graphics/gimp
graphics/gimp
1
0
Fork 0
mirror of https://gitlab.gnome.org/GNOME/gimp synced 2024-10-20 19:43:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

Define accessor function/macro for histogram reads and writes. This slows

Browse source 
2001-03-25 18:16:50  Adam D. Moss <adam@gimp.org>

	* app/convert.c: Define accessor function/macro for histogram
	reads and writes.  This slows us down a little because we avoid
	some of the dirty tricks we used when we knew that the histogram
	was a straight 3d array, so I've recovered some of the speed loss
	by implementing a 5d accessor function with good locality of
	reference.  This change is the first step towards quantizing in a
	more interesting colourspace than frumpy old RGB.
This commit is contained in:
18:16:50 Adam D. Moss 2001-03-25 18:41:54 +00:00 committed by Adam D. Moss
parent 6e2104c68e
commit f05240ac1e
4 changed files with 796 additions and 492 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
ChangeLog
View file

@ -1,3 +1,13 @@
2001-03-25 18:16:50 Adam D. Moss <adam@gimp.org>
* app/convert.c: Define accessor function/macro for histogram
reads and writes. This slows us down a little because we avoid
some of the dirty tricks we used when we knew that the histogram
was a straight 3d array, so I've recovered some of the speed loss
by implementing a 5d accessor function with good locality of
reference. This change is the first step towards quantizing in a
more interesting colourspace than frumpy old RGB.
2001-03-24 Nathan Summers <rock@gimp.org>
* app/tools/gimpscaletool.[ch]: Made the scale tool work again.

426
app/convert.c
View file

@ -19,12 +19,20 @@
/*
* TODO for Convert:
*
* Use palette of another open INDEXED image
*
* Do error-splitting trick for GREY->INDEXED
* . Quantize in (for example) CIE L*a*b*
* . Use palette of another open INDEXED image
* . Do error-splitting trick for GREY->INDEXED
*/
/*
* 2001-03-25 - Define accessor function/macro for histogram reads and
* writes. This slows us down a little because we avoid some of the
* dirty tricks we used when we knew that the histogram was a straight
* 3d array, so I've recovered some of the speed loss by implementing
* a 5d accessor function with good locality of reference. This change
* is the first step towards quantizing in a more interesting colourspace
* than frumpy old RGB. [Adam]
*
* 2000/01/30 - Use palette_selector instead of option_menu for custom
* palette. Use libgimp callback functions. [Sven]
*
@ -125,28 +133,8 @@
#include "libgimp/gimpintl.h"
#define PRECISION_R 6
#define PRECISION_G 6
#define PRECISION_B 6
#define HIST_R_ELEMS (1<<PRECISION_R)
#define HIST_G_ELEMS (1<<PRECISION_G)
#define HIST_B_ELEMS (1<<PRECISION_B)
#define MR HIST_G_ELEMS*HIST_B_ELEMS
#define MG HIST_B_ELEMS
#define BITS_IN_SAMPLE 8
#define R_SHIFT (BITS_IN_SAMPLE-PRECISION_R)
#define G_SHIFT (BITS_IN_SAMPLE-PRECISION_G)
#define B_SHIFT (BITS_IN_SAMPLE-PRECISION_B)
/* this has to match the INTENSITY definition in libgimp/gimpcolorspace.h */
#define R_SCALE 30 /* scale R distances by this much */
#define G_SCALE 59 /* scale G distances by this much */
#define B_SCALE 11 /* and B by this much */
/* bleh! */
static const unsigned char webpal[] =
{
255,255,255,255,255,204,255,255,153,255,255,102,255,255,51,255,255,0,255,
@ -327,6 +315,30 @@ static const guchar DM[128][128] =
{ 51, 14, 61, 29, 59, 20, 55, 31, 0, 49, 11, 60, 3, 26, 22, 56, 0, 40, 12, 43, 41, 8, 36, 0, 17, 57, 24, 2, 46, 26, 61, 18, 0, 38, 12, 59, 6, 49, 3, 57, 19, 63, 5, 33, 18, 54, 28, 56, 0, 43, 26, 46, 63, 27, 56, 22, 27, 54, 38, 28, 63, 24, 10, 45, 0, 31, 42, 21, 12, 25, 44, 49, 59, 6, 26, 50, 3, 34, 27, 59, 0, 35, 62, 16, 4, 58, 47, 0, 43, 24, 37, 2, 54, 20, 46, 31, 0, 56, 34, 5, 55, 45, 60, 37, 0, 40, 10, 38, 63, 46, 15, 20, 0, 53, 21, 62, 30, 11, 24, 27, 40, 0, 57, 26, 3, 45, 27, 35 },
};
#define PRECISION_R 6
#define PRECISION_G 6
#define PRECISION_B 6
#define HIST_R_ELEMS (1<<PRECISION_R)
#define HIST_G_ELEMS (1<<PRECISION_G)
#define HIST_B_ELEMS (1<<PRECISION_B)
#define MR (HIST_G_ELEMS*HIST_B_ELEMS)
#define MG HIST_B_ELEMS
#define BITS_IN_SAMPLE 8
#define R_SHIFT (BITS_IN_SAMPLE-PRECISION_R)
#define G_SHIFT (BITS_IN_SAMPLE-PRECISION_G)
#define B_SHIFT (BITS_IN_SAMPLE-PRECISION_B)
/* this has to match the INTENSITY definition in libgimp/gimpcolorspace.h */
#define R_SCALE 30 /* scale R distances by this much */
#define G_SCALE 59 /* scale G distances by this much */
#define B_SCALE 11 /* and B by this much */
typedef struct _Color Color;
typedef struct _QuantizeObj QuantizeObj;
typedef void (* Pass1_Func) (QuantizeObj *);
@ -336,6 +348,82 @@ typedef void (* Cleanup_Func) (QuantizeObj *);
typedef unsigned long ColorFreq;
typedef ColorFreq *CFHistogram;
/* We effectively pull our three-dimensional space into five dimensions
such that the most-entropic bits lay in the lowest bits of the resulting
array index. This gives significantly better locality of reference
and hence a small speedup despite the extra work involved in calculating
the index.
Why not six dimensions? The expansion of dimensionality is good for random
access such as histogram population and the query pattern typical of
dithering but we have some code which iterates in a scanning manner, for
which the expansion is suboptimal. The compromise is to leave the B
dimension unmolested in the lower-order bits of the index, since this is
the dimension most commonly iterated through in the inner loop of the
scans.
