Unity3D Shader编程之十一 深入理解Unity5中的Standard Shader&屏幕像素化特效的实现

Posted 浅墨_毛星云

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了Unity3D Shader编程之十一 深入理解Unity5中的Standard Shader&屏幕像素化特效的实现相关的知识,希望对你有一定的参考价值。

 

本系列文章由@浅墨_毛星云 出品,转载请注明出处。   
文章链接:  http://blog.csdn.net/poem_qianmo/article/details/50095705
作者:毛星云(浅墨)    微博:http://weibo.com/u/1723155442
本文工程使用的Unity3D版本:   5.2.1 


 
概要:续接上文,本文进一步讲解与分析了上文未讲完的Unity5中Standard Shader正向基础渲染通道源码的片段着色实现部分,以及对屏幕像素化后期特效进行了实现。
 
同样需要声明的是。本文中对Stardard Shader源码的一些分析,全是浅墨自己通过对Unity Shader内建源码的理解,以及Google之后理解与分析而来。如有解释不妥当之处,还请各位及时指出。
 

依然是附上一组本文配套工程的运行截图之后,便开始我们的正文。


傍晚的野外(with 屏幕像素化特效):
 


傍晚的野外(原始场景):



图依然是贴这两张。文章末尾有更多的运行截图,并提供了源工程的下载。先放出可运行的exe下载,如下:

【可运行的本文配套exe游戏场景请点击这里下载】
 
提示:在此游戏场景中按F键可以开关屏幕特效。


 

 
一、关于BRDF(双向反射分布函数)
 

本次源码剖析有涉及BRDF的相关内容,这边简单提一下。
双向反射分布函数(Bidirectional ReflectanceDistribution Function,BRDF)用来定义给定入射方向上的辐射照度(irradiance)如何影响给定出射方向上的辐射率(radiance)。更笼统地说,它描述了入射光线经过某个表面反射后如何在各个出射方向上分布——这可以是从理想镜面反射到漫反射、各向同性(isotropic)或者各向异性(anisotropic)的各种反射。
 
BRDF作为图形学中比较常见的一个知识点,这边暂时不多讲,因为随便拿一本图形学相关的书都可以看到他的身影。这边给出一些参考的链接,大家有需要可以深入了解:
 
1. 如何正确理解 BRDF (双向反射分布函数)? - 计算机 - 知乎
2.图形学理论知识:BRDF 双向反射分布函数
3. An Introduction to BRDF-based Lighting –Nvidia


 



 

二、续Standard Shader中正向基础渲染通道源码分析


 


此部分接上文《【浅墨Unity3D Shader编程】之十 深入理解Unity5中的Standard Shader(二)&屏幕油画特效的实现》的第二部分“Standard Shader中正向基础渲染通道源码分析“。

上文中分析了Standard Shader中正向基础渲染通道的源码,刚好分析完了顶点着色函数vertForwardBase,本文中将对片段着色函数fragForwardBase 进行说明。分析完之后,也就结束了这一系列长得稍微有些离谱的Standard Shader正向基础渲染通道的源码分析。
 

OK,开始吧,先上注释好的片段着色函数fragForwardBase的代码,位于UnityStandardCore.cginc中:

//----------------------------------------【fragForwardBase函数】-------------------------------------------
//  用途:正向渲染基础通道的片段着色函数
//  输入:VertexOutputForwardBase结构体
//  输出:一个half4类型的颜色值
//------------------------------------------------------------------------------------------------------------------
half4 fragForwardBase (VertexOutputForwardBase i) : SV_Target

	//定义并初始化类型为FragmentCommonData的变量s
	FRAGMENT_SETUP(s)
	//若定义了UNITY_OPTIMIZE_TEXCUBELOD,则由输入的顶点参数来设置反射光方向向量
#if UNITY_OPTIMIZE_TEXCUBELOD
	s.reflUVW		= i.reflUVW;
#endif

	//设置主光照
	UnityLight mainLight = MainLight (s.normalWorld);

	//设置阴影的衰减系数
	half atten = SHADOW_ATTENUATION(i);

	//计算全局光照
	half occlusion = Occlusion(i.tex.xy);
	UnityGI gi = FragmentGI (s, occlusion, i.ambientOrLightmapUV, atten, mainLight);

	//加上BRDF-基于物理的光照
	half4 c = UNITY_BRDF_PBS (s.diffColor, s.specColor, s.oneMinusReflectivity, s.oneMinusRoughness, s.normalWorld, -s.eyeVec, gi.light, gi.indirect);
	//加上BRDF-全局光照
	c.rgb += UNITY_BRDF_GI (s.diffColor, s.specColor, s.oneMinusReflectivity, s.oneMinusRoughness, s.normalWorld, -s.eyeVec, occlusion, gi);
	//加上自发光
	c.rgb += Emission(i.tex.xy);

	//设置雾效
	UNITY_APPLY_FOG(i.fogCoord, c.rgb);

	//返回最终的颜色
	return OutputForward (c, s.alpha);
依然是老规矩,把上面代码中新接触到的相关内容进行下分条讲解。





1. FRAGMENT_SETUP(x)宏

FRAGMENT_SETUP(x)宏定义于UnityStandardCore.cginc头文件中,其作用其实就是用FragmentSetup函数初始化括号中的x变量。
#define FRAGMENT_SETUP(x) FragmentCommonData x = \\
	FragmentSetup(i.tex, i.eyeVec, IN_VIEWDIR4PARALLAX(i), i.tangentToWorldAndParallax, IN_WORLDPOS(i));

