CIE: Make the RGB to Lab fast paths as accurate as the reference

The matrices used by the fast paths to convert from RGB to XYZ didn't
exactly match the reference matrices in the Babl space. This caused a
measurable error in their output.

https://bugzilla.gnome.org/show_bug.cgi?id=789695

diff --git a/extensions/CIE.c b/extensions/CIE.c
index e93b9850a4530a3157177a905cf013827e8cfacc..56108fc9b3b30b6adf94cfbebddc2427de4376d3 100644 (file)
--- a/extensions/CIE.c
+++ b/extensions/CIE.c
@@ -625,9 +625,9 @@ rgbf_to_Labf (const Babl *conversion,float *src,
       float g = src[1];
       float b = src[2];
 
-      float xr = 0.43603516f / D50_WHITE_REF_X * r + 0.38511658f / D50_WHITE_REF_X * g + 0.14305115f / D50_WHITE_REF_X * b;
-      float yr = 0.22248840f / D50_WHITE_REF_Y * r + 0.71690369f / D50_WHITE_REF_Y * g + 0.06060791f / D50_WHITE_REF_Y * b;
-      float zr = 0.01391602f / D50_WHITE_REF_Z * r + 0.09706116f / D50_WHITE_REF_Z * g + 0.71392822f / D50_WHITE_REF_Z * b;
+      float xr = 0.43605042f / D50_WHITE_REF_X * r + 0.38511658f / D50_WHITE_REF_X * g + 0.14303589f / D50_WHITE_REF_X * b;
+      float yr = 0.22250366f / D50_WHITE_REF_Y * r + 0.71690369f / D50_WHITE_REF_Y * g + 0.06059265f / D50_WHITE_REF_Y * b;
+      float zr = 0.01393127f / D50_WHITE_REF_Z * r + 0.09707642f / D50_WHITE_REF_Z * g + 0.71389771f / D50_WHITE_REF_Z * b;
 
       float fx = xr > LAB_EPSILON ? _cbrtf (xr) : (LAB_KAPPA * xr + 16.0f) / 116.0f;
       float fy = yr > LAB_EPSILON ? _cbrtf (yr) : (LAB_KAPPA * yr + 16.0f) / 116.0f;
@@ -659,9 +659,9 @@ rgbaf_to_Labf (const Babl *conversion,float *src,
       float g = src[1];
       float b = src[2];
 
-      float xr = 0.43603516f / D50_WHITE_REF_X * r + 0.38511658f / D50_WHITE_REF_X * g + 0.14305115f / D50_WHITE_REF_X * b;
-      float yr = 0.22248840f / D50_WHITE_REF_Y * r + 0.71690369f / D50_WHITE_REF_Y * g + 0.06060791f / D50_WHITE_REF_Y * b;
-      float zr = 0.01391602f / D50_WHITE_REF_Z * r + 0.09706116f / D50_WHITE_REF_Z * g + 0.71392822f / D50_WHITE_REF_Z * b;
+      float xr = 0.43605042f / D50_WHITE_REF_X * r + 0.38511658f / D50_WHITE_REF_X * g + 0.14303589f / D50_WHITE_REF_X * b;
+      float yr = 0.22250366f / D50_WHITE_REF_Y * r + 0.71690369f / D50_WHITE_REF_Y * g + 0.06059265f / D50_WHITE_REF_Y * b;
+      float zr = 0.01393127f / D50_WHITE_REF_Z * r + 0.09707642f / D50_WHITE_REF_Z * g + 0.71389771f / D50_WHITE_REF_Z * b;
 
       float fx = xr > LAB_EPSILON ? _cbrtf (xr) : (LAB_KAPPA * xr + 16.0f) / 116.0f;
       float fy = yr > LAB_EPSILON ? _cbrtf (yr) : (LAB_KAPPA * yr + 16.0f) / 116.0f;
@@ -694,9 +694,9 @@ rgbaf_to_Labaf (const Babl *conversion,float *src,
       float b = src[2];
       float a = src[3];
 
-      float xr = 0.43603516f / D50_WHITE_REF_X * r + 0.38511658f / D50_WHITE_REF_X * g + 0.14305115f / D50_WHITE_REF_X * b;
-      float yr = 0.22248840f / D50_WHITE_REF_Y * r + 0.71690369f / D50_WHITE_REF_Y * g + 0.06060791f / D50_WHITE_REF_Y * b;
-      float zr = 0.01391602f / D50_WHITE_REF_Z * r + 0.09706116f / D50_WHITE_REF_Z * g + 0.71392822f / D50_WHITE_REF_Z * b;
+      float xr = 0.43605042f / D50_WHITE_REF_X * r + 0.38511658f / D50_WHITE_REF_X * g + 0.14303589f / D50_WHITE_REF_X * b;
+      float yr = 0.22250366f / D50_WHITE_REF_Y * r + 0.71690369f / D50_WHITE_REF_Y * g + 0.06059265f / D50_WHITE_REF_Y * b;
+      float zr = 0.01393127f / D50_WHITE_REF_Z * r + 0.09707642f / D50_WHITE_REF_Z * g + 0.71389771f / D50_WHITE_REF_Z * b;
 
       float fx = xr > LAB_EPSILON ? _cbrtf (xr) : (LAB_KAPPA * xr + 16.0f) / 116.0f;
       float fy = yr > LAB_EPSILON ? _cbrtf (yr) : (LAB_KAPPA * yr + 16.0f) / 116.0f;