--adam
RhGhRlGlB
*/
#define VOL_GBITS (PRECISION_G)
#define VOL_BBITS (PRECISION_B)
#define VOL_RBITS (PRECISION_R)
#define VOL_GBITSh (VOL_GBITS - 3)
#define VOL_GBITSl (VOL_GBITS - VOL_GBITSh)
#define VOL_BBITSh (VOL_BBITS - 4)
#define VOL_BBITSl (VOL_BBITS - VOL_BBITSh)
#define VOL_RBITSh (VOL_RBITS - 3)
#define VOL_RBITSl (VOL_RBITS - VOL_RBITSh)
#define VOL_GMASKh (((1<<VOL_GBITSh)-1) << VOL_GBITSl)
#define VOL_GMASKl ((1<<VOL_GBITSl)-1)
#define VOL_BMASKh (((1<<VOL_BBITSh)-1) << VOL_BBITSl)
#define VOL_BMASKl ((1<<VOL_BBITSl)-1)
#define VOL_RMASKh (((1<<VOL_RBITSh)-1) << VOL_RBITSl)
#define VOL_RMASKl ((1<<VOL_RBITSl)-1)
/* The 5d compromise thing. */
#define REF_FUNC(r,g,b) \
( \
(((r) & VOL_RMASKh) << (VOL_BBITS + VOL_GBITS)) | \
(((r) & VOL_RMASKl) << (VOL_GBITSl + VOL_BBITS)) | \
(((g) & VOL_GMASKh) << (VOL_RBITSl + VOL_BBITS)) | \
(((g) & VOL_GMASKl) << (VOL_BBITS)) | \
(b) \
)
/* The full-on 6d thing. */
/*
#define REF_FUNC(r,g,b) \
( \
(((r) & VOL_RMASKh) << (VOL_BBITS + VOL_GBITS)) | \
(((r) & VOL_RMASKl) << (VOL_GBITSl + VOL_BBITSl)) | \
(((g) & VOL_GMASKh) << (VOL_RBITSl + VOL_BBITS)) | \
(((g) & VOL_GMASKl) << (VOL_BBITSl)) | \
(((b) & VOL_BMASKh) << (VOL_RBITSl + VOL_GBITSl)) | \
((b) & VOL_BMASKl) \
)
*/
/* The boring old 3d thing. */
/*
#define REF_FUNC(r,g,b) (((r)<<(PRECISION_G+PRECISION_B)) | ((g)<<(PRECISION_B)) | (b))
*/
/* You even get to choose whether you want the accessor function
implemented as a macro or an inline function. Don't say I never
give you anything. */
/*
#define HIST_RGB(hist_ptr,r,g,b) (&(hist_ptr)[REF_FUNC((r),(g),(b))])
*/
static inline
ColorFreq * HIST_RGB(const ColorFreq *hist_ptr,
const int r, const int g, const int b)
{
return (&(hist_ptr)[
REF_FUNC(r,g,b)
]);
}
struct _Color
{
int red;
@ -1608,12 +1696,8 @@ zero_histogram_gray (CFHistogram histogram)
static void
zero_histogram_rgb (CFHistogram histogram)
{
int r, g, b;
for (r = 0; r < HIST_R_ELEMS; r++)
for (g = 0; g < HIST_G_ELEMS; g++)
for (b = 0; b < HIST_B_ELEMS; b++)
histogram[r*MR + g*MG + b] = 0;
memset(histogram, 0,
HIST_R_ELEMS * HIST_G_ELEMS * HIST_B_ELEMS * sizeof(ColorFreq));
}
@ -1701,9 +1785,10 @@ generate_histogram_rgb (CFHistogram histogram,
)
|| (!has_alpha))
{
colfreq = & histogram[(data[RED_PIX] >> R_SHIFT) * MR +
(data[GREEN_PIX] >> G_SHIFT) * MG +
(data[BLUE_PIX] >> B_SHIFT)];
colfreq = HIST_RGB(histogram,
data[RED_PIX] >> R_SHIFT,
data[GREEN_PIX] >> G_SHIFT,
data[BLUE_PIX] >> B_SHIFT);
(*colfreq)++;
}
@ -1729,9 +1814,10 @@ generate_histogram_rgb (CFHistogram histogram,
)
|| (!has_alpha))
{
colfreq = & histogram[(data[RED_PIX] >> R_SHIFT) * MR +
(data[GREEN_PIX] >> G_SHIFT) * MG +
(data[BLUE_PIX] >> B_SHIFT)];
colfreq = HIST_RGB(histogram,
data[RED_PIX] >> R_SHIFT,
data[GREEN_PIX] >> G_SHIFT,
data[BLUE_PIX] >> B_SHIFT);
(*colfreq)++;
}
data += srcPR.bytes;
@ -1751,9 +1837,10 @@ generate_histogram_rgb (CFHistogram histogram,
((data[ALPHA_PIX] << 6) > (255 * DM[col&DM_WIDTHMASK][row&DM_HEIGHTMASK])) :
(data[ALPHA_PIX] > 127))) || (!has_alpha))
{
colfreq = & histogram[(data[RED_PIX] >> R_SHIFT) * MR +
(data[GREEN_PIX] >> G_SHIFT) * MG +
(data[BLUE_PIX] >> B_SHIFT)];
colfreq = HIST_RGB(histogram,
data[RED_PIX] >> R_SHIFT,
data[GREEN_PIX] >> G_SHIFT,
data[BLUE_PIX] >> B_SHIFT);
(*colfreq)++;
if (!needs_quantize)
@ -1894,7 +1981,7 @@ update_box_gray (CFHistogram histogram,
/* and recompute its volume and population */
{
int i, min, max, dist;
long ccount;
ColorFreq ccount;
min = boxp->Rmin;
max = boxp->Rmax;
@ -1945,11 +2032,10 @@ update_box_rgb (CFHistogram histogram,
/* Shrink the min/max bounds of a box to enclose only nonzero elements, */
/* and recompute its volume, population and error */
{
ColorFreq * histp;
int R,G,B;
int Rmin,Rmax,Gmin,Gmax,Bmin,Bmax;
int dist0,dist1,dist2;
long ccount;
ColorFreq ccount;
guint64 tempRerror;
guint64 tempGerror;
guint64 tempBerror;
@ -1964,78 +2050,84 @@ update_box_rgb (CFHistogram histogram,
for (R = Rmin; R <= Rmax; R++)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Rmin = Rmin = R;
goto have_Rmin;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Rmin = Rmin = R;
goto have_Rmin;
}
}
}
have_Rmin:
if (Rmax > Rmin)
for (R = Rmax; R >= Rmin; R--)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Rmax = Rmax = R;
goto have_Rmax;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Rmax = Rmax = R;
goto have_Rmax;
}
}
}
have_Rmax:
if (Gmax > Gmin)
for (G = Gmin; G <= Gmax; G++)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Gmin = Gmin = G;
goto have_Gmin;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Gmin = Gmin = G;
goto have_Gmin;
}
}
}
have_Gmin:
if (Gmax > Gmin)
for (G = Gmax; G >= Gmin; G--)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Gmax = Gmax = G;
goto have_Gmax;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Gmax = Gmax = G;
goto have_Gmax;
}
}
}
have_Gmax:
if (Bmax > Bmin)
for (B = Bmin; B <= Bmax; B++)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + Gmin*MG + B;
for (G = Gmin; G <= Gmax; G++, histp += MG)
if (*histp != 0)
{
boxp->Bmin = Bmin = B;
goto have_Bmin;
}
for (G = Gmin; G <= Gmax; G++)
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Bmin = Bmin = B;
goto have_Bmin;
}
}
}
have_Bmin:
if (Bmax > Bmin)
for (B = Bmax; B >= Bmin; B--)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + Gmin*MG + B;
for (G = Gmin; G <= Gmax; G++, histp += MG)
if (*histp != 0)
{
boxp->Bmax = Bmax = B;
goto have_Bmax;
}
for (G = Gmin; G <= Gmax; G++)
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Bmax = Bmax = B;
goto have_Bmax;
}
}
}
have_Bmax:
@ -2066,32 +2158,35 @@ update_box_rgb (CFHistogram histogram,
for (R = Rmin; R <= Rmax; R++)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++, histp++)
if (*histp != 0)
{
int ge, be, re;
dummybox.Rmin = dummybox.Rmax = R;
dummybox.Gmin = dummybox.Gmax = G;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
be = dummyqo.cmap[0].blue - dummyqo.cmap[1].blue;
boxp->rerror += (*histp) * re*re;
boxp->gerror += (*histp) * ge*ge;
boxp->berror += (*histp) * be*be;
boxp->error += (*histp) *
(
re*re*R_SCALE + ge*ge*G_SCALE + be*be*B_SCALE
);
ccount += *histp;
}
for (B = Bmin; B <= Bmax; B++)
{
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int ge, be, re;
dummybox.Rmin = dummybox.Rmax = R;
dummybox.Gmin = dummybox.Gmax = G;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