调用此宏,也就是表示写了如下的代码,定义了一个x变量:
FragmentCommonData x =FragmentSetup(i.tex, i.eyeVec, IN_VIEWDIR4PARALLAX(i), i.tangentToWorldAndParallax, IN_WORLDPOS(i));


其中FragmentSetup函数也定义于UnityStandardCore.cginc头文件中,用于填充一个FragmentCommonData结构体并于返回值中返回,也就是进行片段函数相关参数的初始化,相关代码如下:

//函数FragmentSetup:填充一个FragmentCommonData结构体并于返回值中返回,进行片段函数相关参数的初始化
inline FragmentCommonData FragmentSetup (float4 i_tex, half3 i_eyeVec, half3 i_viewDirForParallax, half4 tangentToWorld[3], half3 i_posWorld)

	i_tex = Parallax(i_tex, i_viewDirForParallax);

	half alpha = Alpha(i_tex.xy);
	#if defined(_ALPHATEST_ON)
		clip (alpha - _Cutoff);
	#endif

	FragmentCommonData o = UNITY_SETUP_BRDF_INPUT (i_tex);
	o.normalWorld = PerPixelWorldNormal(i_tex, tangentToWorld);
	o.eyeVec = NormalizePerPixelNormal(i_eyeVec);
	o.posWorld = i_posWorld;

	// NOTE: shader relies on pre-multiply alpha-blend (_SrcBlend = One, _DstBlend = OneMinusSrcAlpha)
	o.diffColor = PreMultiplyAlpha (o.diffColor, alpha, o.oneMinusReflectivity, /*out*/ o.alpha);
	return o;
其中的FragmentCommonData结构体也是定义于UnityStandardCore.cginc头文件中:
//FragmentCommonData结构体:存放片段着色常用变量
struct FragmentCommonData

	half3 diffColor, specColor;//漫反射颜色;镜面反射颜色
	// Note: oneMinusRoughness & 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, oneMinusRoughness;//1减去反射率;1减去粗糙度
	half3 normalWorld, eyeVec, posWorld;//世界空间中的法线向量坐标;视角向量坐标;在世界坐标中的位置坐标
	half alpha;//透明度

#if UNITY_OPTIMIZE_TEXCUBELOD || UNITY_STANDARD_SIMPLE
	half3 reflUVW;//反射率的UVW
#endif

#if UNITY_STANDARD_SIMPLE
	half3 tangentSpaceNormal;//切线空间中的法线向量
#endif
;



 

2. MainLight函数



MainLight函数定义于UnityStandardCore.cginc头文件中,用途是实例化一个UnityLight结构体对象,并进行相应的填充,其返回值作为主光源。相关代码如下:

//  用途:该函数为主光照函数
//  说明:实例化一个UnityLight结构体对象,并进行相应的填充
/*
//注:UnityLight结构体定义于UnityLightingCommon.cginc文件中,原型如下:
struct UnityLight

half3 color;
half3 dir;
half  ndotl;
;
*/

//------------------------------------【函数3】MainLight函数-----------------------------------------
//  用途:该函数为主光照函数
//  说明:实例化一个UnityLight结构体对象,并进行相应的填充
//---------------------------------------------------------------------------------------------------------
UnityLight MainLight (half3 normalWorld)

	//【1】实例化一个UnityLight的对象
	UnityLight l;

	//【2】填充UnityLight的各个参数
	//若光照贴图选项为关,使用Unity内置变量赋值
	#ifdef LIGHTMAP_OFF
		//获取光源的颜色
		l.color = _LightColor0.rgb; 
		//获取光源的方向
		l.dir = _WorldSpaceLightPos0.xyz;
		//获取法线与光源方向的点乘的积
		l.ndotl = LambertTerm (normalWorld, l.dir);

	//光照贴图选项为开,将各项值设为0
	#else
		l.color = half3(0.f, 0.f, 0.f);
		l.ndotl  = 0.f;
		l.dir = half3(0.f, 0.f, 0.f);
	#endif

	//返回赋值完成的UnityLight结构体对象
	return l;


 

3. SHADOW_ATTENUATION宏



SHADOW_ATTENUATION宏相关的代码位于AutoLight.cginc头文件中,用于实现阴影渲染相关的辅助工作,代码如下:

// ----------------
//  阴影相关工具代码 || Shadow helpers
// ----------------

// ---- 屏幕空间阴影 || Screen space shadows
#if defined (SHADOWS_SCREEN)
……
#define SHADOW_ATTENUATION(a) unitySampleShadow(a._ShadowCoord)
#endif


// ----聚光灯光源阴影 || Spot light shadows
#if defined (SHADOWS_DEPTH) && defined (SPOT)
	#define SHADOW_COORDS(idx1) unityShadowCoord4 _ShadowCoord : TEXCOORD##idx1;
	#define TRANSFER_SHADOW(a) a._ShadowCoord = mul (unity_World2Shadow[0], mul(_Object2World,v.vertex));
	#define SHADOW_ATTENUATION(a) UnitySampleShadowmap(a._ShadowCoord)
#endif


// ----点光源阴影 ||  Point light shadows
#if defined (SHADOWS_CUBE)
	#define SHADOW_COORDS(idx1) unityShadowCoord3 _ShadowCoord : TEXCOORD##idx1;
	#define TRANSFER_SHADOW(a) a._ShadowCoord = mul(_Object2World, v.vertex).xyz - _LightPositionRange.xyz;
	#define SHADOW_ATTENUATION(a) UnitySampleShadowmap(a._ShadowCoord)
#endif

// ---- 关闭阴影 || Shadows off
#if !defined (SHADOWS_SCREEN) && !defined (SHADOWS_DEPTH) && !defined (SHADOWS_CUBE)
	#define SHADOW_COORDS(idx1)
	#define TRANSFER_SHADOW(a)
	#define SHADOW_ATTENUATION(a) 1.0
#endif

