋅v))。

SmithJointGGXVisibilityTerm使用了Smith-Joint GGX公式，具体的G项公式见：

https://hal.inria.fr/hal-00942452v1/document

Page 26 ——section 5 Equation (21)

计算lambda的公式见：

Page 13——section 3.2 

χ+(α)：Heaviside function: 1 if a > 0 and 0 if a ≤ 0

上面就是Original formulation这部分被注释掉的原始公式了，正式的代码对上述公式进行了以下优化：

1.每个lambda都+0.5，来抵消 G = 1 / (1 + lambda_v + lambda_l)中分母的1。

2.每项lambda乘(2.0f * NdotL * NdotV)进行化简并整理即可得出。

3.V=G/(4(n⋅l)(n⋅v))的计算，来抵消部分上式乘的(2.0f * NdotL * NdotV)。

当然，因为#if 0以上的代码都没有执行……

Unity最后使用了Smith-Joint的近似公式（UE4使用了同样的公式），因为上述的代码依旧很复杂。考虑到整体的性能，牺牲一部分难以察觉到的精度来提升效率是必要的。

lambdaV = NdotL * sqrt((-NdotV * a2 + NdotV) * NdotV + a2);

lambdaV = NdotL * sqrt((a2-1)* NdotV2  + a2);

lambdaV ≈ NdotL * ((a-1)* NdotV + a）;

1e-5f用来避免分母为0的情况。

inline half SmithJointGGXVisibilityTerm (half NdotL, half NdotV, half roughness)
{
#if 0
    // Original formulation:
    //  lambda_v    = (-1 + sqrt(a2 * (1 - NdotL2) / NdotL2 + 1)) * 0.5f;
    //  lambda_l    = (-1 + sqrt(a2 * (1 - NdotV2) / NdotV2 + 1)) * 0.5f;
    //  G           = 1 / (1 + lambda_v + lambda_l);

    // Reorder code to be more optimal
    half a          = roughness;
    half a2         = a * a;

    half lambdaV    = NdotL * sqrt((-NdotV * a2 + NdotV) * NdotV + a2);
    half lambdaL    = NdotV * sqrt((-NdotL * a2 + NdotL) * NdotL + a2);

    // Simplify visibility term: (2.0f * NdotL * NdotV) /  ((4.0f * NdotL * NdotV) * (lambda_v + lambda_l + 1e-5f));
    return 0.5f / (lambdaV + lambdaL + 1e-5f);  // This function is not intended to be running on Mobile,
                                                // therefore epsilon is smaller than can be represented by half
#else
    // Approximation of the above formulation (simplify the sqrt, not mathematically correct but close enough)
    half a = roughness;
    half lambdaV = NdotL * (NdotV * (1 - a) + a);
    half lambdaL = NdotV * (NdotL * (1 - a) + a);

    return 0.5f / (lambdaV + lambdaL + 1e-5f);
#endif
}

 然后是下一个V项的实现，SmithBeckmannVisibilityTerm函数，并在函数中调用了SmithVisibilityTerm函数。

使用的是SmithBeckmann公式，公式具体见首段链接。

SmithBeckmannVisibilityTerm函数计算了k的值，并调用SmithVisibilityTerm函数进一步计算，*0.25是先抵消V=G/(4(n⋅l)(n⋅v))中的4。

SmithBeckmann公式的分子在SmithVisibilityTerm中的gL&gV计算中与V=G/(4(n⋅l)(n⋅v))中的(n⋅l)(n⋅v)消项。

inline half SmithBeckmannVisibilityTerm (half NdotL, half NdotV, half roughness)
{
    half c = 0.797884560802865h; // c = sqrt(2 / Pi)
    half k = roughness * c;
    return SmithVisibilityTerm (NdotL, NdotV, k) * 0.25f; // * 0.25 is the 1/4 of the visibility term
}

inline half SmithVisibilityTerm (half NdotL, half NdotV, half k)
{
    half gL = NdotL * (1-k) + k;
    half gV = NdotV * (1-k) + k;
    return 1.0 / (gL * gV + 1e-5f); // This function is not intended to be running on Mobile,
                                    // therefore epsilon is smaller than can be represented by half
}

微表面分布项D

微表面分布项依旧有两个实现，依旧从GGX的实现开始。

没什么好说的，这里直接根据首段链接里的公式转换为代码即可。

UNITY_INV_PI =1 / UNITY_PI

inline float GGXTerm (float NdotH, float roughness)
{
    float a2 = roughness * roughness;
    float d = (NdotH * a2 - NdotH) * NdotH + 1.0f; // 2 mad
    return UNITY_INV_PI * a2 / (d * d + 1e-7f); // This function is not intended to be running on Mobile,
                                            // therefore epsilon is smaller than what can be represented by half
}