be = dummyqo.cmap[0].blue - dummyqo.cmap[1].blue;
boxp->rerror += freq_here * re*re;
boxp->gerror += freq_here * ge*ge;
boxp->berror += freq_here * be*be;
boxp->error += freq_here *
(
re*re*R_SCALE + ge*ge*G_SCALE + be*be*B_SCALE
);
ccount += freq_here;
}
}
}
/* Scan again, taking note of halfway error point for red axis */
@ -2103,23 +2198,26 @@ update_box_rgb (CFHistogram histogram,
for (G = Gmin; G <= Gmax; G++)
{
dummybox.Gmin = dummybox.Gmax = G;
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++, histp++)
if (*histp != 0)
{
int re;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
tempRerror += (*histp)*re*re;
if (tempRerror*2 > boxp->rerror)
goto green_axisscan;
else
boxp->Rhalferror = R;
}
for (B = Bmin; B <= Bmax; B++)
{
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int re;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
tempRerror += freq_here * re*re;
if (tempRerror*2 > boxp->rerror)
goto green_axisscan;
else
boxp->Rhalferror = R;
}
}
}
}
@ -2133,23 +2231,26 @@ update_box_rgb (CFHistogram histogram,
for (R = Rmin; R <= Rmax; R++)
{
dummybox.Rmin = dummybox.Rmax = R;
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++, histp++)
if (*histp != 0)
{
int ge;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
tempGerror += (*histp)*ge*ge;
if (tempGerror*2 > boxp->gerror)
goto blue_axisscan;
else
boxp->Ghalferror = G;
}
for (B = Bmin; B <= Bmax; B++)
{
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int ge;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
tempGerror += freq_here * ge*ge;
if (tempGerror*2 > boxp->gerror)
goto blue_axisscan;
else
boxp->Ghalferror = G;
}
}
}
}
@ -2165,8 +2266,9 @@ update_box_rgb (CFHistogram histogram,
dummybox.Rmin = dummybox.Rmax = R;
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + B;
if (*histp != 0)
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int be;
dummybox.Gmin = dummybox.Gmax = G;
@ -2174,7 +2276,7 @@ update_box_rgb (CFHistogram histogram,
be = dummyqo.cmap[0].blue - dummyqo.cmap[1].blue;
tempBerror += (*histp)*be*be;
tempBerror += freq_here * be*be;
if (tempBerror*2 > boxp->berror)
goto finished_axesscan;
@ -2402,16 +2504,14 @@ compute_color_rgb (QuantizeObj *quantobj,
{
/* Current algorithm: mean weighted by pixels (not colors) */
/* Note it is important to get the rounding correct! */
ColorFreq * histp;
int R, G, B;
int Rmin, Rmax;
int Gmin, Gmax;
int Bmin, Bmax;
long count;
long total = 0;
long Rtotal = 0;
long Gtotal = 0;
long Btotal = 0;
ColorFreq total = 0;
ColorFreq Rtotal = 0;
ColorFreq Gtotal = 0;
ColorFreq Btotal = 0;
Rmin = boxp->Rmin; Rmax = boxp->Rmax;
Gmin = boxp->Gmin; Gmax = boxp->Gmax;
@ -2420,24 +2520,24 @@ compute_color_rgb (QuantizeObj *quantobj,
for (R = Rmin; R <= Rmax; R++)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
{
if ((count = *histp++) != 0)
ColorFreq this_freq = *HIST_RGB(histogram, R, G, B);
if (this_freq != 0)
{
total += count;
Rtotal += ((R << R_SHIFT) + ((1<<R_SHIFT)>>1)) * count;
Gtotal += ((G << G_SHIFT) + ((1<<G_SHIFT)>>1)) * count;
Btotal += ((B << B_SHIFT) + ((1<<B_SHIFT)>>1)) * count;
total += this_freq;
Rtotal += ((R << R_SHIFT) + ((1<<R_SHIFT)>>1)) * this_freq;
Gtotal += ((G << G_SHIFT) + ((1<<G_SHIFT)>>1)) * this_freq;
Btotal += ((B << B_SHIFT) + ((1<<B_SHIFT)>>1)) * this_freq;
}
}
}
if (total != 0)
{
quantobj->cmap[icolor].red = (Rtotal + (total>>1)) / total;
quantobj->cmap[icolor].red = (Rtotal + (total>>1)) / total;
quantobj->cmap[icolor].green = (Gtotal + (total>>1)) / total;
quantobj->cmap[icolor].blue = (Btotal + (total>>1)) / total;
quantobj->cmap[icolor].blue = (Btotal + (total>>1)) / total;
}
else /* The only situation where total==0 is if the image was null or
* all-transparent. In that case we just put a dummy value in
@ -2856,7 +2956,6 @@ fill_inverse_cmap_rgb (QuantizeObj *quantobj,
int minR, minG, minB; /* lower left corner of update box */
int iR, iG, iB;
int * cptr; /* pointer into bestcolor[] array */
ColorFreq * cachep; /* pointer into main cache array */
/* This array lists the candidate colormap indexes. */
int colorlist[MAXNUMCOLORS];
int numcolors; /* number of candidate colors */
@ -2892,9 +2991,8 @@ fill_inverse_cmap_rgb (QuantizeObj *quantobj,
cptr = bestcolor;
for (iR = 0; iR < BOX_R_ELEMS; iR++) {
for (iG = 0; iG < BOX_G_ELEMS; iG++) {
cachep = & histogram[(R+iR)*MR+(G+iG)*MG+B];
for (iB = 0; iB < BOX_B_ELEMS; iB++) {
*cachep++ = (*cptr++) + 1;
*HIST_RGB(histogram, R+iR, G+iG, B+iB) = (*cptr++) + 1;
}
}
}
@ -3165,7 +3263,7 @@ median_cut_pass2_no_dither_rgb (QuantizeObj *quantobj,
R = (src[red_pix]) >> R_SHIFT;
G = (src[green_pix]) >> G_SHIFT;
B = (src[blue_pix]) >> B_SHIFT;
cachep = &histogram[R*MR + G*MG + B];
cachep = HIST_RGB(histogram,R,G,B);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)
@ -3246,7 +3344,7 @@ median_cut_pass2_fixed_dither_rgb (QuantizeObj *quantobj,
R = (src[red_pix]) >> R_SHIFT;
G = (src[green_pix]) >> G_SHIFT;
B = (src[blue_pix]) >> B_SHIFT;
cachep = &histogram[R*MR + G*MG + B];
cachep = HIST_RGB(histogram,R,G,B);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)
@ -3271,7 +3369,7 @@ median_cut_pass2_fixed_dither_rgb (QuantizeObj *quantobj,
G = (CLAMP0255(color->green + ge)) >> G_SHIFT;
B = (CLAMP0255(color->blue + be)) >> B_SHIFT;
cachep = &histogram[R*MR + G*MG + B];
cachep = HIST_RGB(histogram,R,G,B);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)
@ -3824,7 +3922,7 @@ median_cut_pass2_fs_dither_rgb (QuantizeObj *quantobj,
ge = g >> G_SHIFT;
be = b >> B_SHIFT;
cachep = &histogram[re*MR + ge*MG + be];
cachep = HIST_RGB(histogram, re, ge, be);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)

426
app/core/gimpimage-convert.c
View file

@ -19,12 +19,20 @@
/*
* TODO for Convert:
*
* Use palette of another open INDEXED image
*
* Do error-splitting trick for GREY->INDEXED
* . Quantize in (for example) CIE L*a*b*
* . Use palette of another open INDEXED image
* . Do error-splitting trick for GREY->INDEXED
*/
/*
* 2001-03-25 - Define accessor function/macro for histogram reads and
* writes. This slows us down a little because we avoid some of the
* dirty tricks we used when we knew that the histogram was a straight
* 3d array, so I've recovered some of the speed loss by implementing
* a 5d accessor function with good locality of reference. This change