可以发现,SHADOW_ATTENUATION(a)宏除了在关闭阴影的状态是等于1以外,其他几种情况都是等价于UnitySampleShadowmap(a._ShadowCoord)函数的调用。而这里的UnitySampleShadowmap函数,定于于UnityShadowLibrary.cginc函数中。实现代码如下。

//------------------------------【UnitySampleShadowmap函数】---------------------------------
// 用途:采样阴影贴图,得到阴影衰减值
// 输入参数: float3型的阴影向量坐标vec
// 返回值:阴影衰减值
//-------------------------------------------------------------------------------------------------------
inline half UnitySampleShadowmap (float3 vec)

	float mydist = length(vec) * _LightPositionRange.w;
	mydist *= 0.97; // bias

	#if defined (SHADOWS_SOFT)
		float z = 1.0/128.0;
		float4 shadowVals;
		shadowVals.x = SampleCubeDistance (vec+float3( z, z, z));
		shadowVals.y = SampleCubeDistance (vec+float3(-z,-z, z));
		shadowVals.z = SampleCubeDistance (vec+float3(-z, z,-z));
		shadowVals.w = SampleCubeDistance (vec+float3( z,-z,-z));
		half4 shadows = (shadowVals < mydist.xxxx) ? _LightShadowData.rrrr : 1.0f;
		return dot(shadows,0.25);
	#else
		float dist = SampleCubeDistance (vec);
		return dist < mydist ? _LightShadowData.r : 1.0;
	#endif



 

4. Occlusion函数


Occlusion函数用于进行全局光照的第一步。其输入参数为一个float2型的纹理坐标,而其half型的返回值将作为FragmentGI函数的一个输入参数。Occlusion函数的原型如下:

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

其中的LerpOneTo函数很简单,用于线性插值,输入两个值b和t,返回1+(b-1)*t,具体定义如下:
half LerpOneTo(half b, half t)

	half oneMinusT = 1 - t;
	return oneMinusT + b * t;



5. UnityGI结构体
 

UnityGI结构体是Unity中存放全局光照光源信息的结构体,定义于UnityLightingCommon.cginc头文件中,如下。

//全局光照结构体
struct UnityGI

	UnityLight light;//定义第一个光源参数结构体,表示第一个光源
	//若定义了DIRLIGHTMAP_SEPARATE(单独的方向光源光照贴图)
	#ifdef DIRLIGHTMAP_SEPARATE
		//若定义了LIGHTMAP_ON(打开光照贴图)
		#ifdef LIGHTMAP_ON
			UnityLight light2;//定义第二个光源参数结构体,表示第二个光源
		#endif
		//若定义了DYNAMICLIGHTMAP_ON(打开动态光照贴图)
		#ifdef DYNAMICLIGHTMAP_ON
			UnityLight light3;//定义第三个光源参数结构体,表示第三个光源
		#endif
	#endif
	UnityIndirect indirect;//Unity中间接光源参数的结构体
;

其中包含了UnityLight结构体和UnityIndirect结构体,其中UnityLight结构体是Unity Shader中最基本的光照结构体,而UnityIndirect是Unity中存放间接光源信息的结构体。它们两者也定义于UnityLightingCommon.cginc头文件中,代码如下。

//Unity中光源参数的结构体
struct UnityLight

	half3 color;//光源颜色
	half3 dir;//光源方向
	half  ndotl; //入射光方向和当前表面法线方向的点积
;

//Unity中间接光源参数的结构体
struct UnityIndirect

	half3 diffuse;//漫反射颜色
	half3 specular;//镜面反射颜色
;




 
6. FragmentGI函数


FragmentGI函数是片段着色部分全局光照的处理函数,定义于UnityStandardCore.cginc头文件中。相关代码如下:

//函数:片段着色部分全局光照的处理函数
inline UnityGI FragmentGI (FragmentCommonData s, half occlusion, half4 i_ambientOrLightmapUV, half atten, UnityLight light, bool reflections)

	//【1】实例化一个UnityGIInput的对象
	UnityGIInput d;
	//【2】填充此UnityGIInput对象的各个值
	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.boxMax[0] = unity_SpecCube0_BoxMax;
	d.boxMin[0] = unity_SpecCube0_BoxMin;
	d.probePosition[0] = unity_SpecCube0_ProbePosition;
	d.probeHDR[0] = unity_SpecCube0_HDR;

	d.boxMax[1] = unity_SpecCube1_BoxMax;
	d.boxMin[1] = unity_SpecCube1_BoxMin;
	d.probePosition[1] = unity_SpecCube1_ProbePosition;
	d.probeHDR[1] = unity_SpecCube1_HDR;

	//【3】根据填充好的UnityGIInput结构体对象,调用一下UnityGlobalIllumination函数
	if(reflections)
	
		Unity_GlossyEnvironmentData g;
		g.roughness		= 1 - s.oneMinusRoughness;
	#if UNITY_OPTIMIZE_TEXCUBELOD || UNITY_STANDARD_SIMPLE
		g.reflUVW 		= s.reflUVW;
	#else
		g.reflUVW		= reflect(s.eyeVec, s.normalWorld);
	#endif

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


inline UnityGI FragmentGI (FragmentCommonData s, half occlusion, half4 i_ambientOrLightmapUV, half atten, UnityLight light)

	return FragmentGI(s, occlusion, i_ambientOrLightmapUV, atten, light, true);


其中的UnityGIInput结构体定义了全局光照所需要的一些函数,定义为如下:

