liz/rawTherapee
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Finished implementation of Lanczos resize itself

Browse Source
This commit is contained in:
Ilia Popov
2010-10-28 01:07:49 +02:00
parent 52c822ba85
commit 6399e6d03e
1 changed files with 39 additions and 38 deletions
Download Patch File Download Diff File Expand all files Collapse all files

77
rtengine/ipresize.cc
View File

@@ -35,21 +35,35 @@ namespace rtengine {
#define CLIP(a) ((a)>0?((a)<CMAXVAL?(a):CMAXVAL):0)
#define CLIPTO(a,b,c) ((a)>(b)?((a)<(c)?(a):(c)):(b))
double Lanc(double x, double a)
{
if (x * x < 1e-6)
return 1.0;
else if (x * x > a * a)
return 0.0;
else {
x = M_PI * x;
return sin(x) * sin(x / a) / (x * x / a);
}
}
void ImProcFunctions::resize (Image16* src, Image16* dst) {
if(true) {
//if(params->resize.method == "Lanczos") {
double delta = 1.0 / params->resize.scale;
time_t t1 = clock();
if(params->resize.method == "Lanczos") {
const double delta = 1.0 / params->resize.scale;
const double a = 3.0;
const int support = 6;
const int kc = 2;
const double sc = std::min(params->resize.scale, 1.0);
const int support = (int)(2.0 * a / sc) + 1;
Image16 * tmp = new Image16(src->width, dst->height);
//#pragma omp parallel for
for (int i = 0; i < tmp->height; i++) {
// y coord of the center of pixel on src image
double y0 = (i + 0.5) * delta - 0.5;
int i0 = floor(y0);
// weights for interpolation in y direction
double w[support];
@@ -57,20 +71,17 @@ void ImProcFunctions::resize (Image16* src, Image16* dst) {
// sum of weights used for normalization
double ww = 0.0;
int ii0 = std::max(0, i0 - kc);
int ii1 = std::min(src->height, i0 - kc + support);
int ii0 = std::max(0, (int)floor(y0 - a / sc) + 1);
int ii1 = std::min(src->height, (int)floor(y0 + a / sc) + 1);
// calculate weights
for (int ii = ii0; ii < ii1; ii++) {
int k = ii - i0 + kc;
double z = M_PI * (y0 - (i0 + k - kc));
if (z * z > 1e-6)
w[k] = sin(z) * sin(a*z) / (a * z * z);
else
w[k] = 1.0;
int k = ii - ii0;
double z = sc * (y0 - ii);
w[k] = Lanc(z, a);
ww += w[k];
}
// normalize weights
for (int k = 0; k < support; k++) {
w[k] /= ww;
@@ -81,7 +92,7 @@ void ImProcFunctions::resize (Image16* src, Image16* dst) {
double r = 0.0, g = 0.0, b = 0.0;
for (int ii = ii0; ii < ii1; ii++) {
int k = ii - i0 + kc;
int k = ii - ii0;
r += w[k] * src->r[ii][j];
g += w[k] * src->g[ii][j];
@@ -94,10 +105,10 @@ void ImProcFunctions::resize (Image16* src, Image16* dst) {
}
}
//#pragma omp parallel for
for (int j = 0; j < dst->width; j++) {
// y coord of the center of pixel on src image
// x coord of the center of pixel on src image
double x0 = (j + 0.5) * delta - 0.5;
int j0 = floor(x0);
// weights for interpolation in y direction
double w[support];
@@ -105,17 +116,14 @@ void ImProcFunctions::resize (Image16* src, Image16* dst) {
// sum of weights used for normalization
double ww = 0.0;
int jj0 = std::max(0, j0 - kc);
int jj1 = std::min(tmp->width, j0 - kc + support);
int jj0 = std::max(0, (int)floor(x0 - a / sc) + 1);
int jj1 = std::min(tmp->width, (int)floor(x0 + a / sc) + 1);
// calculate weights
for (int jj = jj0; jj < jj1; jj++) {
int k = jj - j0 + kc;
double z = M_PI * (x0 - (j0 + k - kc));
if (z * z > 1e-6)
w[k] = sin(z) * sin(a*z) / (a * z * z);
else
w[k] = 1.0;
int k = jj - jj0;
double z = sc * (x0 - jj);
w[k] = Lanc(z, a);
ww += w[k];
}
@@ -129,7 +137,7 @@ void ImProcFunctions::resize (Image16* src, Image16* dst) {
double r = 0.0, g = 0.0, b = 0.0;
for (int jj = jj0; jj < jj1; jj++) {
int k = jj - j0 + kc;
int k = jj - jj0;
r += w[k] * tmp->r[i][jj];
g += w[k] * tmp->g[i][jj];
@@ -144,10 +152,7 @@ void ImProcFunctions::resize (Image16* src, Image16* dst) {
delete tmp;
}
return;
if(params->resize.method == "Downscale (Better)") {
else if(params->resize.method == "Downscale (Better)") {
// small-scale algorithm by Ilia
// provides much better quality on small scales
// calculates mean value over source pixels which current destination pixel covers
@@ -229,11 +234,8 @@ void ImProcFunctions::resize (Image16* src, Image16* dst) {
dst->b[i][j] = CLIP((int)b);
}
}
return;
}
if(params->resize.method == "Downscale (Faster)")
{
else if(params->resize.method == "Downscale (Faster)") {
// faster version of algo above, does not take into account border pixels,
// which are summed with non-unity weights in slow algo. So, no need
// for weights at all
@@ -302,9 +304,8 @@ void ImProcFunctions::resize (Image16* src, Image16* dst) {
dst->b[i][j] = CLIP( b * k / divider);
}
}
return;
}
if (params->resize.method.substr(0,7)=="Bicubic") {
else if (params->resize.method.substr(0,7)=="Bicubic") {
double Av = -0.5;
if (params->resize.method=="Bicubic (Sharper)")
Av = -0.75;