* is the first step towards quantizing in a more interesting colourspace
* than frumpy old RGB. [Adam]
*
* 2000/01/30 - Use palette_selector instead of option_menu for custom
* palette. Use libgimp callback functions. [Sven]
*
@ -125,28 +133,8 @@
#include "libgimp/gimpintl.h"
#define PRECISION_R 6
#define PRECISION_G 6
#define PRECISION_B 6
#define HIST_R_ELEMS (1<<PRECISION_R)
#define HIST_G_ELEMS (1<<PRECISION_G)
#define HIST_B_ELEMS (1<<PRECISION_B)
#define MR HIST_G_ELEMS*HIST_B_ELEMS
#define MG HIST_B_ELEMS
#define BITS_IN_SAMPLE 8
#define R_SHIFT (BITS_IN_SAMPLE-PRECISION_R)
#define G_SHIFT (BITS_IN_SAMPLE-PRECISION_G)
#define B_SHIFT (BITS_IN_SAMPLE-PRECISION_B)
/* this has to match the INTENSITY definition in libgimp/gimpcolorspace.h */
#define R_SCALE 30 /* scale R distances by this much */
#define G_SCALE 59 /* scale G distances by this much */
#define B_SCALE 11 /* and B by this much */
/* bleh! */
static const unsigned char webpal[] =
{
255,255,255,255,255,204,255,255,153,255,255,102,255,255,51,255,255,0,255,
@ -327,6 +315,30 @@ static const guchar DM[128][128] =
{ 51, 14, 61, 29, 59, 20, 55, 31, 0, 49, 11, 60, 3, 26, 22, 56, 0, 40, 12, 43, 41, 8, 36, 0, 17, 57, 24, 2, 46, 26, 61, 18, 0, 38, 12, 59, 6, 49, 3, 57, 19, 63, 5, 33, 18, 54, 28, 56, 0, 43, 26, 46, 63, 27, 56, 22, 27, 54, 38, 28, 63, 24, 10, 45, 0, 31, 42, 21, 12, 25, 44, 49, 59, 6, 26, 50, 3, 34, 27, 59, 0, 35, 62, 16, 4, 58, 47, 0, 43, 24, 37, 2, 54, 20, 46, 31, 0, 56, 34, 5, 55, 45, 60, 37, 0, 40, 10, 38, 63, 46, 15, 20, 0, 53, 21, 62, 30, 11, 24, 27, 40, 0, 57, 26, 3, 45, 27, 35 },
};
#define PRECISION_R 6
#define PRECISION_G 6
#define PRECISION_B 6
#define HIST_R_ELEMS (1<<PRECISION_R)
#define HIST_G_ELEMS (1<<PRECISION_G)
#define HIST_B_ELEMS (1<<PRECISION_B)
#define MR (HIST_G_ELEMS*HIST_B_ELEMS)
#define MG HIST_B_ELEMS
#define BITS_IN_SAMPLE 8
#define R_SHIFT (BITS_IN_SAMPLE-PRECISION_R)
#define G_SHIFT (BITS_IN_SAMPLE-PRECISION_G)
#define B_SHIFT (BITS_IN_SAMPLE-PRECISION_B)
/* this has to match the INTENSITY definition in libgimp/gimpcolorspace.h */
#define R_SCALE 30 /* scale R distances by this much */
#define G_SCALE 59 /* scale G distances by this much */
#define B_SCALE 11 /* and B by this much */
typedef struct _Color Color;
typedef struct _QuantizeObj QuantizeObj;
typedef void (* Pass1_Func) (QuantizeObj *);
@ -336,6 +348,82 @@ typedef void (* Cleanup_Func) (QuantizeObj *);
typedef unsigned long ColorFreq;
typedef ColorFreq *CFHistogram;
/* We effectively pull our three-dimensional space into five dimensions
such that the most-entropic bits lay in the lowest bits of the resulting
array index. This gives significantly better locality of reference
and hence a small speedup despite the extra work involved in calculating
the index.
Why not six dimensions? The expansion of dimensionality is good for random
access such as histogram population and the query pattern typical of
dithering but we have some code which iterates in a scanning manner, for
which the expansion is suboptimal. The compromise is to leave the B
dimension unmolested in the lower-order bits of the index, since this is
the dimension most commonly iterated through in the inner loop of the
scans.
--adam
RhGhRlGlB
*/
#define VOL_GBITS (PRECISION_G)
#define VOL_BBITS (PRECISION_B)
#define VOL_RBITS (PRECISION_R)
#define VOL_GBITSh (VOL_GBITS - 3)
#define VOL_GBITSl (VOL_GBITS - VOL_GBITSh)
#define VOL_BBITSh (VOL_BBITS - 4)
#define VOL_BBITSl (VOL_BBITS - VOL_BBITSh)
#define VOL_RBITSh (VOL_RBITS - 3)
#define VOL_RBITSl (VOL_RBITS - VOL_RBITSh)
#define VOL_GMASKh (((1<<VOL_GBITSh)-1) << VOL_GBITSl)
#define VOL_GMASKl ((1<<VOL_GBITSl)-1)
#define VOL_BMASKh (((1<<VOL_BBITSh)-1) << VOL_BBITSl)
#define VOL_BMASKl ((1<<VOL_BBITSl)-1)
#define VOL_RMASKh (((1<<VOL_RBITSh)-1) << VOL_RBITSl)
#define VOL_RMASKl ((1<<VOL_RBITSl)-1)
/* The 5d compromise thing. */
#define REF_FUNC(r,g,b) \
( \
(((r) & VOL_RMASKh) << (VOL_BBITS + VOL_GBITS)) | \
(((r) & VOL_RMASKl) << (VOL_GBITSl + VOL_BBITS)) | \
(((g) & VOL_GMASKh) << (VOL_RBITSl + VOL_BBITS)) | \
(((g) & VOL_GMASKl) << (VOL_BBITS)) | \
(b) \
)
/* The full-on 6d thing. */
/*
#define REF_FUNC(r,g,b) \
( \
(((r) & VOL_RMASKh) << (VOL_BBITS + VOL_GBITS)) | \
(((r) & VOL_RMASKl) << (VOL_GBITSl + VOL_BBITSl)) | \
(((g) & VOL_GMASKh) << (VOL_RBITSl + VOL_BBITS)) | \
(((g) & VOL_GMASKl) << (VOL_BBITSl)) | \
(((b) & VOL_BMASKh) << (VOL_RBITSl + VOL_GBITSl)) | \
((b) & VOL_BMASKl) \
)
*/
/* The boring old 3d thing. */
/*
#define REF_FUNC(r,g,b) (((r)<<(PRECISION_G+PRECISION_B)) | ((g)<<(PRECISION_B)) | (b))
*/
/* You even get to choose whether you want the accessor function
implemented as a macro or an inline function. Don't say I never
give you anything. */
/*
#define HIST_RGB(hist_ptr,r,g,b) (&(hist_ptr)[REF_FUNC((r),(g),(b))])
*/
static inline
ColorFreq * HIST_RGB(const ColorFreq *hist_ptr,
const int r, const int g, const int b)
{
return (&(hist_ptr)[
REF_FUNC(r,g,b)
]);
}
struct _Color
{
int red;
@ -1608,12 +1696,8 @@ zero_histogram_gray (CFHistogram histogram)
static void
zero_histogram_rgb (CFHistogram histogram)
{
int r, g, b;
for (r = 0; r < HIST_R_ELEMS; r++)
for (g = 0; g < HIST_G_ELEMS; g++)
for (b = 0; b < HIST_B_ELEMS; b++)
histogram[r*MR + g*MG + b] = 0;
memset(histogram, 0,
HIST_R_ELEMS * HIST_G_ELEMS * HIST_B_ELEMS * sizeof(ColorFreq));
}
@ -1701,9 +1785,10 @@ generate_histogram_rgb (CFHistogram histogram,
)
|| (!has_alpha))
{
colfreq = & histogram[(data[RED_PIX] >> R_SHIFT) * MR +
(data[GREEN_PIX] >> G_SHIFT) * MG +
(data[BLUE_PIX] >> B_SHIFT)];
colfreq = HIST_RGB(histogram,
data[RED_PIX] >> R_SHIFT,
data[GREEN_PIX] >> G_SHIFT,
data[BLUE_PIX] >> B_SHIFT);
(*colfreq)++;
}
@ -1729,9 +1814,10 @@ generate_histogram_rgb (CFHistogram histogram,
)
|| (!has_alpha))
{
colfreq = & histogram[(data[RED_PIX] >> R_SHIFT) * MR +
(data[GREEN_PIX] >> G_SHIFT) * MG +
(data[BLUE_PIX] >> B_SHIFT)];
colfreq = HIST_RGB(histogram,
data[RED_PIX] >> R_SHIFT,
data[GREEN_PIX] >> G_SHIFT,
data[BLUE_PIX] >> B_SHIFT);
(*colfreq)++;
}
data += srcPR.bytes;
@ -1751,9 +1837,10 @@ generate_histogram_rgb (CFHistogram histogram,
((data[ALPHA_PIX] << 6) > (255 * DM[col&DM_WIDTHMASK][row&DM_HEIGHTMASK])) :
(data[ALPHA_PIX] > 127))) || (!has_alpha))
{
colfreq = & histogram[(data[RED_PIX] >> R_SHIFT) * MR +
(data[GREEN_PIX] >> G_SHIFT) * MG +
(data[BLUE_PIX] >> B_SHIFT)];
colfreq = HIST_RGB(histogram,
