[NV8] Improved FXAA when stipple filter is enabled (definition off)

robi29 · Oct 17, 2023 · 47377fb · 47377fb
1 parent 7138a97
commit 47377fb
Show file tree

Hide file tree

Showing 8 changed files with 17 additions and 16 deletions.
diff --git a/README.md b/README.md
@@ -661,6 +661,7 @@ Download 7zip archive from [releases](https://github.com/robi29/GTA-IV-Shaders/r
     - Added FXAA
     - Improved depth of field filter
     - Fixed aspect ratio of noise effect
+    - Improved FXAA when stipple filter is enabled (definition off)
 
 - rmptfx_collision
     - Fixed z-fighting making depth buffer logarithmic, requires FusionFix ASI script (kudos to Parallellines)

diff --git a/win32_30_nv8/rage_postfx/rage_postfxPS12.asm b/win32_30_nv8/rage_postfx/rage_postfxPS12.asm
@@ -57,7 +57,7 @@
     def c1, 0, 0.212500006, 0.715399981, 0.0720999986
     def c2, 0.25, 1, 256, 0
     def c3, 0, 2, 4, 8
-    def c4, -0.5, -1.5, 1.5, 0.5
+    def c4, 0, -1.5, 0.8, 0.5
     def c5, 2, -1, 0.125, 0
     def c6, 256, 2, 4, 8
     // NVIDIA FXAA 3.11 by Timothy Lottes
@@ -567,7 +567,7 @@
     mul r4.zw, r3.xyxy, c3.w
     mad r4.zw, v0.xyxy, c0.xyxy, r4
     texld r5, r4.zwzw, s5
-    add r2.w, r5.x, c4.x
+    add r2.w, r5.x, -c4.w
     mad r4.zw, r4.xyxy, r2.w, v0.xyxy
     mov r5.xyz, r3
     mov r2.w, c2.y

diff --git a/win32_30_nv8/rage_postfx/rage_postfxPS14.asm b/win32_30_nv8/rage_postfx/rage_postfxPS14.asm
@@ -57,7 +57,7 @@
     def c1, 0, 0.212500006, 0.715399981, 0.0720999986
     def c2, 0.25, 1, 256, 0
     def c3, 0, 2, 4, 8
-    def c4, -0.5, -1.5, 1.5, 0.5
+    def c4, 0, -1.5, 0.8, 0.5
     def c5, 2, -1, 0.125, 0
     def c6, 256, 2, 4, 8
     // NVIDIA FXAA 3.11 by Timothy Lottes
@@ -575,7 +575,7 @@
     mul r4.zw, r3.xyxy, c3.w
     mad r4.zw, v0.xyxy, c0.xyxy, r4
     texld r5, r4.zwzw, s5
-    add r2.w, r5.x, c4.x
+    add r2.w, r5.x, -c4.w
     mad r4.zw, r4.xyxy, r2.w, v0.xyxy
     mov r5.xyz, r3
     mov r2.w, c2.y

diff --git a/win32_30_nv8/rage_postfx/rage_postfxPS16.asm b/win32_30_nv8/rage_postfx/rage_postfxPS16.asm
@@ -57,7 +57,7 @@
     def c1, 0, 0.212500006, 0.715399981, 0.0720999986
     def c2, 0.25, 1, 256, 0
     def c3, 0, 2, 4, 8
-    def c4, -0.5, -1.5, 1.5, 0.5
+    def c4, 0, -1.5, 0.8, 0.5
     def c5, 2, -1, 0.125, 0
     def c6, 256, 2, 4, 8
     // NVIDIA FXAA 3.11 by Timothy Lottes
@@ -568,7 +568,7 @@
     mul r4.zw, r3.xyxz, c3.w
     mad r4.zw, v0.xyxy, c0.xyxy, r4
     texld r5, r4.zwzw, s5
-    add r2.w, r5.x, c4.x
+    add r2.w, r5.x, -c4.w
     mad r4.zw, r4.xyxy, r2.w, v0.xyxy
     mov r5.xyz, r3.xzww
     mov r2.w, c2.y