//全局光照的输入参数结构体
struct UnityGIInput 

	UnityLight light; // 像素光源,由引擎准备并传输过来 || pixel light, sent from the engine

	float3 worldPos;//世界空间中的位置坐标
	half3 worldViewDir;//世界空间中的视角方向向量坐标
	half atten;//衰减值
	half3 ambient;//环境光颜色
	half4 lightmapUV; //光照贴图的UV坐标,其中 取.xy = static lightmapUV(静态光照贴图的UV) , .zw = dynamic lightmap UV(动态光照贴图的UV)

	float4 boxMax[2];//box最大值
	float4 boxMin[2];//box最小值
	float4 probePosition[2];//光照探针的位置
	float4 probeHDR[2];//光照探针的高动态范围图像(High-Dynamic Range)
;

UnityGIInput 中还包含了UnityLight结构体,其定义和代码实现上文刚刚已经有提到过。

FragmentGI函数最终利用了UnityGlobalIllumination函数,其定义于UnityGlobalIllumination.cginc头文件中,实现如下。

inline UnityGI UnityGlobalIllumination (UnityGIInput data, half occlusion, half3 normalWorld)

	return UnityGI_Base(data, occlusion, normalWorld);


FragmentGI 函数就是嵌套了一层UnityGI_Base函数,那我们继续溯源,找到UnityGI_Base函数的定义,也是位于UnityGlobalIllumination.cginc头文件中:
//UnityGI_Base函数:Unity的全局光照Base版
inline UnityGI UnityGI_Base(UnityGIInput data, half occlusion, half3 normalWorld)

	//【1】实例化一个UnityGI类型的结构体
	UnityGI o_gi;
	//【2】重置此UnityGI的结构体
	ResetUnityGI(o_gi);

	//【3】开始逐个填充参数
	#if !defined(LIGHTMAP_ON)
		o_gi.light = data.light;
		o_gi.light.color *= data.atten;
	#endif


	#if UNITY_SHOULD_SAMPLE_SH
		#if UNITY_SAMPLE_FULL_SH_PER_PIXEL
			half3 sh = ShadeSH9(half4(normalWorld, 1.0));
		#elif (SHADER_TARGET >= 30) && !UNITY_STANDARD_SIMPLE
			half3 sh = data.ambient + ShadeSH12Order(half4(normalWorld, 1.0));
		#else
			half3 sh = data.ambient;
		#endif

		o_gi.indirect.diffuse = sh;
	#endif


	#if defined(LIGHTMAP_ON)
		// Baked lightmaps
		fixed4 bakedColorTex = UNITY_SAMPLE_TEX2D(unity_Lightmap, data.lightmapUV.xy);
		half3 bakedColor = DecodeLightmap(bakedColorTex);

		#ifdef DIRLIGHTMAP_OFF
			o_gi.indirect.diffuse = bakedColor;

			#ifdef SHADOWS_SCREEN
				o_gi.indirect.diffuse = MixLightmapWithRealtimeAttenuation (o_gi.indirect.diffuse, data.atten, bakedColorTex);
			#endif // SHADOWS_SCREEN

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

			#ifdef SHADOWS_SCREEN
				o_gi.indirect.diffuse = MixLightmapWithRealtimeAttenuation (o_gi.indirect.diffuse, data.atten, bakedColorTex);
			#endif // SHADOWS_SCREEN

		#elif DIRLIGHTMAP_SEPARATE
			// Left halves of both intensity and direction lightmaps store direct light; right halves - indirect.

			// Direct
			fixed4 bakedDirTex = UNITY_SAMPLE_TEX2D_SAMPLER(unity_LightmapInd, unity_Lightmap, data.lightmapUV.xy);
			o_gi.indirect.diffuse = DecodeDirectionalSpecularLightmap (bakedColor, bakedDirTex, normalWorld, false, 0, o_gi.light);

			// Indirect
			half2 uvIndirect = data.lightmapUV.xy + half2(0.5, 0);
			bakedColor = DecodeLightmap(UNITY_SAMPLE_TEX2D(unity_Lightmap, uvIndirect));
			bakedDirTex = UNITY_SAMPLE_TEX2D_SAMPLER(unity_LightmapInd, unity_Lightmap, uvIndirect);
			o_gi.indirect.diffuse += DecodeDirectionalSpecularLightmap (bakedColor, bakedDirTex, normalWorld, false, 0, o_gi.light2);
		#endif
	#endif

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

		#ifdef DIRLIGHTMAP_OFF
			o_gi.indirect.diffuse += realtimeColor;

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

		#elif DIRLIGHTMAP_SEPARATE
			half4 realtimeDirTex = UNITY_SAMPLE_TEX2D_SAMPLER(unity_DynamicDirectionality, unity_DynamicLightmap, data.lightmapUV.zw);
			half4 realtimeNormalTex = UNITY_SAMPLE_TEX2D_SAMPLER(unity_DynamicNormal, unity_DynamicLightmap, data.lightmapUV.zw);
			o_gi.indirect.diffuse += DecodeDirectionalSpecularLightmap (realtimeColor, realtimeDirTex, normalWorld, true, realtimeNormalTex, o_gi.light3);
		#endif
	#endif

	o_gi.indirect.diffuse *= occlusion;

	//【4】返回此UnityGI类型的结构体
	return o_gi;
不难发现,此FragmentGI函数的实现,也就是实例化一个UnityGIInput结构体对象,然后依次填充了此结构体对象的每个各个参数,最后调用一下基于UnityGI_Base函数的UnityGlobalIllumination函数而已。

 



7. UNITY_BRDF_PBS宏


首先,这边有一段宏,定义于UnityPBSLighting.cginc头文件中,根据不同的情况,将UNITY_BRDF_PBS宏定义为不同版本的UNITY_BRDF_PBS宏——是BRDF3_Unity_PBS、BRDF2_Unity_PBS还是BRDF1_Unity_PBS。