data[RED_PIX] >> R_SHIFT,
data[GREEN_PIX] >> G_SHIFT,
data[BLUE_PIX] >> B_SHIFT);
(*colfreq)++;
if (!needs_quantize)
@ -1894,7 +1981,7 @@ update_box_gray (CFHistogram histogram,
/* and recompute its volume and population */
{
int i, min, max, dist;
long ccount;
ColorFreq ccount;
min = boxp->Rmin;
max = boxp->Rmax;
@ -1945,11 +2032,10 @@ update_box_rgb (CFHistogram histogram,
/* Shrink the min/max bounds of a box to enclose only nonzero elements, */
/* and recompute its volume, population and error */
{
ColorFreq * histp;
int R,G,B;
int Rmin,Rmax,Gmin,Gmax,Bmin,Bmax;
int dist0,dist1,dist2;
long ccount;
ColorFreq ccount;
guint64 tempRerror;
guint64 tempGerror;
guint64 tempBerror;
@ -1964,78 +2050,84 @@ update_box_rgb (CFHistogram histogram,
for (R = Rmin; R <= Rmax; R++)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Rmin = Rmin = R;
goto have_Rmin;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Rmin = Rmin = R;
goto have_Rmin;
}
}
}
have_Rmin:
if (Rmax > Rmin)
for (R = Rmax; R >= Rmin; R--)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Rmax = Rmax = R;
goto have_Rmax;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Rmax = Rmax = R;
goto have_Rmax;
}
}
}
have_Rmax:
if (Gmax > Gmin)
for (G = Gmin; G <= Gmax; G++)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Gmin = Gmin = G;
goto have_Gmin;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Gmin = Gmin = G;
goto have_Gmin;
}
}
}
have_Gmin:
if (Gmax > Gmin)
for (G = Gmax; G >= Gmin; G--)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Gmax = Gmax = G;
goto have_Gmax;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Gmax = Gmax = G;
goto have_Gmax;
}
}
}
have_Gmax:
if (Bmax > Bmin)
for (B = Bmin; B <= Bmax; B++)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + Gmin*MG + B;
for (G = Gmin; G <= Gmax; G++, histp += MG)
if (*histp != 0)
{
boxp->Bmin = Bmin = B;
goto have_Bmin;
}
for (G = Gmin; G <= Gmax; G++)
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Bmin = Bmin = B;
goto have_Bmin;
}
}
}
have_Bmin:
if (Bmax > Bmin)
for (B = Bmax; B >= Bmin; B--)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + Gmin*MG + B;
for (G = Gmin; G <= Gmax; G++, histp += MG)
if (*histp != 0)
{
boxp->Bmax = Bmax = B;
goto have_Bmax;
}
for (G = Gmin; G <= Gmax; G++)
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Bmax = Bmax = B;
goto have_Bmax;
}
}
}
have_Bmax:
@ -2066,32 +2158,35 @@ update_box_rgb (CFHistogram histogram,
for (R = Rmin; R <= Rmax; R++)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++, histp++)
if (*histp != 0)
{
int ge, be, re;
dummybox.Rmin = dummybox.Rmax = R;
dummybox.Gmin = dummybox.Gmax = G;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
be = dummyqo.cmap[0].blue - dummyqo.cmap[1].blue;
boxp->rerror += (*histp) * re*re;
boxp->gerror += (*histp) * ge*ge;
boxp->berror += (*histp) * be*be;
boxp->error += (*histp) *
(
re*re*R_SCALE + ge*ge*G_SCALE + be*be*B_SCALE
);
ccount += *histp;
}
for (B = Bmin; B <= Bmax; B++)
{
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int ge, be, re;
dummybox.Rmin = dummybox.Rmax = R;
dummybox.Gmin = dummybox.Gmax = G;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
be = dummyqo.cmap[0].blue - dummyqo.cmap[1].blue;
boxp->rerror += freq_here * re*re;
boxp->gerror += freq_here * ge*ge;
boxp->berror += freq_here * be*be;
boxp->error += freq_here *
(
re*re*R_SCALE + ge*ge*G_SCALE + be*be*B_SCALE
);
ccount += freq_here;
}
}
}
/* Scan again, taking note of halfway error point for red axis */
@ -2103,23 +2198,26 @@ update_box_rgb (CFHistogram histogram,
for (G = Gmin; G <= Gmax; G++)
{
dummybox.Gmin = dummybox.Gmax = G;
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++, histp++)
if (*histp != 0)
{
int re;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
tempRerror += (*histp)*re*re;
if (tempRerror*2 > boxp->rerror)
goto green_axisscan;
else
boxp->Rhalferror = R;
}
for (B = Bmin; B <= Bmax; B++)
{
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int re;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
tempRerror += freq_here * re*re;
if (tempRerror*2 > boxp->rerror)
goto green_axisscan;
else
boxp->Rhalferror = R;
}
}
}
}
@ -2133,23 +2231,26 @@ update_box_rgb (CFHistogram histogram,
for (R = Rmin; R <= Rmax; R++)
{
dummybox.Rmin = dummybox.Rmax = R;
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++, histp++)
if (*histp != 0)
{
int ge;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
tempGerror += (*histp)*ge*ge;
if (tempGerror*2 > boxp->gerror)
goto blue_axisscan;
else
boxp->Ghalferror = G;
}
for (B = Bmin; B <= Bmax; B++)
{
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int ge;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
tempGerror += freq_here * ge*ge;
if (tempGerror*2 > boxp->gerror)
goto blue_axisscan;
else
boxp->Ghalferror = G;
}
}
}
}
@ -2165,8 +2266,9 @@ update_box_rgb (CFHistogram histogram,
dummybox.Rmin = dummybox.Rmax = R;
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + B;
if (*histp != 0)
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int be;
dummybox.Gmin = dummybox.Gmax = G;
@ -2174,7 +2276,7 @@ update_box_rgb (CFHistogram histogram,
be = dummyqo.cmap[0].blue - dummyqo.cmap[1].blue;
tempBerror += (*histp)*be*be;
tempBerror += freq_here * be*be;
if (tempBerror*2 > boxp->berror)
goto finished_axesscan;
@ -2402,16 +2504,14 @@ compute_color_rgb (QuantizeObj *quantobj,
{
/* Current algorithm: mean weighted by pixels (not colors) */
/* Note it is important to get the rounding correct! */
ColorFreq * histp;
int R, G, B;
int Rmin, Rmax;
int Gmin, Gmax;
int Bmin, Bmax;
long count;
long total = 0;
long Rtotal = 0;
long Gtotal = 0;
long Btotal = 0;
ColorFreq total = 0;
ColorFreq Rtotal = 0;
ColorFreq Gtotal = 0;
ColorFreq Btotal = 0;
Rmin = boxp->Rmin; Rmax = boxp->Rmax;
Gmin = boxp->Gmin; Gmax = boxp->Gmax;
@ -2420,24 +2520,24 @@ compute_color_rgb (QuantizeObj *quantobj,
for (R = Rmin; R <= Rmax; R++)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
{
if ((count = *histp++) != 0)
ColorFreq this_freq = *HIST_RGB(histogram, R, G, B);
if (this_freq != 0)
{
total += count;
Rtotal += ((R << R_SHIFT) + ((1<<R_SHIFT)>>1)) * count;
Gtotal += ((G << G_SHIFT) + ((1<<G_SHIFT)>>1)) * count;
Btotal += ((B << B_SHIFT) + ((1<<B_SHIFT)>>1)) * count;
total += this_freq;
Rtotal += ((R << R_SHIFT) + ((1<<R_SHIFT)>>1)) * this_freq;