然后是BlinnPhong的D项实现，NDFBlinnPhongNormalizedTerm函数。这里需要注意的是函数的输入参数n，是公式中的Power项的计算，实现函数为PerceptualRoughnessToSpecPower。

normTerm即公式中的正态分布项（首项）。

inline half NDFBlinnPhongNormalizedTerm (half NdotH, half n)
{
    // norm = (n+2)/(2*pi)
    half normTerm = (n + 2.0) * (0.5/UNITY_PI);

    half specTerm = pow (NdotH, n);
    return specTerm * normTerm;
}

PerceptualRoughnessToSpecPower函数的具体实现如下，PerceptualRoughnessToRoughness函数是返回perceptualRoughness的平方（UE4的roughness=Unity的perceptualRoughness）。

inline half PerceptualRoughnessToSpecPower (half perceptualRoughness)
{
    half m = PerceptualRoughnessToRoughness(perceptualRoughness);   // m is the true academic roughness.
    half sq = max(1e-4f, m*m);
    half n = (2.0 / sq) - 2.0;                          // https://dl.dropboxusercontent.com/u/55891920/papers/mm_brdf.pdf
    n = max(n, 1e-4f);                                  // prevent possible cases of pow(0,0), which could happen when roughness is 1.0 and NdotH is zero
    return n;
}

PerceptualRoughnessToRoughness函数的具体实现如下，用于将感性粗糙度计算为学术意义上的粗糙度。perceptualRoughness的值等于1-smoothness，在SmoothnessToPerceptualRoughness函数中实现。

float PerceptualRoughnessToRoughness(float perceptualRoughness)
{
    return perceptualRoughness * perceptualRoughness;
}

SmoothnessToPerceptualRoughness函数的具体实现，用于计算感性粗糙度，smoothness即材质的光滑度贴图/参数。

float SmoothnessToPerceptualRoughness(float smoothness)
{
    return (1 - smoothness);
}

以上，Unity BRDF公式中的D/V/F项全部实现。

BRDF1_Unity_PBS函数

接下来，开始对代码中的BRDF1_Unity_PBS函数进行分析。由于函数代码比较长，在分析时，会对函数进行分块解释。同时，还会对调用的函数进行展示和讲解，所以为方便观看，我会把每块代码都标序号。

Part-1

输入：漫反射颜色，镜面反射颜色，一减反射率，平滑度，法线，视线，灯光，GI

1.计算感性粗糙度，实现函数见上文。

2.计算半角向量。

3.这段注释涉及到NdotV的取值问题。对于可见的像素来说，NdotV的取值不能为负值，但是因为透视投影和法线映射会造成这种情况。在这种情况下，法线应该修改的有效（例如面向摄像机）而不会造成奇怪的异常。但是这个修改的操作会占用一些ALU，是用户所不希望的。另一种方法是取NdotV的绝对值（不太正确，但还可以）。下面定义的宏用来控制两种实现方式，如果你的平台算术逻辑单元紧张的话，那就会将负值设置为0。这种校正对于使用Smith-Joint GGX能见度函数是很有用的，因为会导致粗糙表面的高光边缘异常会更明显。

Note：默认情况下禁用这块代码，因为跟SpeedTree使用的双面光照不兼容。

定义宏UNITY_HANDLE_CORRECTLY_NEGATIVE_NDOTV ，值为0。

0：取NdotV的绝对值。

1：计算NdotV，然后判断值的正负。若值为正，返回normal；若值为负，返回normal + viewDir * (-shiftAmount + 1e-5f)。最后，根据返回的值计算NdotV。

normal + viewDir * (-shiftAmount + 1e-5f)其实是对normal向量向视线向量接近的计算。对返回的结果应该进行重新规范化，为了节省ALU并没有这么做。之后的NdotV计算，saturate()其实没有必要了。但是因为对于操作的输出应用saturate()是没有开销的，因此这里依然使用了saturate()。

half4 BRDF1_Unity_PBS (half3 diffColor, half3 specColor, half oneMinusReflectivity, half smoothness,
    float3 normal, float3 viewDir,
    UnityLight light, UnityIndirect gi)
{
    float perceptualRoughness = SmoothnessToPerceptualRoughness (smoothness);
    float3 halfDir = Unity_SafeNormalize (float3(light.dir) + viewDir);