//-------------------------------------------------------------------------------------
// 默认使用BRDF || Default BRDF to use:
#if !defined (UNITY_BRDF_PBS) // 允许显式地在自定义着色器中重写BRDF的实现细节 || allow to explicitly override BRDF in custom shader
	//满足着色目标模型的版本小于Shader Model 3.0,或者是PlayStation 2平台
	#if (SHADER_TARGET < 30) || defined(SHADER_API_PSP2)
		// 为小于SM3.0的着色模型回退为低保真度的BRDF版本 || Fallback to low fidelity one for pre-SM3.0
		#define UNITY_BRDF_PBS BRDF3_Unity_PBS
	#elif defined(SHADER_API_MOBILE)
		// 为移动平台简化的BRDF版本 || Somewhat simplified for mobile
		#define UNITY_BRDF_PBS BRDF2_Unity_PBS
	#else
		//最高特效的SM3、PC平台或者游戏主机平台的BRDF版本 || Full quality for SM3+ PC / consoles
		#define UNITY_BRDF_PBS BRDF1_Unity_PBS
	#endif
#endif
三种情况下,BRDF3_Unity_PBS、BRDF2_Unity_PBS、 BRDF1_Unity_PBS三个函数的参数和返回值都一样,区别仅仅是内部的实现。在这边,以BRDF1_Unity_PBS为例,讲一下参数值的含义。
 
half4 BRDF1_Unity_PBS (half3 diffColor,half3 specColor, half oneMinusReflectivity, half oneMinusRoughness,half3 normal,half3 viewDir,UnityLight light, UnityIndirect gi)
 
第一个参数,half3型的diffColor,表示漫反射颜色的值。
第二个参数,half3型的specColor,表示镜面反射颜色值。
第三个参数,half型的oneMinusReflectivity,表示1减去反射率的值。
第四个参数,half型的oneMinusRoughness,表示1减去粗糙度的值。
第五次参数,half3型的normal,表示法线的方向。
第六个参数,half3型的viewDir,表示视线的方向。

第七个参数,UnityLight型的light,表示Unity中光源参数的结构体,包含half3型的光源颜色color,half3型的光源方向dir,half型的入射光方向和当前表面法线方向的点乘的积ndotl。上文有贴出过其实现代码,都几次提到了,这边就再贴一下。

struct UnityLight

	half3 color;//光源颜色
	half3 dir;//光源方向
	half  ndotl; //入射光方向和当前表面法线方向的点积
;
第八个参数,UnityIndirect类型的gi ,一个包含了half3型的漫反射颜色diffuse和half3型的镜面反射颜色specular的光线反射结构体, 表示间接光照信息
struct UnityIndirect

	half3 diffuse;//漫反射颜色
	half3 specular;//镜面反射颜色
;

下面将三种版本的函数分别贴出来,它们都定义于UnityStandardBRDF.cginc头文件中。


 
7.1 BRDF1_Unity_PBS


//最高特效的SM3、PC平台或者游戏主机平台的BRDF版本 || Full quality for SM3+ PC / consoles
//-------------------------------------------------------------------------------------

// Note: BRDF entry points use oneMinusRoughness (aka "smoothness") and 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.


// Main Physically Based BRDF
// Derived from Disney work and based on Torrance-Sparrow micro-facet model
//
//   BRDF = kD / pi + kS * (D * V * F) / 4
//   I = BRDF * NdotL
//
// * NDF (depending on UNITY_BRDF_GGX):
//  a) Normalized BlinnPhong
//  b) GGX
// * Smith for Visiblity term
// * Schlick approximation for Fresnel
half4 BRDF1_Unity_PBS (half3 diffColor, half3 specColor, half oneMinusReflectivity, half oneMinusRoughness,
	half3 normal, half3 viewDir,
	UnityLight light, UnityIndirect gi)

	half roughness = 1-oneMinusRoughness;
	half3 halfDir = Unity_SafeNormalize (light.dir + viewDir);

	half nl = light.ndotl;
	half nh = BlinnTerm (normal, halfDir);
	half nv = DotClamped (normal, viewDir);
	half lv = DotClamped (light.dir, viewDir);
	half lh = DotClamped (light.dir, halfDir);

#if UNITY_BRDF_GGX
	half V = SmithGGXVisibilityTerm (nl, nv, roughness);
	half D = GGXTerm (nh, roughness);
#else
	half V = SmithBeckmannVisibilityTerm (nl, nv, roughness);
	half D = NDFBlinnPhongNormalizedTerm (nh, RoughnessToSpecPower (roughness));
#endif

	half nlPow5 = Pow5 (1-nl);
	half nvPow5 = Pow5 (1-nv);
	half Fd90 = 0.5 + 2 * lh * lh * roughness;
	half disneyDiffuse = (1 + (Fd90-1) * nlPow5) * (1 + (Fd90-1) * nvPow5);
	
	// HACK: theoretically we should divide by Pi diffuseTerm 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 to in cases when they are injected into ambient SH
	// NOTE: multiplication by Pi is part of single constant together with 1/4 now

	half specularTerm = max(0, (V * D * nl) * unity_LightGammaCorrectionConsts_PIDiv4);// Torrance-Sparrow model, Fresnel is applied later (for optimization reasons)
	half diffuseTerm = disneyDiffuse * nl;
	
	half grazingTerm = saturate(oneMinusRoughness + (1-oneMinusReflectivity));
    half3 color =	diffColor * (gi.diffuse + light.color * diffuseTerm)
                    + specularTerm * light.color * FresnelTerm (specColor, lh)
					+ gi.specular * FresnelLerp (specColor, grazingTerm, nv);

	return half4(color, 1);