Gtotal += ((G << G_SHIFT) + ((1<<G_SHIFT)>>1)) * this_freq;
Btotal += ((B << B_SHIFT) + ((1<<B_SHIFT)>>1)) * this_freq;
}
}
}
if (total != 0)
{
quantobj->cmap[icolor].red = (Rtotal + (total>>1)) / total;
quantobj->cmap[icolor].red = (Rtotal + (total>>1)) / total;
quantobj->cmap[icolor].green = (Gtotal + (total>>1)) / total;
quantobj->cmap[icolor].blue = (Btotal + (total>>1)) / total;
quantobj->cmap[icolor].blue = (Btotal + (total>>1)) / total;
}
else /* The only situation where total==0 is if the image was null or
* all-transparent. In that case we just put a dummy value in
@ -2856,7 +2956,6 @@ fill_inverse_cmap_rgb (QuantizeObj *quantobj,
int minR, minG, minB; /* lower left corner of update box */
int iR, iG, iB;
int * cptr; /* pointer into bestcolor[] array */
ColorFreq * cachep; /* pointer into main cache array */
/* This array lists the candidate colormap indexes. */
int colorlist[MAXNUMCOLORS];
int numcolors; /* number of candidate colors */
@ -2892,9 +2991,8 @@ fill_inverse_cmap_rgb (QuantizeObj *quantobj,
cptr = bestcolor;
for (iR = 0; iR < BOX_R_ELEMS; iR++) {
for (iG = 0; iG < BOX_G_ELEMS; iG++) {
cachep = & histogram[(R+iR)*MR+(G+iG)*MG+B];
for (iB = 0; iB < BOX_B_ELEMS; iB++) {
*cachep++ = (*cptr++) + 1;
*HIST_RGB(histogram, R+iR, G+iG, B+iB) = (*cptr++) + 1;
}
}
}
@ -3165,7 +3263,7 @@ median_cut_pass2_no_dither_rgb (QuantizeObj *quantobj,
R = (src[red_pix]) >> R_SHIFT;
G = (src[green_pix]) >> G_SHIFT;
B = (src[blue_pix]) >> B_SHIFT;
cachep = &histogram[R*MR + G*MG + B];
cachep = HIST_RGB(histogram,R,G,B);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)
@ -3246,7 +3344,7 @@ median_cut_pass2_fixed_dither_rgb (QuantizeObj *quantobj,
R = (src[red_pix]) >> R_SHIFT;
G = (src[green_pix]) >> G_SHIFT;
B = (src[blue_pix]) >> B_SHIFT;
cachep = &histogram[R*MR + G*MG + B];
cachep = HIST_RGB(histogram,R,G,B);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)
@ -3271,7 +3369,7 @@ median_cut_pass2_fixed_dither_rgb (QuantizeObj *quantobj,
G = (CLAMP0255(color->green + ge)) >> G_SHIFT;
B = (CLAMP0255(color->blue + be)) >> B_SHIFT;
cachep = &histogram[R*MR + G*MG + B];
cachep = HIST_RGB(histogram,R,G,B);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)
@ -3824,7 +3922,7 @@ median_cut_pass2_fs_dither_rgb (QuantizeObj *quantobj,
ge = g >> G_SHIFT;
be = b >> B_SHIFT;
cachep = &histogram[re*MR + ge*MG + be];
cachep = HIST_RGB(histogram, re, ge, be);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)

426
app/gimpimage-convert.c
View file

@ -19,12 +19,20 @@
/*
* TODO for Convert:
*
* Use palette of another open INDEXED image
*
* Do error-splitting trick for GREY->INDEXED
* . Quantize in (for example) CIE L*a*b*
* . Use palette of another open INDEXED image
* . Do error-splitting trick for GREY->INDEXED
*/
/*
* 2001-03-25 - Define accessor function/macro for histogram reads and
* writes. This slows us down a little because we avoid some of the
* dirty tricks we used when we knew that the histogram was a straight
* 3d array, so I've recovered some of the speed loss by implementing
* a 5d accessor function with good locality of reference. This change
* is the first step towards quantizing in a more interesting colourspace
* than frumpy old RGB. [Adam]
*
* 2000/01/30 - Use palette_selector instead of option_menu for custom
* palette. Use libgimp callback functions. [Sven]
*
@ -125,28 +133,8 @@
#include "libgimp/gimpintl.h"
#define PRECISION_R 6
#define PRECISION_G 6
#define PRECISION_B 6
#define HIST_R_ELEMS (1<<PRECISION_R)
#define HIST_G_ELEMS (1<<PRECISION_G)
#define HIST_B_ELEMS (1<<PRECISION_B)
#define MR HIST_G_ELEMS*HIST_B_ELEMS
#define MG HIST_B_ELEMS
#define BITS_IN_SAMPLE 8
#define R_SHIFT (BITS_IN_SAMPLE-PRECISION_R)
#define G_SHIFT (BITS_IN_SAMPLE-PRECISION_G)
#define B_SHIFT (BITS_IN_SAMPLE-PRECISION_B)
/* this has to match the INTENSITY definition in libgimp/gimpcolorspace.h */
#define R_SCALE 30 /* scale R distances by this much */
#define G_SCALE 59 /* scale G distances by this much */
#define B_SCALE 11 /* and B by this much */
/* bleh! */
static const unsigned char webpal[] =
{
255,255,255,255,255,204,255,255,153,255,255,102,255,255,51,255,255,0,255,
@ -327,6 +315,30 @@ static const guchar DM[128][128] =
{ 51, 14, 61, 29, 59, 20, 55, 31, 0, 49, 11, 60, 3, 26, 22, 56, 0, 40, 12, 43, 41, 8, 36, 0, 17, 57, 24, 2, 46, 26, 61, 18, 0, 38, 12, 59, 6, 49, 3, 57, 19, 63, 5, 33, 18, 54, 28, 56, 0, 43, 26, 46, 63, 27, 56, 22, 27, 54, 38, 28, 63, 24, 10, 45, 0, 31, 42, 21, 12, 25, 44, 49, 59, 6, 26, 50, 3, 34, 27, 59, 0, 35, 62, 16, 4, 58, 47, 0, 43, 24, 37, 2, 54, 20, 46, 31, 0, 56, 34, 5, 55, 45, 60, 37, 0, 40, 10, 38, 63, 46, 15, 20, 0, 53, 21, 62, 30, 11, 24, 27, 40, 0, 57, 26, 3, 45, 27, 35 },
};
#define PRECISION_R 6
#define PRECISION_G 6
#define PRECISION_B 6
#define HIST_R_ELEMS (1<<PRECISION_R)
#define HIST_G_ELEMS (1<<PRECISION_G)
#define HIST_B_ELEMS (1<<PRECISION_B)
#define MR (HIST_G_ELEMS*HIST_B_ELEMS)
#define MG HIST_B_ELEMS
#define BITS_IN_SAMPLE 8
#define R_SHIFT (BITS_IN_SAMPLE-PRECISION_R)
#define G_SHIFT (BITS_IN_SAMPLE-PRECISION_G)
#define B_SHIFT (BITS_IN_SAMPLE-PRECISION_B)
/* this has to match the INTENSITY definition in libgimp/gimpcolorspace.h */
#define R_SCALE 30 /* scale R distances by this much */
#define G_SCALE 59 /* scale G distances by this much */
#define B_SCALE 11 /* and B by this much */
typedef struct _Color Color;
typedef struct _QuantizeObj QuantizeObj;
typedef void (* Pass1_Func) (QuantizeObj *);
@ -336,6 +348,82 @@ typedef void (* Cleanup_Func) (QuantizeObj *);
typedef unsigned long ColorFreq;
typedef ColorFreq *CFHistogram;
/* We effectively pull our three-dimensional space into five dimensions
such that the most-entropic bits lay in the lowest bits of the resulting
array index. This gives significantly better locality of reference
and hence a small speedup despite the extra work involved in calculating
the index.
Why not six dimensions? The expansion of dimensionality is good for random
access such as histogram population and the query pattern typical of
dithering but we have some code which iterates in a scanning manner, for
which the expansion is suboptimal. The compromise is to leave the B
dimension unmolested in the lower-order bits of the index, since this is
the dimension most commonly iterated through in the inner loop of the
scans.