// NdotV should not be negative for visible pixels, but it can happen due to perspective projection and normal mapping
// In this case normal should be modified to become valid (i.e facing camera) and not cause weird artifacts.
// but this operation adds few ALU and users may not want it. Alternative is to simply take the abs of NdotV (less correct but works too).
// Following define allow to control this. Set it to 0 if ALU is critical on your platform.
// This correction is interesting for GGX with SmithJoint visibility function because artifacts are more visible in this case due to highlight edge of rough surface
// Edit: Disable this code by default for now as it is not compatible with two sided lighting used in SpeedTree.
#define UNITY_HANDLE_CORRECTLY_NEGATIVE_NDOTV 0

#if UNITY_HANDLE_CORRECTLY_NEGATIVE_NDOTV
    // The amount we shift the normal toward the view vector is defined by the dot product.
    half shiftAmount = dot(normal, viewDir);
    normal = shiftAmount < 0.0f ? normal + viewDir * (-shiftAmount + 1e-5f) : normal;
    // A re-normalization should be applied here but as the shift is small we don't do it to save ALU.
    //normal = normalize(normal);

    half nv = saturate(dot(normal, viewDir)); // TODO: this saturate should no be necessary here
#else
    half nv = abs(dot(normal, viewDir));    // This abs allow to limit artifact
#endif

Part-2

1.计算NdotL/NdotH/LdotV/LdotH用于后续计算。

2.计算漫反射项，使用的是DisneyDiffuse函数，之后会提到。

3.计算粗糙度，使用PerceptualRoughnessToRoughness函数将感性粗糙度转换到学术意义上的粗糙度，上文已讲解。

4.选择V项和D项的不同实现函数，上文已讲解。

理论上，应该对diffuse项除π（见首段BRDF公式），并且不乘specularTerm。

但是——1.这会导致shader看起来比原来的颜色暗很多。2.在引擎看来，设置为Non-importance的灯光当加入ambient SH计算时需要除π。（ambient SH解释见后文）

所以，Unity中的diffuseTerm 计算，并没有除π，反而乘了NdotL。

half nl = saturate(dot(normal, light.dir));
    float nh = saturate(dot(normal, halfDir));

    half lv = saturate(dot(light.dir, viewDir));
    half lh = saturate(dot(light.dir, halfDir));

    // Diffuse term
    half diffuseTerm = DisneyDiffuse(nv, nl, lh, perceptualRoughness) * nl;

    // Specular term
    // HACK: theoretically we should divide diffuseTerm by Pi and not multiply specularTerm!
    // BUT 1) that will make shader look significantly darker than Legacy ones
    // and 2) on engine side "Non-important" lights have to be divided by Pi too in cases when they are injected into ambient SH
    float roughness = PerceptualRoughnessToRoughness(perceptualRoughness);
#if UNITY_BRDF_GGX
    // GGX with roughtness to 0 would mean no specular at all, using max(roughness, 0.002) here to match HDrenderloop roughtness remapping.
    roughness = max(roughness, 0.002);
    half V = SmithJointGGXVisibilityTerm (nl, nv, roughness);
    float D = GGXTerm (nh, roughness);
#else
    // Legacy
    half V = SmithBeckmannVisibilityTerm (nl, nv, roughness);
    half D = NDFBlinnPhongNormalizedTerm (nh, PerceptualRoughnessToSpecPower(perceptualRoughness));
#endif

 DisneyDiffuse函数

输入：NdotV，NdotL，LdotH，感性粗糙度。

传统的漫反射计算使用的是Lambert模型，但是使用Lambert模型会使得物体的边缘过暗，和真实的表现有差异。因此，Disney的Diffuse计算采用了schlick的近似菲涅尔计算公式来弥补效果，公式见：

https://disney-animation.s3.amazonaws.com/library/s2012_pbs_disney_brdf_notes_v2.pdf

