Gamut compression - reduce artifacts resulting from out of gamut (#7205)
* First functions needs for ACES * New file GUI compressgamut * GUI first step * GUI first step * Gui step 2 * GUI procparams and paramsedit * GUI read * GUI step 4 * GUI step 5 * First tooltip * Gamut compression tooltips * Various GUI improvment * History msg * Comment code with Aces remarks * First change improccoordinator and events * Save work on matrix * Compress gamut next work * First try gamut compress * Replace cout by printf in invertmatrix * Change tooltips and events * Added namespace std to iplab2rgb * Comment code * Active rtthumbnail * Change tooltip * Various improvment GUI and rolloff * Added adobeRGB gamut * Appimage windows yml * Remove rttumbnail gamutcompr * Change event to COMPR * Change tooltip and verbose * Restore raw de-haze history message * Refactor ACES gamut compression functions * Fix gamut compression color space history message Correctly display the color space name. * Add "unchanged" for gamut compression color space Allow "unchanged" in batch mode. * Fix gamut compression yellow distance for batch * Update copyright for gamut compression GUI * Fix gamut compression color space names * Refactor gamut compression code * Remove comment rtthumnail.cc - Acesp1 default * Change matrix DCI-P3 - threshold maximum to 1 in GUI and in gamut compression - tooltips * Change tooltip * Clean code - change tooltips * Remove appimage windows yml --------- Co-authored-by: Lawrence Lee <45837045+Lawrence37@users.noreply.github.com>
This commit is contained in:
Desmis
GitHub
@@ -16,6 +16,10 @@
|
||||
* You should have received a copy of the GNU General Public License
|
||||
* along with RawTherapee. If not, see <https://www.gnu.org/licenses/>.
|
||||
*/
|
||||
#include <algorithm>
|
||||
#include <array>
|
||||
#include <utility>
|
||||
|
||||
#include <glibmm/ustring.h>
|
||||
|
||||
#include "rtengine.h"
|
||||
@@ -26,6 +30,7 @@
|
||||
#include "iccstore.h"
|
||||
#include <iostream>
|
||||
#include "linalgebra.h"
|
||||
#include "procparams.h"
|
||||
|
||||
using namespace std;
|
||||
|
||||
@@ -2105,6 +2110,181 @@ float Color::eval_ACEScct_curve(float x, bool forward)
|
||||
|
||||
// end code take in ART thanks to Alberto Griggio
|
||||
|
||||
//functions needs to use ACES
|
||||
|
||||
// transpose Matrix
|
||||
void Color::transpose(const Matrix &ma, Matrix &R)
|
||||
{
|
||||
if (&ma == &R) {
|
||||
std::swap(R[0][1], R[1][0]);
|
||||
std::swap(R[0][2], R[2][0]);
|
||||
std::swap(R[1][0], R[0][1]);
|
||||
std::swap(R[1][2], R[2][1]);
|
||||
std::swap(R[2][0], R[0][2]);
|
||||
std::swap(R[2][1], R[1][2]);
|
||||
} else {
|
||||
R[0][0] = ma[0][0];
|
||||
R[0][1] = ma[1][0];
|
||||
R[0][2] = ma[2][0];
|
||||
R[1][0] = ma[0][1];
|
||||
R[1][1] = ma[1][1];
|
||||
R[1][2] = ma[2][1];
|
||||
R[2][0] = ma[0][2];
|
||||
R[2][1] = ma[1][2];
|
||||
R[2][2] = ma[2][2];
|
||||
}
|
||||
}
|
||||
|
||||
// multiply Matrix x Matrix
|
||||
void Color::multip(const Matrix &ma, const Matrix &mb, Matrix &R)
|
||||
{
|
||||
const bool overwrite = &ma == &R || &mb == &R;
|
||||
if (overwrite) {
|
||||
// Use buffer to hold result so the input doesn't get overwritten while
|
||||
// the multiplication is happening.
|
||||
Matrix buf;
|
||||
multip(ma, mb, buf);
|
||||
for (int i = 0; i < 3; ++i) {
|
||||
for (int j = 0; j < 3; ++j) {
|
||||
R[i][j] = buf[i][j];
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for (int i = 0; i < 3; ++i) {
|
||||
for (int j = 0; j < 3; ++j) {
|
||||
double sum = 0.0;
|
||||
for (int k = 0; k < 3; ++k) {
|
||||
sum += ma[i][k] * mb[k][j];
|
||||
}
|
||||
R[i][j] = sum;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
//multiply Matrix
|
||||
void Color::mult3(std::array<float, 3> &in, const Matrix &ma, std::array<float, 3> &out)
|
||||
{
|
||||
// Use buffer for result in case in and out overlap.