--adam
RhGhRlGlB
*/
#define VOL_GBITS (PRECISION_G)
#define VOL_BBITS (PRECISION_B)
#define VOL_RBITS (PRECISION_R)
#define VOL_GBITSh (VOL_GBITS - 3)
#define VOL_GBITSl (VOL_GBITS - VOL_GBITSh)
#define VOL_BBITSh (VOL_BBITS - 4)
#define VOL_BBITSl (VOL_BBITS - VOL_BBITSh)
#define VOL_RBITSh (VOL_RBITS - 3)
#define VOL_RBITSl (VOL_RBITS - VOL_RBITSh)
#define VOL_GMASKh (((1<<VOL_GBITSh)-1) << VOL_GBITSl)
#define VOL_GMASKl ((1<<VOL_GBITSl)-1)
#define VOL_BMASKh (((1<<VOL_BBITSh)-1) << VOL_BBITSl)
#define VOL_BMASKl ((1<<VOL_BBITSl)-1)
#define VOL_RMASKh (((1<<VOL_RBITSh)-1) << VOL_RBITSl)
#define VOL_RMASKl ((1<<VOL_RBITSl)-1)
/* The 5d compromise thing. */
#define REF_FUNC(r,g,b) \
( \
(((r) & VOL_RMASKh) << (VOL_BBITS + VOL_GBITS)) | \
(((r) & VOL_RMASKl) << (VOL_GBITSl + VOL_BBITS)) | \
(((g) & VOL_GMASKh) << (VOL_RBITSl + VOL_BBITS)) | \
(((g) & VOL_GMASKl) << (VOL_BBITS)) | \
(b) \
)
/* The full-on 6d thing. */
/*
#define REF_FUNC(r,g,b) \
( \
(((r) & VOL_RMASKh) << (VOL_BBITS + VOL_GBITS)) | \
(((r) & VOL_RMASKl) << (VOL_GBITSl + VOL_BBITSl)) | \
(((g) & VOL_GMASKh) << (VOL_RBITSl + VOL_BBITS)) | \
(((g) & VOL_GMASKl) << (VOL_BBITSl)) | \
(((b) & VOL_BMASKh) << (VOL_RBITSl + VOL_GBITSl)) | \
((b) & VOL_BMASKl) \
)
*/
/* The boring old 3d thing. */
/*
#define REF_FUNC(r,g,b) (((r)<<(PRECISION_G+PRECISION_B)) | ((g)<<(PRECISION_B)) | (b))
*/
/* You even get to choose whether you want the accessor function
implemented as a macro or an inline function. Don't say I never
give you anything. */
/*
#define HIST_RGB(hist_ptr,r,g,b) (&(hist_ptr)[REF_FUNC((r),(g),(b))])
*/
static inline
ColorFreq * HIST_RGB(const ColorFreq *hist_ptr,
const int r, const int g, const int b)
{
return (&(hist_ptr)[
REF_FUNC(r,g,b)
]);
}
struct _Color
{
int red;
@ -1608,12 +1696,8 @@ zero_histogram_gray (CFHistogram histogram)
static void
zero_histogram_rgb (CFHistogram histogram)
{
int r, g, b;
for (r = 0; r < HIST_R_ELEMS; r++)
for (g = 0; g < HIST_G_ELEMS; g++)
for (b = 0; b < HIST_B_ELEMS; b++)
histogram[r*MR + g*MG + b] = 0;
memset(histogram, 0,
HIST_R_ELEMS * HIST_G_ELEMS * HIST_B_ELEMS * sizeof(ColorFreq));
}
@ -1701,9 +1785,10 @@ generate_histogram_rgb (CFHistogram histogram,
)
|| (!has_alpha))
{
colfreq = & histogram[(data[RED_PIX] >> R_SHIFT) * MR +
(data[GREEN_PIX] >> G_SHIFT) * MG +
(data[BLUE_PIX] >> B_SHIFT)];
colfreq = HIST_RGB(histogram,
data[RED_PIX] >> R_SHIFT,
data[GREEN_PIX] >> G_SHIFT,
data[BLUE_PIX] >> B_SHIFT);
(*colfreq)++;
}
@ -1729,9 +1814,10 @@ generate_histogram_rgb (CFHistogram histogram,
)
|| (!has_alpha))
{
colfreq = & histogram[(data[RED_PIX] >> R_SHIFT) * MR +
(data[GREEN_PIX] >> G_SHIFT) * MG +
(data[BLUE_PIX] >> B_SHIFT)];
colfreq = HIST_RGB(histogram,
data[RED_PIX] >> R_SHIFT,
data[GREEN_PIX] >> G_SHIFT,
data[BLUE_PIX] >> B_SHIFT);
(*colfreq)++;
}
data += srcPR.bytes;
@ -1751,9 +1837,10 @@ generate_histogram_rgb (CFHistogram histogram,
((data[ALPHA_PIX] << 6) > (255 * DM[col&DM_WIDTHMASK][row&DM_HEIGHTMASK])) :
(data[ALPHA_PIX] > 127))) || (!has_alpha))
{
colfreq = & histogram[(data[RED_PIX] >> R_SHIFT) * MR +
(data[GREEN_PIX] >> G_SHIFT) * MG +
(data[BLUE_PIX] >> B_SHIFT)];
colfreq = HIST_RGB(histogram,
data[RED_PIX] >> R_SHIFT,
data[GREEN_PIX] >> G_SHIFT,
data[BLUE_PIX] >> B_SHIFT);
(*colfreq)++;
if (!needs_quantize)
@ -1894,7 +1981,7 @@ update_box_gray (CFHistogram histogram,
/* and recompute its volume and population */
{
int i, min, max, dist;
long ccount;
ColorFreq ccount;
min = boxp->Rmin;
max = boxp->Rmax;
@ -1945,11 +2032,10 @@ update_box_rgb (CFHistogram histogram,
/* Shrink the min/max bounds of a box to enclose only nonzero elements, */
/* and recompute its volume, population and error */
{
ColorFreq * histp;
int R,G,B;
int Rmin,Rmax,Gmin,Gmax,Bmin,Bmax;
int dist0,dist1,dist2;
long ccount;
ColorFreq ccount;
guint64 tempRerror;
guint64 tempGerror;
guint64 tempBerror;
@ -1964,78 +2050,84 @@ update_box_rgb (CFHistogram histogram,
for (R = Rmin; R <= Rmax; R++)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Rmin = Rmin = R;
goto have_Rmin;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Rmin = Rmin = R;
goto have_Rmin;
}
}
}
have_Rmin:
if (Rmax > Rmin)
for (R = Rmax; R >= Rmin; R--)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Rmax = Rmax = R;
goto have_Rmax;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Rmax = Rmax = R;
goto have_Rmax;
}
}
}
have_Rmax:
if (Gmax > Gmin)
for (G = Gmin; G <= Gmax; G++)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Gmin = Gmin = G;
goto have_Gmin;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Gmin = Gmin = G;
goto have_Gmin;
}
}
}
have_Gmin:
if (Gmax > Gmin)
for (G = Gmax; G >= Gmin; G--)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
if (*histp++ != 0)
{
boxp->Gmax = Gmax = G;
goto have_Gmax;
}
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Gmax = Gmax = G;
goto have_Gmax;
}
}
}
have_Gmax:
if (Bmax > Bmin)
for (B = Bmin; B <= Bmax; B++)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + Gmin*MG + B;
for (G = Gmin; G <= Gmax; G++, histp += MG)
if (*histp != 0)
{
boxp->Bmin = Bmin = B;
goto have_Bmin;
}
for (G = Gmin; G <= Gmax; G++)
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Bmin = Bmin = B;
goto have_Bmin;
}
}
}
have_Bmin:
if (Bmax > Bmin)
for (B = Bmax; B >= Bmin; B--)
for (R = Rmin; R <= Rmax; R++)
{
histp = histogram + R*MR + Gmin*MG + B;
for (G = Gmin; G <= Gmax; G++, histp += MG)
if (*histp != 0)
{
boxp->Bmax = Bmax = B;
goto have_Bmax;
}
for (G = Gmin; G <= Gmax; G++)
{
if (*HIST_RGB(histogram, R, G, B) != 0)
{
boxp->Bmax = Bmax = B;
goto have_Bmax;
}
}
}
have_Bmax:
@ -2066,32 +2158,35 @@ update_box_rgb (CFHistogram histogram,
for (R = Rmin; R <= Rmax; R++)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++, histp++)
if (*histp != 0)
{
int ge, be, re;
dummybox.Rmin = dummybox.Rmax = R;
dummybox.Gmin = dummybox.Gmax = G;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
be = dummyqo.cmap[0].blue - dummyqo.cmap[1].blue;
boxp->rerror += (*histp) * re*re;
boxp->gerror += (*histp) * ge*ge;
boxp->berror += (*histp) * be*be;
boxp->error += (*histp) *
(
re*re*R_SCALE + ge*ge*G_SCALE + be*be*B_SCALE
);