7.2 BRDF2_Unity_PBS
12

// 为移动平台简化的BRDF版本 || Somewhat simplified for mobile
// Based on Minimalist CookTorrance BRDF
// Implementation is slightly different from original derivation: http://www.thetenthplanet.de/archives/255
//
// * BlinnPhong as NDF
// * Modified Kelemen and Szirmay-Kalos for Visibility term
// * Fresnel approximated with 1/LdotH
half4 BRDF2_Unity_PBS (half3 diffColor, half3 specColor, half oneMinusReflectivity, half oneMinusRoughness,
	half3 normal, half3 viewDir,
	UnityLight light, UnityIndirect gi)

	half3 halfDir = Unity_SafeNormalize (light.dir + viewDir);

	half nl = light.ndotl;
	half nh = BlinnTerm (normal, halfDir);
	half nv = DotClamped (normal, viewDir);
	half lh = DotClamped (light.dir, halfDir);

	half roughness = 1-oneMinusRoughness;
	half specularPower = RoughnessToSpecPower (roughness);
	// Modified with approximate Visibility function that takes roughness into account
	// Original ((n+1)*N.H^n) / (8*Pi * L.H^3) didn't take into account roughness 
	// and produced extremely bright specular at grazing angles

	// HACK: theoretically we should divide by Pi diffuseTerm 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 to in cases when they are injected into ambient SH
	// NOTE: multiplication by Pi is cancelled with Pi in denominator

	half invV = lh * lh * oneMinusRoughness + roughness * roughness; // approx ModifiedKelemenVisibilityTerm(lh, 1-oneMinusRoughness);
	half invF = lh;
	half specular = ((specularPower + 1) * pow (nh, specularPower)) / (unity_LightGammaCorrectionConsts_8 * invV * invF + 1e-4h); // @TODO: might still need saturate(nl*specular) on Adreno/Mali

	// Prevent FP16 overflow on mobiles
#if SHADER_API_GLES || SHADER_API_GLES3
	specular = clamp(specular, 0.0, 100.0);
#endif

	half grazingTerm = saturate(oneMinusRoughness + (1-oneMinusReflectivity));
    half3 color =	(diffColor + specular * specColor) * light.color * nl
    				+ gi.diffuse * diffColor
					+ gi.specular * FresnelLerpFast (specColor, grazingTerm, nv);

	return half4(color, 1);


 
7.3 BRDF3_Unity_PBS

// 为小于SM3.0的着色模型回退为低保真度的BRDF版本 || Fallback to low fidelity one for pre-SM3.0
// Old school, not microfacet based Modified Normalized Blinn-Phong BRDF
// Implementation uses Lookup texture for performance
//
// * Normalized BlinnPhong in RDF form
// * Implicit Visibility term
// * No Fresnel term
//
// TODO: specular is too weak in Linear rendering mode
half4 BRDF3_Unity_PBS (half3 diffColor, half3 specColor, half oneMinusReflectivity, half oneMinusRoughness,
	half3 normal, half3 viewDir,
	UnityLight light, UnityIndirect gi)

	half3 reflDir = reflect (viewDir, normal);

	half nl = light.ndotl;
	half nv = DotClamped (normal, viewDir);

	// Vectorize Pow4 to save instructions
	half2 rlPow4AndFresnelTerm = Pow4 (half2(dot(reflDir, light.dir), 1-nv));  // use R.L instead of N.H to save couple of instructions
	half rlPow4 = rlPow4AndFresnelTerm.x; // power exponent must match kHorizontalWarpExp in NHxRoughness() function in GeneratedTextures.cpp
	half fresnelTerm = rlPow4AndFresnelTerm.y;

	half grazingTerm = saturate(oneMinusRoughness + (1-oneMinusReflectivity));

	half3 color = BRDF3_Direct(diffColor, specColor, rlPow4, oneMinusRoughness);
	color *= light.color * nl;
	color += BRDF3_Indirect(diffColor, specColor, gi, grazingTerm, fresnelTerm);

	return half4(color, 1);


BRDF1_Unity_PBS函数的实现部分用到了最多的变量,最终表现效果最好,主要用于Shader Model 3.0、PC平台或者游戏主机平台。BRDF2_Unity_PBS简化了一部分计算,主要用于移动平台,而BRDF3_Unity_PBS是为Shader Model 小于3.0的着色模型提供基本版的BRDF,实现细节最为简陋。


 



8. UNITY_BRDF_GI宏



UNITY_BRDF_GI宏位于UnityPBSLighting.cginc头文件中,相关代码如下。

//-------------------------------------------------------------------------------------
// 从间接的方向光照贴图中进行BRDF(双向反射分布函数)的光照提取 || BRDF for lights extracted from *indirect* directional lightmaps (baked and realtime).
// 使用UNITY_BRDF_PBS从方向光源烘焙方向光照贴图, || Baked directional lightmap with *direct* light uses UNITY_BRDF_PBS.
// 若想得到更好的效果,可以使用BRDF1_Unity_PBS || For better quality change to BRDF1_Unity_PBS.
// SM2.0中的非方向光照贴图|| No directional lightmaps in SM2.0.

