#include "ReShade.fxh" /* Lanczos3 - Multipass code by Hyllian 2022. */ /* Copyright (C) 2010 Team XBMC http://www.xbmc.org Copyright (C) 2011 Stefanos A. http://www.opentk.com This Program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. This Program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with XBMC; see the file COPYING. If not, write to the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. http://www.gnu.org/copyleft/gpl.html */ uniform float L3_PRESCALE < ui_type = "drag"; ui_min = 1.0; ui_max = 8.0; ui_step = 1.0; ui_label = "Prescale factor"; > = 1.0; uniform bool LANCZOS3_ANTI_RINGING < ui_type = "radio"; ui_label = "Lanczos3 Anti-Ringing"; > = true; uniform float2 NormalizedNativePixelSize < source = "normalized_native_pixel_size"; >; uniform float BufferWidth < source = "bufferwidth"; >; texture2D tLanczos3_P0{Width=BUFFER_WIDTH;Height=BUFFER_HEIGHT;Format=RGBA8;}; sampler2D sLanczos3_P0{Texture=tLanczos3_P0;AddressU=CLAMP;AddressV=CLAMP;AddressW=CLAMP;MagFilter=POINT;MinFilter=POINT;}; #define AR_STRENGTH 1.0 #define FIX(c) (max(abs(c),1e-5)) #define PI 3.1415926535897932384626433832795 #define radius 3.0 float3 weight3(float x) { float3 Sampling = FIX(2.0 * PI * float3(x - 1.5, x - 0.5, x + 0.5)); // Lanczos3. Note: we normalize outside this function, so no point in multiplying by radius. return sin(Sampling) * sin(Sampling / radius) / (Sampling * Sampling); } float3 lanczos3ar(float fp, float3 C0, float3 C1, float3 C2, float3 C3, float3 C4, float3 C5) { float3 w1 = weight3(0.5 - fp * 0.5); float3 w2 = weight3(1.0 - fp * 0.5); float sum = dot(w1, 1.0.xxx) + dot(w2, 1.0.xxx); w1 /= sum; w2 /= sum; float3 color = mul(w1, float3x3( C0, C2, C4 )) + mul(w2, float3x3( C1, C3, C5)); // Anti-ringing if (LANCZOS3_ANTI_RINGING == true) { float3 aux = color; float3 min_sample = min(min(C1, C2), min(C3, C4)); float3 max_sample = max(max(C1, C2), max(C3, C4)); color = clamp(color, min_sample, max_sample); color = lerp(aux, color, AR_STRENGTH*step(0.0, (C1-C2)*(C3-C4))); } return color; } float4 PS_Lanczos3_X(float4 vpos: SV_Position, float2 uv_tx : TEXCOORD) : SV_Target { // Both dimensions are unfiltered, so it looks for lores pixels. float2 ps = NormalizedNativePixelSize/L3_PRESCALE; float2 pos = uv_tx.xy/ps - float2(0.5, 0.0); float2 tc = (floor(pos) + 0.5.xx) * ps; float2 fp = frac(pos); float3 C0 = tex2D(ReShade::BackBuffer, tc + ps*float2(-2.0, 0.0)).rgb; float3 C1 = tex2D(ReShade::BackBuffer, tc + ps*float2(-1.0, 0.0)).rgb; float3 C2 = tex2D(ReShade::BackBuffer, tc + ps*float2( 0.0, 0.0)).rgb; float3 C3 = tex2D(ReShade::BackBuffer, tc + ps*float2( 1.0, 0.0)).rgb; float3 C4 = tex2D(ReShade::BackBuffer, tc + ps*float2( 2.0, 0.0)).rgb; float3 C5 = tex2D(ReShade::BackBuffer, tc + ps*float2( 3.0, 0.0)).rgb; float3 color = lanczos3ar(fp.x, C0, C1, C2, C3, C4, C5); return float4(color, 1.0); } float4 PS_Lanczos3_Y(float4 vpos: SV_Position, float2 uv_tx : TEXCOORD) : SV_Target { // One must be careful here. Horizontal dimension is already filtered, so it looks for x in hires. float2 ps = float2(1.0/BufferWidth, NormalizedNativePixelSize.y/L3_PRESCALE); float2 pos = uv_tx.xy/ps - float2(0.0, 0.5); float2 tc = (floor(pos) + 0.5.xx) * ps; float2 fp = frac(pos); float3 C0 = tex2D(sLanczos3_P0, tc + ps*float2(0.0, -2.0)).rgb; float3 C1 = tex2D(sLanczos3_P0, tc + ps*float2(0.0, -1.0)).rgb; float3 C2 = tex2D(sLanczos3_P0, tc + ps*float2(0.0, 0.0)).rgb; float3 C3 = tex2D(sLanczos3_P0, tc + ps*float2(0.0, 1.0)).rgb; float3 C4 = tex2D(sLanczos3_P0, tc + ps*float2(0.0, 2.0)).rgb; float3 C5 = tex2D(sLanczos3_P0, tc + ps*float2(0.0, 3.0)).rgb; float3 color = lanczos3ar(fp.y, C0, C1, C2, C3, C4, C5); return float4(color, 1.0); } technique Lanczos3 { pass { VertexShader = PostProcessVS; PixelShader = PS_Lanczos3_X; RenderTarget = tLanczos3_P0; } pass { VertexShader = PostProcessVS; PixelShader = PS_Lanczos3_Y; } }