diff --git a/win32_30_nv8/rage_postfx/rage_postfxPS18.asm b/win32_30_nv8/rage_postfx/rage_postfxPS18.asm
@@ -57,7 +57,7 @@
     def c1, 0, 0.212500006, 0.715399981, 0.0720999986
     def c2, 0.25, 1, 256, 0
     def c3, 0, 2, 4, 8
-    def c4, -0.5, -1.5, 1.5, 0.5
+    def c4, 0, -1.5, 0.8, 0.5
     def c5, 2, -1, 0.125, 0
     def c6, 256, 2, 4, 8
     // NVIDIA FXAA 3.11 by Timothy Lottes
@@ -576,7 +576,7 @@
     mul r4.zw, r3.xyxz, c3.w
     mad r4.zw, v0.xyxy, c0.xyxy, r4
     texld r5, r4.zwzw, s5
-    add r2.w, r5.x, c4.x
+    add r2.w, r5.x, -c4.w
     mad r4.zw, r4.xyxy, r2.w, v0.xyxy
     mov r5.xyz, r3.xzww
     mov r2.w, c2.y

diff --git a/win32_30_nv8/rage_postfx/rage_postfxPS24.asm b/win32_30_nv8/rage_postfx/rage_postfxPS24.asm
@@ -61,7 +61,7 @@
     def c1, 0, 0.212500006, 0.715399981, 0.0720999986
     def c2, 0.25, 1, 256, 0
     def c3, 0, 2, 4, 8
-    def c4, -0.5, -1.5, 1.5, 0.5
+    def c4, 0, -1.5, 0.8, 0.5
     def c5, 2, -1, 0.125, 0
     def c6, 0, 0, 0, 0
     def c7, 256, 2, 4, 8
@@ -573,7 +573,7 @@
     mul r4.zw, c0.xyxy, v0.xyxy
     mad r4.zw, r3.xyxy, c3.w, r4
     texld r5, r4.zwzw, s6
-    add r2.w, r5.x, c4.x
+    add r2.w, r5.x, -c4.w
     mad r4.zw, r4.xyxy, r2.w, v0.xyxy
     mov r5.xyz, r3
     mov r2.w, c2.y
@@ -606,7 +606,7 @@
     mad r1.xy, v0, r1.zwzw, c85
     frc r1.xy, r1
     texld r1, r1, s3
-    add r1.z, r1.z, c4.x
+    add r1.z, r1.z, -c4.w
     mad r0.xyz, r1.z, c85.z, r0
     texld r1, c1.x, s5
     rcp r0.w, r1.x

diff --git a/win32_30_nv8/rage_postfx/rage_postfxPS26.asm b/win32_30_nv8/rage_postfx/rage_postfxPS26.asm
@@ -61,7 +61,7 @@
     def c1, 0, 0.212500006, 0.715399981, 0.0720999986
     def c2, 0.25, 1, 256, 0
     def c3, 0, 2, 4, 8
-    def c4, -0.5, -1.5, 1.5, 0.5
+    def c4, 0, -1.5, 0.8, 0.5
     def c5, 2, -1, 0.125, 0
     def c6, 0, 0, 0, 0
     def c7, 256, 2, 4, 8
@@ -581,7 +581,7 @@
     mul r4.zw, c0.xyxy, v0.xyxy
     mad r4.zw, r3.xyxy, c3.w, r4
     texld r5, r4.zwzw, s6
-    add r2.w, r5.x, c4.x
+    add r2.w, r5.x, -c4.w
     mad r4.zw, r4.xyxy, r2.w, v0.xyxy
     mov r5.xyz, r3
     mov r2.w, c2.y
@@ -614,7 +614,7 @@
     mad r1.xy, v0, r1.zwzw, c85
     frc r1.xy, r1
     texld r1, r1, s3
-    add r1.z, r1.z, c4.x
+    add r1.z, r1.z, -c4.w
     mad r0.xyz, r1.z, c85.z, r0
     texld r1, c1.x, s5
     rcp r0.w, r1.x

diff --git a/win32_30_nv8/rage_postfx/rage_postfxPS28.asm b/win32_30_nv8/rage_postfx/rage_postfxPS28.asm
@@ -53,7 +53,7 @@
     def c1, 0, 0.212500006, 0.715399981, 0.0720999986
     def c2, 0.25, 256, 1, 0
     def c3, 0, 2, 4, 8
-    def c4, -0.5, -1.5, 1.5, 0.5
+    def c4, 0, -1.5, 0.8, 0.5
     def c5, 2, -1, 0.125, 0
     def c6, 256, 2, 4, 8
     def c7, 0, 0.25, 0.5, 1
@@ -552,7 +552,7 @@
     mul r3.zw, r2.xyxy, c3.w
     mad r3.zw, v0.xyxy, c0.xyxy, r3
     texld r4, r3.zwzw, s4
-    add r1.w, r4.x, c4.x
+    add r1.w, r4.x, -c4.w
     mad r3.zw, r3.xyxy, r1.w, v0.xyxy
     mov r4.xyz, r2
     mov r1.w, c2.z