Page 14——section 5.3

在Unity的计算中，将公式的diffuseAlbedo / PI，即baseColor/π挪到函数外进行计算。

half DisneyDiffuse(half NdotV, half NdotL, half LdotH, half perceptualRoughness)
{
    half fd90 = 0.5 + 2 * LdotH * LdotH * perceptualRoughness;
    // Two schlick fresnel term
    half lightScatter   = (1 + (fd90 - 1) * Pow5(1 - NdotL));
    half viewScatter    = (1 + (fd90 - 1) * Pow5(1 - NdotV));

    return lightScatter * viewScatter;
}

 Part-3

BRDF1_Unity_PBS函数的最后一段代码，反而内容比较多。

1.计算高光项的一部分，菲涅尔项最后再添加。如果镜面反射高光关闭，那么镜面反射项为0。

2.如果开启了颜色空间GAMMA校正，那么这里会进行一次计算。

3.计算surfaceReduction参数。Unity在注释中给出了它的公式,但我并没有查到计算他的目的，在这里它用于间接光照的计算。

4.为了提供真正的Lambert照明，如果SpecColor的各个通道值均为0，那么就是全漫反射。

 any - returns true if a boolean scalar or any component of a boolean vector is true.

5.最后的color输出，分为三个部分：漫反射+镜面反射+表面衰减。

漫反射：输入的漫反射颜色（纹理）*GI的漫反射颜色（间接光照）+输入的漫反射颜色（纹理）*光照颜色（直接光照）*漫反射项

镜面反射：镜面反射项（V项和D项）*光照颜色（直接光照）*菲涅尔项（F项）

表面衰减：表面衰减系数*GI镜面反射（间接光照）*菲涅尔插值

gi的类型为UnityIndirect结构体，解释见下文。

 half specularTerm = V*D * UNITY_PI; // Torrance-Sparrow model, Fresnel is applied later

#   ifdef UNITY_COLORSPACE_GAMMA
        specularTerm = sqrt(max(1e-4h, specularTerm));
#   endif

    // specularTerm * nl can be NaN on Metal in some cases, use max() to make sure it's a sane value
    specularTerm = max(0, specularTerm * nl);
#if defined(_SPECULARHIGHLIGHTS_OFF)
    specularTerm = 0.0;
#endif

    // surfaceReduction = Int D(NdotH) * NdotH * Id(NdotL>0) dH = 1/(roughness^2+1)
    half surfaceReduction;
#   ifdef UNITY_COLORSPACE_GAMMA
        surfaceReduction = 1.0-0.28*roughness*perceptualRoughness;      // 1-0.28*x^3 as approximation for (1/(x^4+1))^(1/2.2) on the domain [0;1]
#   else
        surfaceReduction = 1.0 / (roughness*roughness + 1.0);           // fade in [0.5;1]
#   endif

    // To provide true Lambert lighting, we need to be able to kill specular completely.
    specularTerm *= any(specColor) ? 1.0 : 0.0;

    half grazingTerm = saturate(smoothness + (1-oneMinusReflectivity));
    half3 color =   diffColor * (gi.diffuse + light.color * diffuseTerm)
                    + specularTerm * light.color * FresnelTerm (specColor, lh)
                    + surfaceReduction * gi.specular * FresnelLerp (specColor, grazingTerm, nv);

    return half4(color, 1);
}

FresnelLerp函数

返回F0到F90之间的线性插值，t的实现和菲涅尔项中的实现一致。

inline half3 FresnelLerp (half3 F0, half3 F90, half cosA)
{
    half t = Pow5 (1 - cosA);   // ala Schlick interpoliation
    return lerp (F0, F90, t);
}

计算oneMinusReflectivity的函数是EnergyConservationBetweenDiffuseAndSpecular，位于UnityStandardUtils.cginc。余下代码用在他处，暂不讨论。

inline half3 EnergyConservationBetweenDiffuseAndSpecular (half3 albedo, half3 specColor, out half oneMinusReflectivity)
{
    oneMinusReflectivity = 1 - SpecularStrength(specColor);
    #if !UNITY_CONSERVE_ENERGY
        return albedo;
    #elif UNITY_CONSERVE_ENERGY_MONOCHROME
        return albedo * oneMinusReflectivity;
    #else
        return albedo * (half3(1,1,1) - specColor);
    #endif
}

SpecularStrength函数

如果Shader Mode<3.0，返回镜面反射颜色的R通道。

否则，返回镜面反射颜色RGB三个通道中最大的通道值。

Note：Shader Mode 2.0由于指令数的限制，简化了这项运算，直接返回R通道值（因为大多数的金属要么是单色，要么是淡黄色/黄色调的，主要影响的是R通道）。

half SpecularStrength(half3 specular)
{
    #if (SHADER_TARGET < 30)
        // SM2.0: instruction count limitation
        // SM2.0: simplified SpecularStrength
        return specular.r; // Red channel - because most metals are either monocrhome or with redish/yellowish tint
    #else
        return max (max (specular.r, specular.g), specular.b);
    #endif
}