[HDRP] Add custom pass buffer scaling functions (#5809)

* Add functions to sample/load custom buffer for custom pass

* Updated changelog

* Updated changelog

* Fix scaling issues with dynamic res

* Update CHANGELOG.md

* Fix custom pass sampling functions

* Fix after post scaling

* Update CHANGELOG.md

Co-authored-by: sebastienlagarde <sebastien@unity3d.com>

Author: alelievr

com.unity.render-pipelines.high-definition/CHANGELOG.md

@@ -9,11 +9,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 ### Added
 - Added support for orthographic camera in path tracing.
 - Added public API to edit materials from script at runtime.
-
-### Changed
-- Use RayTracingAccelerationStructure.CullInstances to filter Renderers and populate the acceleration structure with ray tracing instances for improved CPU performance on the main thread.
-- Changed the max distance for Light Anchors to avoid unstability with high values (case 1362802).
-- PrepareLightsForGPU CPU Light loop performance improvement (40% to 70% faster), utilizing burst and optimized. Utilizing better sorting, distributing work in jobs and improving cache access of light data.
+- Added new functions that sample the custom buffer in custom passes (CustomPassSampleCustomColor and CustomPassLoadCustomColor) to handle the RTHandleScale automatically.
 
 ### Fixed
 - Fixed decal position when created from context menu. (case 1368987)
@@ -27,13 +23,18 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 - Fixed depth pyramid being incorrect when having multiple cameras (scene view and gameview) and when hardware DRS was activated.
 - Fixed the cloudlayer not using depth buffer.
 - Fixed crossfade not working on the HD ST8 ShaderGraph [case 1369586](https://fogbugz.unity3d.com/f/cases/1369586/)
-
 - Fixed range compression factor being clamped. (case 1365707)
 - Fixed tooltip not showing on labels in ShaderGraphs (1358483).
 - Fix API warnings in Matcap mode on Metal.
 - Fix D3D validation layer errors w.r.t shadow textures when an atlas is not used.
 - Fixed anchor position offset property for the Light Anchor component. (case 1362809)
 - Fixed minor performance issues in SSGI (case 1367144).
+- Fixed scaling issues with dynamic resolution and the CustomPassSampleCameraColor function.
+
+### Changed
+- Use RayTracingAccelerationStructure.CullInstances to filter Renderers and populate the acceleration structure with ray tracing instances for improved CPU performance on the main thread.
+- Changed the max distance for Light Anchors to avoid unstability with high values (case 1362802).
+- PrepareLightsForGPU CPU Light loop performance improvement (40% to 70% faster), utilizing burst and optimized. Utilizing better sorting, distributing work in jobs and improving cache access of light data.
 
 ## [13.1.0] - 2021-09-24
 

com.unity.render-pipelines.high-definition/Documentation~/Custom-Pass-Creating.md

@@ -463,6 +463,11 @@ You can also load custom buffers using the following functions:
 - `LoadCustomColor(uint2 pixelCoords)`
 - `LoadCustomDepth(uint2 pixelCoords)`
 
+Note that depending on the injection point used for the Fullscreen custom pass, sampling the custom buffer can result in incorrect scaling. Thus it's recommended to use these functions instead:
+
+- `CustomPassSampleCustomColor(float2 uv)`
+- `CustomPassLoadCustomColor(uint2 pixelCoords)`
+
 HDRP sets the custom pass target buffers to the Camera buffers by default. However, you can select a custom buffer in the UI of the Custom Pass. To do this, go to your Custom Pass component and change the **Target Color Buffer** or **Target Depth Buffer** properties.
 
 To change the buffer format of the Custom Pass component in your HDRP asset, go to **Rendering > Custom Pass > Custom Buffer Format** and select one of the following formats from the drop down menu:

com.unity.render-pipelines.high-definition/Editor/RenderPipeline/CustomPass/CustomPassFullScreenShader.template

@@ -23,8 +23,8 @@ Shader "FullScreen/#SCRIPTNAME#"
     // To sample custom buffers, you have access to these functions:
     // But be careful, on most platforms you can't sample to the bound color buffer. It means that you
     // can't use the SampleCustomColor when the pass color buffer is set to custom (and same for camera the buffer).
-    // float4 SampleCustomColor(float2 uv);
-    // float4 LoadCustomColor(uint2 pixelCoords);
+    // float4 CustomPassSampleCustomColor(float2 uv);
+    // float4 CustomPassLoadCustomColor(uint2 pixelCoords);
     // float LoadCustomDepth(uint2 pixelCoords);
     // float SampleCustomDepth(float2 uv);
 

com.unity.render-pipelines.high-definition/Runtime/RenderPipeline/RenderPass/CustomPass/CustomPassCommon.hlsl

@@ -24,9 +24,10 @@ float3 CustomPassSampleCameraColor(float2 uv, float lod, bool uvGuards = true)
         case CUSTOMPASSINJECTIONPOINT_BEFORE_RENDERING: return float3(0, 0, 0);
         // there is no color pyramid yet for before transparent so we can't sample with mips.
         // Also, we don't use _RTHandleScaleHistory to sample because the color pyramid bound is the actual camera color buffer which is at the resolution of the camera
-        case CUSTOMPASSINJECTIONPOINT_BEFORE_TRANSPARENT:
-        case CUSTOMPASSINJECTIONPOINT_BEFORE_PRE_REFRACTION: return SAMPLE_TEXTURE2D_X_LOD(_ColorPyramidTexture, s_trilinear_clamp_sampler, uv * _RTHandleScaleHistory.xy, 0).rgb;
-        case CUSTOMPASSINJECTIONPOINT_AFTER_POST_PROCESS: return SAMPLE_TEXTURE2D_X_LOD(_AfterPostProcessColorBuffer, s_trilinear_clamp_sampler, uv * _RTHandleScaleHistory.xy, 0).rgb;
+        case CUSTOMPASSINJECTIONPOINT_BEFORE_TRANSPARENT: return SAMPLE_TEXTURE2D_X_LOD(_ColorPyramidTexture, s_trilinear_clamp_sampler, uv * _RTHandleScaleHistory.xy, 0).rgb;
+        case CUSTOMPASSINJECTIONPOINT_BEFORE_POST_PROCESS:
+        case CUSTOMPASSINJECTIONPOINT_BEFORE_PRE_REFRACTION: return SAMPLE_TEXTURE2D_X_LOD(_ColorPyramidTexture, s_trilinear_clamp_sampler, uv * _RTHandleScale.xy, 0).rgb;
+        case CUSTOMPASSINJECTIONPOINT_AFTER_POST_PROCESS: return SAMPLE_TEXTURE2D_X_LOD(_AfterPostProcessColorBuffer, s_trilinear_clamp_sampler, uv * _RTHandleScale.zw, 0).rgb;
         default: return SampleCameraColor(uv, lod);
     }
 }
@@ -44,6 +45,20 @@ float3 CustomPassLoadCameraColor(uint2 pixelCoords, float lod)
     }
 }
 
+float4 CustomPassSampleCustomColor(float2 uv)
+{
+    switch ((int)_CustomPassInjectionPoint)
+    {
+        case CUSTOMPASSINJECTIONPOINT_AFTER_POST_PROCESS: return LOAD_TEXTURE2D_X_LOD(_CustomColorTexture, uv * _ScreenSize.xy, 0);
+        default: return SampleCustomColor(uv);
+    }
+}
+
+float4 CustomPassLoadCustomColor(uint2 pixelCoords)
+{
+    return LoadCustomColor(pixelCoords);
+}
+
 struct Attributes
 {
     uint vertexID : SV_VertexID;