//若没有定义UNITY_BRDF_PBS_LIGHTMAP_INDIRECT宏
#if !defined(UNITY_BRDF_PBS_LIGHTMAP_INDIRECT)
	//定义UNITY_BRDF_PBS_LIGHTMAP_INDIRECT = BRDF2_Unity_PBS
	#define UNITY_BRDF_PBS_LIGHTMAP_INDIRECT BRDF2_Unity_PBS
#endif
//若没有定义UNITY_BRDF_GI宏
#if !defined (UNITY_BRDF_GI)
	//定义UNITY_BRDF_GI = BRDF_Unity_Indirect
	#define UNITY_BRDF_GI BRDF_Unity_Indirect
#endif

上面这段代码中关于UNITY_BRDF_GI宏的地方,就是说若没有定义UNITY_BRDF_GI宏,就定义一个UNITY_BRDF_GI宏等价于BRDF_Unity_Indirect。这边的BRDF_Unity_Indirect是一个函数名,就紧紧跟在上面这段宏代码的后面:
 
//间接光照的BRDF
inline half3 BRDF_Unity_Indirect (half3 baseColor, half3 specColor, half oneMinusReflectivity, half oneMinusRoughness, half3 normal, half3 viewDir, half occlusion, UnityGI gi)

	half3 c = 0;
	#if defined(DIRLIGHTMAP_SEPARATE)
		gi.indirect.diffuse = 0;
		gi.indirect.specular = 0;

		#ifdef LIGHTMAP_ON
			c += UNITY_BRDF_PBS_LIGHTMAP_INDIRECT (baseColor, specColor, oneMinusReflectivity, oneMinusRoughness, normal, viewDir, gi.light2, gi.indirect).rgb * occlusion;
		#endif
		#ifdef DYNAMICLIGHTMAP_ON
			c += UNITY_BRDF_PBS_LIGHTMAP_INDIRECT (baseColor, specColor, oneMinusReflectivity, oneMinusRoughness, normal, viewDir, gi.light3, gi.indirect).rgb * occlusion;
		#endif
	#endif
	return c;


关于此段代码,BRDF_Unity_Indirect 函数的核心部分其实就是在调用UNITY_BRDF_PBS_LIGHTMAP_INDIRECT,而上文的宏有交代过,UNITY_BRDF_PBS_LIGHTMAP_INDIRECT宏等价于 BRDF2_Unity_PBS。而BRDF2_Unity_PBS函数,其定义于UnityStandardBRDF.cginc中,是为移动平台简化的BRDF版本,这个上文刚刚提到过,这边就不多交代。
 

 


 
9.Emission函数


Emission函数定于于UnityStandardInput.cginc头文件中,根据指定的自发光光照贴图,利用tex2D函数,对输入的纹理进行光照贴图的采样,相关代码如下:

//---------------------------------------【Emission函数】-----------------------------------------
// 用途:根据指定的自发光光照贴图,利用tex2D函数,对输入的纹理进行光照贴图的采样
// 输入参数:float2型的纹理坐标
// 输出参数:经过将自发光纹理和输入纹理进行tex2D采样得到的half3型的自发光颜色
//-----------------------------------------------------------------------------------------------
half3 Emission(float2 uv)

#ifndef _EMISSION
	return 0;
#else
	return tex2D(_EmissionMap, uv).rgb * _EmissionColor.rgb;
#endif

其中用于采样的自发光贴图对应的函数定义于UnityStandardInput.cginc头文件的一开始部分。
sampler2D	_EmissionMap;

这边这句代码其实是相当于在CGPROGRAM中的顶点和片段着色函数之前,对这个变量进行声明,以便于CG语言块中使用的时候,能识别到他的含义。 因为在Standard.shader源码的一开始,Properties块也就是属性值声明部分,对其进行了属性的声明:
//自发光纹理图
_EmissionMap("Emission", 2D) = "white" 


 
 
10.UNITY_APPLY_FOG宏
 


UNITY_APPLY_FOG宏相关的一些代码用于雾效的启用与否的辅助工作,定义于UnityCG.cginc头文件中,这边贴出注释好的代码即可。

//UNITY_FOG_LERP_COLOR宏的定义
#define UNITY_FOG_LERP_COLOR(col,fogCol,fogFac) col.rgb = lerp((fogCol).rgb, (col).rgb, saturate(fogFac))

//【1】若已经定义了FOG_LINEAR、FOG_EXP、FOG_EXP2宏三者至少之一,便可以进行到此#if实现部分
#if defined(FOG_LINEAR) || defined(FOG_EXP) || defined(FOG_EXP2)
	//【1-1】若满足着色目标模型的版本小于Shader Model 3.0,或者定义了SHADER_API_MOBILE宏,便可以进行到此#if实现部分
	#if (SHADER_TARGET < 30) || defined(SHADER_API_MOBILE)
		//移动平台和Shader Model 2.0:已经计算了每顶点的雾效因子,所以插一下值就可以了 ||mobile or SM2.0: fog factor was already calculated per-vertex, so just lerp the color
		//定义 UNITY_APPLY_FOG_COLOR(coord,col,fogCol) 等价于UNITY_FOG_LERP_COLOR(col,fogCol,coord)
	#define UNITY_APPLY_FOG_COLOR(coord,col,fogCol) UNITY_FOG_LERP_COLOR(col,fogCol,coord)

