mirror of
https://github.com/hrydgard/ppsspp.git
synced 2026-10-01 14:58:14 +00:00
Create fsr_rcas.fsh
Add AMD FSR1
This commit is contained in:
1 parent
2e07d5d3f6
commit
0ff1f6483c
1 file changed
+117
@@ -0,0 +1,117 @@
|
||||
// MIT License
|
||||
//
|
||||
// Copyright (c) 2021 Advanced Micro Devices, Inc. All rights reserved.
|
||||
//
|
||||
// Permission is hereby granted, free of charge, to any person obtaining a copy
|
||||
// of this software and associated documentation files (the "Software"), to deal
|
||||
// in the Software without restriction, including without limitation the rights
|
||||
// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
||||
// copies of the Software, and to permit persons to whom the Software is
|
||||
// furnished to do so, subject to the following conditions:
|
||||
//
|
||||
// The above copyright notice and this permission notice shall be included in all
|
||||
// copies or substantial portions of the Software.
|
||||
//
|
||||
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
||||
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
||||
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
||||
// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
||||
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
||||
// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
||||
// SOFTWARE.
|
||||
|
||||
// Set precision for OpenGL ES
|
||||
#ifdef GL_ES
|
||||
precision mediump float;
|
||||
precision mediump int;
|
||||
#endif
|
||||
|
||||
// Texture sampler for input image
|
||||
uniform sampler2D sampler0;
|
||||
|
||||
// Texture coordinate delta (1.0 / viewport size)
|
||||
uniform vec2 u_texelDelta;
|
||||
|
||||
// User settings:
|
||||
// u_setting.x = sharpness value (0.0 to 1.0)
|
||||
uniform vec4 u_setting;
|
||||
|
||||
// Interpolated texture coordinates from vertex shader
|
||||
varying vec2 v_texcoord0;
|
||||
|
||||
// Split-screen divider width for comparison views
|
||||
const float lineWidth = 0.005;
|
||||
|
||||
// Luminance coefficients based on Rec. 709 standard
|
||||
// Used for converting RGB to perceptually weighted luminance
|
||||
const vec3 lumCoef = vec3(0.2126, 0.7152, 0.0722);
|
||||
|
||||
// RCAS (Robust Contrast Adaptive Sharpening) constants
|
||||
// Peak negative lobe strength and its inverse
|
||||
const float rcasPeak = 8.0 - 3.0; // Peak negative lobe strength
|
||||
const float rcasInvPeak = 1.0 / rcasPeak; // Inverse of peak strength
|
||||
|
||||
// Small epsilon value for numerical stability
|
||||
// Matches AMD's reference implementation
|
||||
const float FSR_EPS = 0.0001;
|
||||
|
||||
// Cross-shaped sampling pattern offsets (N, W, E, S)
|
||||
// Used for 4-tap cross sampling around center pixel
|
||||
const vec2 crossOffsets[4] = vec2[4](
|
||||
vec2( 0.0, -1.0), // North
|
||||
vec2(-1.0, 0.0), // West
|
||||
vec2( 1.0, 0.0), // East
|
||||
vec2( 0.0, 1.0) // South
|
||||
);
|
||||
|
||||
// Vanilla RCAS kernel implementation (no edge-aware weighting)
|
||||
// Performs contrast adaptive sharpening on the input texture
|
||||
vec4 FsrRcasVanilla(vec2 uv) {
|
||||
// Sample center pixel and convert to luminance
|
||||
vec3 C = texture2D(sampler0, uv).rgb;
|
||||
float CL = dot(C, lumCoef);
|
||||
|
||||
// Sample the 4 cross neighbors and convert to luminance
|
||||
vec3 N = texture2D(sampler0, uv + crossOffsets[0] * u_texelDelta).rgb; // North
|
||||
vec3 W = texture2D(sampler0, uv + crossOffsets[1] * u_texelDelta).rgb; // West
|
||||
vec3 E = texture2D(sampler0, uv + crossOffsets[2] * u_texelDelta).rgb; // East
|
||||
vec3 S = texture2D(sampler0, uv + crossOffsets[3] * u_texelDelta).rgb; // South
|
||||
|
||||
float NL = dot(N, lumCoef); // North luminance
|
||||
float WL = dot(W, lumCoef); // West luminance
|
||||
float EL = dot(E, lumCoef); // East luminance
|
||||
float SL = dot(S, lumCoef); // South luminance
|
||||
|
||||
// Calculate adaptive amplification factor to prevent oversharpening
|
||||
// Uses min/max range analysis to determine safe sharpening strength
|
||||
vec3 minRGB = min(min(min(N, W), min(E, S)), C); // Minimum RGB in 5-tap neighborhood
|
||||
vec3 maxRGB = max(max(max(N, W), max(E, S)), C); // Maximum RGB in 5-tap neighborhood
|
||||
vec3 invMax = 1.0 / (maxRGB + FSR_EPS); // Inverse of maximum (with epsilon)
|
||||
vec3 amp = clamp(min(minRGB, 2.0 - maxRGB) * invMax, 0.0, 1.0); // Amplification factor
|
||||
amp = inversesqrt(amp + FSR_EPS); // Inverse square root for non-linearity
|
||||
|
||||
// Calculate sharpening weight based on amplification
|
||||
float w = -rcasInvPeak / dot(amp, lumCoef);
|
||||
|
||||
// Compute sharpened luminance using contrast adaptive formula
|
||||
float sumL = NL + WL + EL + SL; // Sum of neighbor luminances
|
||||
float invDen = 1.0 / (4.0 * w + 1.0); // Inverse denominator
|
||||
float sharpL = clamp((sumL * w + CL) * invDen, 0.0, 1.0); // Sharpened luminance
|
||||
|
||||
// Reconstruct color by preserving chroma (hue/saturation)
|
||||
// This prevents color shifts during sharpening
|
||||
vec3 chroma = C - vec3(CL); // Extract chroma (color without brightness)
|
||||
vec3 sharpColor = chroma + vec3(sharpL); // Apply sharpened luminance to chroma
|
||||
|
||||
// Blend between original and sharpened based on user sharpness setting
|
||||
// u_setting.x controls the blend: 0.0 = original, 1.0 = fully sharpened
|
||||
vec3 outColor = mix(C, sharpColor, u_setting.x);
|
||||
|
||||
return vec4(outColor, 1.0);
|
||||
}
|
||||
|
||||
// Main fragment shader entry point
|
||||
void main() {
|
||||
// Apply RCAS sharpening to the current fragment
|
||||
gl_FragColor = FsrRcasVanilla(v_texcoord0);
|
||||
}
|
||||
Reference in new issue
Block a user