|
||||
std::array<float, 3> buf{0.f, 0.f, 0.f};
|
||||
for (int i = 0; i < 3; ++i) {
|
||||
for( int j = 0; j < 3; ++j){
|
||||
buf[i] += static_cast<float>(in[j] * ma[j][i]);
|
||||
}
|
||||
}
|
||||
std::copy(buf.cbegin(), buf.cend(), out.begin());
|
||||
}
|
||||
|
||||
// ACES-style gamut compression
|
||||
//
|
||||
// tweaked from the original from https://github.com/jedypod/gamut-compress
|
||||
// tweaked from CTL in ART thanks to Alberto Griggio
|
||||
|
||||
//from ACES https://docs.acescentral.com/specifications/rgc/#appendix-c-illustrations
|
||||
// https://docs.acescentral.com/specifications/rgc/#appendix-d-ctl-reference-implementation
|
||||
// https://docs.acescentral.com/specifications/rgc/
|
||||
// Distance from achromatic which will be compressed to the gamut boundary
|
||||
// Values calculated to encompass the encoding gamuts of common digital cinema cameras
|
||||
//const float LIM_CYAN = 1.147;
|
||||
//const float LIM_MAGENTA = 1.264;
|
||||
//const float LIM_YELLOW = 1.312;
|
||||
|
||||
//Percentage of the core gamut to protect
|
||||
// Values calculated to protect all the colors of the ColorChecker Classic 24 as given by
|
||||
// ISO 17321-1 and Ohta (1997)
|
||||
//const float THR_CYAN = 0.815;
|
||||
//const float THR_MAGENTA = 0.803;
|
||||
//const float THR_YELLOW = 0.880;
|
||||
|
||||
// Aggressiveness of the compression curve
|
||||
//const float PWR = 1.2;
|
||||
//https://www.gujinwei.org/research/camspec/
|
||||
|
||||
void Color::aces_reference_gamut_compression(
|
||||
const std::array<float, 3> &rgb_in,
|
||||
const std::array<float, 3> &threshold,
|
||||
const std::array<float, 3> &distance_limit,
|
||||
const Matrix &to_out, const Matrix &from_out,
|
||||
float pwr, bool rolloff,
|
||||
float &R, float &G, float &B)
|
||||
{
|
||||
std::array<float, 3> rgb{rgb_in[0], rgb_in[1], rgb_in[2]};
|
||||
|
||||
// Calculate scale so compression function passes through distance limit:
|
||||
// (x=distance_limit, y=1)
|
||||
std::array<float, 3> s;
|
||||
for (unsigned i = 0; i < s.size(); ++i) {
|
||||
// Scale factor: c = (1 - t) / sqrt(l - 1)
|
||||
s[i] = (1.0f - threshold[i]) / sqrt(fmax(1.001f, distance_limit[i]) - 1.0f);
|
||||
}
|
||||
// target colorspace
|
||||
Color::mult3(rgb, to_out, rgb);
|
||||
|
||||
// Achromatic axis
|
||||
const float ac = fmax(rgb[0], fmax(rgb[1], rgb[2]));
|
||||
|
||||
// Inverse RGB Ratios: distance from achromatic axis
|
||||
std::array<float, 3> d{0.f, 0.f, 0.f};
|
||||
if (ac != 0) {
|
||||
for (unsigned i = 0; i < d.size(); ++i) {
|
||||
d[i] = (ac - rgb[i]) / fabs(ac);
|
||||
}
|
||||
}
|
||||
std::array<float, 3> cd{d[0], d[1], d[2]}; // Compressed distance
|
||||
if (!rolloff) {
|
||||
// Parabolic compression function:
|
||||
// https://www.desmos.com/calculator/nvhp63hmtj
|
||||
// y = { x < t: x
|
||||
// x >= t: c sqrt(x - t + c^2 / 4) - c sqrt(c^2 / 4) + t }
|
||||
// The second piece is equal to
|
||||
// c (sqrt(x - t + c^2 / 4) - |c / 2|) + t
|
||||
for (unsigned i = 0; i < cd.size(); ++i) {
|
||||
if (d[i] >= threshold[i]) {
|
||||
const float c_2 = s[i] / 2.f;
|
||||
const float c2_4 = c_2 * c_2;
|
||||
cd[i] = s[i] * (sqrt(d[i] - threshold[i] + c2_4) - fabs(c_2)) +
|
||||
threshold[i];
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for (unsigned i = 0; i < cd.size(); ++i) {
|
||||
if (d[i] >= threshold[i]) {
|
||||
if (threshold[i] == 1.f) {
|
||||
cd[i] = 1.f;
|
||||
} else {
|
||||
// Calculate scale factor for y = 1 intersect
|
||||
const float limit = distance_limit[i];
|
||||
const float thres = threshold[i];
|
||||
// l - t
|
||||
// Scale s = --------------------------
|
||||
// ( ( 1 - t )-p )(1 / p)
|
||||
// ( ( ----- ) - 1 )
|
||||
// ( ( l - t ) )
|
||||
const float scale = (limit - thres) / pow(pow((1.0f - thres) / (limit - thres), - pwr) - 1.0f, 1.0f / pwr);
|
||||
// Normalize distance outside threshold by scale factor
|
||||
// x' = (x - t) / s
|
||||
const float nd = (d[i] - thres) / scale;
|
||||
// x'
|
||||
// y = t + s ----------------
|
||||
// (1 + x'^p)^(1/p)
|
||||
const float po = pow(nd, pwr);
|
||||
cd[i] = thres + scale * nd / (pow(1.0f + po, 1.0f / pwr));
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
// Inverse RGB Ratios to RGB
|
||||
|
||||
for (unsigned i = 0; i < rgb.size(); ++i) {
|
||||
rgb[i] = ac - cd[i] * fabs(ac);
|
||||
}
|
||||
//working colorspace from_out
|
||||
Color::mult3(rgb, from_out, rgb);
|
||||
R = rgb[0];
|
||||
G = rgb[1];
|
||||
B = rgb[2];
|
||||
}
|
||||
|
||||
|
||||
void Color::primaries_to_xyz(double p[6], double Wx, double Wz, double *pxyz, int cat)
|
||||
{
|
||||
|
||||
Reference in New Issue
Block a user