	//【1-2】 Shader Model 3.0和PC/游戏主机平台:计算雾效因子以及进行雾颜色的插值 ||SM3.0 and PC/console: calculate fog factor and lerp fog color
	#else
		//定义 UNITY_APPLY_FOG_COLOR(coord,col,fogCol)等价于UNITY_CALC_FOG_FACTOR(coord); UNITY_FOG_LERP_COLOR(col,fogCol,unityFogFactor)
		#define UNITY_APPLY_FOG_COLOR(coord,col,fogCol) UNITY_CALC_FOG_FACTOR(coord); UNITY_FOG_LERP_COLOR(col,fogCol,unityFogFactor)
	#endif
//【2】否则,直接定义UNITY_APPLY_FOG_COLOR宏
#else
	#define UNITY_APPLY_FOG_COLOR(coord,col,fogCol)
#endif
//【3】若定义了UNITY_PASS_FORWARDADD(正向附加渲染通道)宏
#ifdef UNITY_PASS_FORWARDADD
	//定义UNITY_APPLY_FOG(coord,col) 等价于UNITY_APPLY_FOG_COLOR(coord,col,fixed4(0,0,0,0))
	#define UNITY_APPLY_FOG(coord,col) UNITY_APPLY_FOG_COLOR(coord,col,fixed4(0,0,0,0))
//【4】否则,UNITY_APPLY_FOG(coord,col) 等价于 UNITY_APPLY_FOG_COLOR(coord,col,unity_FogColor)
#else
	#define UNITY_APPLY_FOG(coord,col) UNITY_APPLY_FOG_COLOR(coord,col,unity_FogColor)
#endif



 
11.OutputForward函数

OutputForward函数定义于UnityStandardCore.cginc头文件中,其为正向渲染通道的输出函数。
//-----------------------------【函数OutputForward】----------------------------------------------
// 用途:正向渲染通道输出函数
//	输入参数:一个half4类型的一个颜色值output,一个half型的透明度值alphaFromSurface
// 返回值:经过透明处理的half4型的输出颜色值
//-------------------------------------------------------------------------------------------------
half4 OutputForward (half4 output, half alphaFromSurface)

	#if defined(_ALPHABLEND_ON) || defined(_ALPHAPREMULTIPLY_ON)
		output.a = alphaFromSurface;
	#else
		UNITY_OPAQUE_ALPHA(output.a);
	#endif
	return output;

其中UNITY_OPAQUE_ALPHA宏的定义为:

#define UNITY_OPAQUE_ALPHA(outputAlpha) outputAlpha = 1.0



 
 
三、屏幕像素化特效的实现
 
 

我们都知道,Unity中的屏幕特效通常分为两部分来实现:

  • Shader实现部分
  • 脚本实现部分
下面依旧是从这两个方面对本次的特效进行实现。
 
 
3.1 Shader实现部分
 
 
国际惯例,上注释好的Shader代码。
Shader "浅墨Shader编程/Volume11/PixelEffect"

	//------------------------------------【属性值】------------------------------------
	Properties
	
	//主纹理
	_MainTex("Texture", 2D) = "white" 
	//封装的变量值
	_Params("PixelNumPerRow (X) Ratio (Y)", Vector) = (80, 1, 1, 1.5)


	//------------------------------------【唯一的子着色器】------------------------------------
	SubShader
	
		//关闭剔除操作
		Cull Off
		//关闭深度写入模式
		ZWrite Off
		//设置深度测试模式:渲染所有像素.等同于关闭透明度测试(AlphaTest Off)
		ZTest Always

		//--------------------------------唯一的通道-------------------------------
		Pass
		
			//===========开启CG着色器语言编写模块===========
			CGPROGRAM

			//编译指令:告知编译器顶点和片段着色函数的名称
			#pragma vertex vert
			#pragma fragment frag

			//包含头文件
			#include "UnityCG.cginc"

			//顶点着色器输入结构
			struct vertexInput
			
				float4 vertex : POSITION;//顶点位置
				float2 uv : TEXCOORD0;//一级纹理坐标
			;

			//顶点着色器输出结构
			struct vertexOutput
			
				float4 vertex : SV_POSITION;//像素位置
				float2 uv : TEXCOORD0;//一级纹理坐标
			;

			//--------------------------------【顶点着色函数】-----------------------------
			// 输入:顶点输入结构体
			// 输出:顶点输出结构体
			//---------------------------------------------------------------------------------
			//顶点着色函数
			vertexOutput vert(vertexInput   v)
			
				//【1】实例化一个输入结构体
				vertexOutput o;
				//【2】填充此输出结构
				//输出的顶点位置(像素位置)为模型视图投影矩阵乘以顶点位置,也就是将三维空间中的坐标投影到了二维窗口
				o.vertex = mul(UNITY_MATRIX_MVP, v.vertex);
				//输入的UV纹理坐标为顶点输出的坐标
				o.uv = v.uv;

				//【3】返回此输出结构对象
				return o;
			

			//变量的声明
			sampler2D _MainTex;
			half4 _Params;

			//进行像素化操作的自定义函数PixelateOperation
			half4 PixelateOperation(sampler2D tex, half2 uv, half scale, half ratio)
			
				//【1】计算每个像素块的尺寸
				half PixelSize = 1.0 / scale;
				//【2】取整计算每个像素块的坐标值,ceil函数,对输入参数向上取整
				half coordX=PixelSize * ceil(uv.x / PixelSize);
				half coordY = (ratio * PixelSize)* ceil(uv.y / PixelSize / ratio);
				//【3】组合坐标值
				half2 coord = half2(coordX,coordY);
				//【4】返回坐标值
				return half4(tex2D(tex,

以上是关于Unity3D Shader编程之十一 深入理解Unity5中的Standard Shader&屏幕像素化特效的实现的主要内容,如果未能解决你的问题,请参考以下文章

浅墨Unity3D Shader编程之十三 单色透明Shader & 标准镜面高光Shader

Unity Shader编程之十四 边缘发光Shader(Rim Shader)的两种实现形态

深入理解Spring系列之十一:SpringMVC-@RequestBody接收json数据报415

unity shader 基础之十一 动画屏幕后处理深度纹理

转:Java并发编程之十一:线程间通信中notify通知的遗漏(含代码)

浅墨Unity3D Shader编程之六 暗黑城堡篇: 表面着色器(Surface Shader)的写法