ccount += *histp;
}
for (B = Bmin; B <= Bmax; B++)
{
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int ge, be, re;
dummybox.Rmin = dummybox.Rmax = R;
dummybox.Gmin = dummybox.Gmax = G;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
be = dummyqo.cmap[0].blue - dummyqo.cmap[1].blue;
boxp->rerror += freq_here * re*re;
boxp->gerror += freq_here * ge*ge;
boxp->berror += freq_here * be*be;
boxp->error += freq_here *
(
re*re*R_SCALE + ge*ge*G_SCALE + be*be*B_SCALE
);
ccount += freq_here;
}
}
}
/* Scan again, taking note of halfway error point for red axis */
@ -2103,23 +2198,26 @@ update_box_rgb (CFHistogram histogram,
for (G = Gmin; G <= Gmax; G++)
{
dummybox.Gmin = dummybox.Gmax = G;
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++, histp++)
if (*histp != 0)
{
int re;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
tempRerror += (*histp)*re*re;
if (tempRerror*2 > boxp->rerror)
goto green_axisscan;
else
boxp->Rhalferror = R;
}
for (B = Bmin; B <= Bmax; B++)
{
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int re;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
re = dummyqo.cmap[0].red - dummyqo.cmap[1].red;
tempRerror += freq_here * re*re;
if (tempRerror*2 > boxp->rerror)
goto green_axisscan;
else
boxp->Rhalferror = R;
}
}
}
}
@ -2133,23 +2231,26 @@ update_box_rgb (CFHistogram histogram,
for (R = Rmin; R <= Rmax; R++)
{
dummybox.Rmin = dummybox.Rmax = R;
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++, histp++)
if (*histp != 0)
{
int ge;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
tempGerror += (*histp)*ge*ge;
if (tempGerror*2 > boxp->gerror)
goto blue_axisscan;
else
boxp->Ghalferror = G;
}
for (B = Bmin; B <= Bmax; B++)
{
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int ge;
dummybox.Bmin = dummybox.Bmax = B;
compute_color_rgb(&dummyqo, histogram, &dummybox, 1);
ge = dummyqo.cmap[0].green - dummyqo.cmap[1].green;
tempGerror += freq_here * ge*ge;
if (tempGerror*2 > boxp->gerror)
goto blue_axisscan;
else
boxp->Ghalferror = G;
}
}
}
}
@ -2165,8 +2266,9 @@ update_box_rgb (CFHistogram histogram,
dummybox.Rmin = dummybox.Rmax = R;
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + B;
if (*histp != 0)
ColorFreq freq_here;
freq_here = *HIST_RGB(histogram, R, G, B);
if (freq_here != 0)
{
int be;
dummybox.Gmin = dummybox.Gmax = G;
@ -2174,7 +2276,7 @@ update_box_rgb (CFHistogram histogram,
be = dummyqo.cmap[0].blue - dummyqo.cmap[1].blue;
tempBerror += (*histp)*be*be;
tempBerror += freq_here * be*be;
if (tempBerror*2 > boxp->berror)
goto finished_axesscan;
@ -2402,16 +2504,14 @@ compute_color_rgb (QuantizeObj *quantobj,
{
/* Current algorithm: mean weighted by pixels (not colors) */
/* Note it is important to get the rounding correct! */
ColorFreq * histp;
int R, G, B;
int Rmin, Rmax;
int Gmin, Gmax;
int Bmin, Bmax;
long count;
long total = 0;
long Rtotal = 0;
long Gtotal = 0;
long Btotal = 0;
ColorFreq total = 0;
ColorFreq Rtotal = 0;
ColorFreq Gtotal = 0;
ColorFreq Btotal = 0;
Rmin = boxp->Rmin; Rmax = boxp->Rmax;
Gmin = boxp->Gmin; Gmax = boxp->Gmax;
@ -2420,24 +2520,24 @@ compute_color_rgb (QuantizeObj *quantobj,
for (R = Rmin; R <= Rmax; R++)
for (G = Gmin; G <= Gmax; G++)
{
histp = histogram + R*MR + G*MG + Bmin;
for (B = Bmin; B <= Bmax; B++)
{
if ((count = *histp++) != 0)
ColorFreq this_freq = *HIST_RGB(histogram, R, G, B);
if (this_freq != 0)
{
total += count;
Rtotal += ((R << R_SHIFT) + ((1<<R_SHIFT)>>1)) * count;
Gtotal += ((G << G_SHIFT) + ((1<<G_SHIFT)>>1)) * count;
Btotal += ((B << B_SHIFT) + ((1<<B_SHIFT)>>1)) * count;
total += this_freq;
Rtotal += ((R << R_SHIFT) + ((1<<R_SHIFT)>>1)) * this_freq;
Gtotal += ((G << G_SHIFT) + ((1<<G_SHIFT)>>1)) * this_freq;
Btotal += ((B << B_SHIFT) + ((1<<B_SHIFT)>>1)) * this_freq;
}
}
}
if (total != 0)
{
quantobj->cmap[icolor].red = (Rtotal + (total>>1)) / total;
quantobj->cmap[icolor].red = (Rtotal + (total>>1)) / total;
quantobj->cmap[icolor].green = (Gtotal + (total>>1)) / total;
quantobj->cmap[icolor].blue = (Btotal + (total>>1)) / total;
quantobj->cmap[icolor].blue = (Btotal + (total>>1)) / total;
}
else /* The only situation where total==0 is if the image was null or
* all-transparent. In that case we just put a dummy value in
@ -2856,7 +2956,6 @@ fill_inverse_cmap_rgb (QuantizeObj *quantobj,
int minR, minG, minB; /* lower left corner of update box */
int iR, iG, iB;
int * cptr; /* pointer into bestcolor[] array */
ColorFreq * cachep; /* pointer into main cache array */
/* This array lists the candidate colormap indexes. */
int colorlist[MAXNUMCOLORS];
int numcolors; /* number of candidate colors */
@ -2892,9 +2991,8 @@ fill_inverse_cmap_rgb (QuantizeObj *quantobj,
cptr = bestcolor;
for (iR = 0; iR < BOX_R_ELEMS; iR++) {
for (iG = 0; iG < BOX_G_ELEMS; iG++) {
cachep = & histogram[(R+iR)*MR+(G+iG)*MG+B];
for (iB = 0; iB < BOX_B_ELEMS; iB++) {
*cachep++ = (*cptr++) + 1;
*HIST_RGB(histogram, R+iR, G+iG, B+iB) = (*cptr++) + 1;
}
}
}
@ -3165,7 +3263,7 @@ median_cut_pass2_no_dither_rgb (QuantizeObj *quantobj,
R = (src[red_pix]) >> R_SHIFT;
G = (src[green_pix]) >> G_SHIFT;
B = (src[blue_pix]) >> B_SHIFT;
cachep = &histogram[R*MR + G*MG + B];
cachep = HIST_RGB(histogram,R,G,B);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)
@ -3246,7 +3344,7 @@ median_cut_pass2_fixed_dither_rgb (QuantizeObj *quantobj,
R = (src[red_pix]) >> R_SHIFT;
G = (src[green_pix]) >> G_SHIFT;
B = (src[blue_pix]) >> B_SHIFT;
cachep = &histogram[R*MR + G*MG + B];
cachep = HIST_RGB(histogram,R,G,B);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)
@ -3271,7 +3369,7 @@ median_cut_pass2_fixed_dither_rgb (QuantizeObj *quantobj,
G = (CLAMP0255(color->green + ge)) >> G_SHIFT;
B = (CLAMP0255(color->blue + be)) >> B_SHIFT;
cachep = &histogram[R*MR + G*MG + B];
cachep = HIST_RGB(histogram,R,G,B);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)
@ -3824,7 +3922,7 @@ median_cut_pass2_fs_dither_rgb (QuantizeObj *quantobj,
ge = g >> G_SHIFT;
be = b >> B_SHIFT;
cachep = &histogram[re*MR + ge*MG + be];
cachep = HIST_RGB(histogram, re, ge, be);
/* If we have not seen this color before, find nearest
colormap entry and update the cache */
if (*cachep == 0)