FrameworkDocs

Developer documentation for "Zenteon: Framework" ReShade reconstruction framework.

NOTE: This page is a work in process

While the included .fhx header file(s) and assets are under the FRL license, the original shader still falls under the restrictions of the AGNYA License: https://github.com/nvb-uy/AGNYA-License. A "Hello World" file is included to verify everything is working properly.

Dependencies: Framework is currently dependent on the standard "ReShade.fxh", this dependency is planned to be removed in the future.

v0.2

Buffers

In addition to the standard COLOR and DEPTH Buffers provided by the ReShade.fxh, framework generates a few additional buffers for shader usage.

Full Resolution Buffers:

1. zfw::tNormal - Octahedral encoded normal buffer, optionally processed with smoothing and texturing via Framework settings.
2. zfw::tAlbedo - Estimated albedo buffer.
3. zfw::tRoughness - Estimated roughness buffer.
4. zfw::tVelocity - 1:1 Motion vectors, with dissoclusion stored in the z channel.

Reduced Resolution Buffers:

1. zfw::tLowNormal - A quarter resolution octahedral encoded normal buffer, meant to save performance when distant sampling is required.
2. zfw::tLowDepth - A quarter resolution depth buffer, meant to save performance when distant sampling is required. Uses the max depth in each 4x4 block to prevent halos.

Functions

In order to streamline development and allow ease of access to the generated buffers, Framework provides some additional functions.

Projections
float zfw::getDepth(float2 uv) Returns the linearized depth value, identical to ReShade::GetLinearizedDepth, just faster to type.
float3 zfw::uvzToView(float3 uvz) Projects uvz coordinates to viewspace.
float3 zfw::uvToView(float2 uv) Projects uv coordinates to viewspace, automatically grabbing depth from the specified uv coordinates.
float3 zfw::viewToUv(float3 xyz) Projects viewspace coordinates to uvz space.

Sampling Functions
float3 zfw::getNormal(float2 uv) Samples and decodes the full resolution normal value at the specified uv coordinates.
float zfw::getRoughness(float2 uv) Returns the roughness buffer value at the specified uv coordinates.
float3 zfw::getVelocity(float2 uv) Returns the velocity buffer value at the specified uv coordinates. Sample confidence (1.0 for valid samples, 0.0 for invalid) is stored in the z channel for sample rejection.
float3 zfw::getBackBuffer(float2 uv) Returns the rgb values of the COLOR buffer at the specified uv coordinates.
float3 zfw::getAlbedo(float2 uv) Returns the estimated albedo value at the specified uv coordinates.

float3 zfw::sampleNormal(float2 uv, float mip) Samples and decodes the quarter resolution normal value at the specified uv coordinates and miplevel.
float zfw::sampleDepth(float2 uv, float mip) Samples the quarter resolution depth value at the specified uv coordinates and miplevel. Note that this currently produces averaged mips past the first level, exercise caution to avoid haloing.

Tonemapping
The FrameworkResources.fxh also includes a luminance preserving tonemapper and inverse tonemapper that prevents oversaturation significantly. Note that these are meant to be used in tandem: using different operators for tonemapping and inverting is not recommended.
float3 zfw::toneMap(float3 x, float whitepoint) Tonemaps the input color with the specified whitepoint.
float3 toneMapInverse(float3 x, float whitepoint) Inverse tonemaps the input color with the specified whitepoint.

Internal functions
Framework also has a few functions used internally, that while they aren't specifically intended for the user, may still be of use.
float3 zfw::UVtoOCT(float2 uv) Converts from uv to octahedral coordinates.
float2 zfw::OCTtoUV(float3 xyz) Converts from octahedral coordinated to uv.
float2 zfw::NormalEncode(float3 n) Compresses normals to octahedral.
float3 zfw::NormalDecode(float2 n) Decompresses octahedral normals.
float zfw::getLuminance(float3 x) Returns the luminance of the input (linear) color.