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

Speedup for BadpixelsLab()

Browse Source
This commit is contained in:
heckflosse 2018-02-23 16:56:54 +01:00
parent cfbcd6cd5b
commit cde2796569
1 changed files with 50 additions and 54 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

104
rtengine/PF_correct_RT.cc
View File

@ -38,6 +38,7 @@
namespace rtengine
{
// Defringe in Lab mode
void ImProcFunctions::PF_correct_RT(LabImage * lab, double radius, int thresh)
{
BENCHFUN
@ -46,10 +47,9 @@ void ImProcFunctions::PF_correct_RT(LabImage * lab, double radius, int thresh)
chCurve.reset(new FlatCurve(params->defringe.huecurve));
}
// local variables
const int width = lab->W, height = lab->H;
//temporary array to store chromaticity
// temporary array to store chromaticity
const std::unique_ptr<float[]> fringe(new float[width * height]);
JaggedArray<float> tmpa(width, height);
@ -99,7 +99,7 @@ void ImProcFunctions::PF_correct_RT(LabImage * lab, double radius, int thresh)
float chparam = chCurve->getVal((Color::huelab_to_huehsv2(HH))) - 0.5f; // get C=f(H)
if (chparam < 0.f) {
chparam *= 2.f; // increased action if chparam < 0
chparam *= 2.f; // increased action if chparam < 0
}
chromaChfactor = SQR(1.f + chparam);
@ -128,7 +128,7 @@ void ImProcFunctions::PF_correct_RT(LabImage * lab, double radius, int thresh)
const int halfwin = std::ceil(2 * radius) + 1;
// Issue 1674:
// often, CA is not evenly distributed, e.g. a lot in contrasty regions and none in the sky.
// often, colour fringe is not evenly distributed, e.g. a lot in contrasty regions and none in the sky.
// so it's better to schedule dynamic and let every thread only process 16 rows, to avoid running big threads out of work
// Measured it and in fact gives better performance than without schedule(dynamic,16). Of course, there could be a better
// choice for the chunk_size than 16
@ -141,13 +141,13 @@ void ImProcFunctions::PF_correct_RT(LabImage * lab, double radius, int thresh)
int j = 0;
for (; j < halfwin - 1; j++) {
//test for pixel darker than some fraction of neighbourhood ave, near an edge, more saturated than average
// test for pixel darker than some fraction of neighbourhood ave, near an edge, more saturated than average
if (fringe[i * width + j] < threshfactor) {
float atot = 0.f, btot = 0.f, norm = 0.f;
for (int i1 = std::max(0, i - halfwin + 1); i1 < std::min(height, i + halfwin); i1++)
for (int j1 = 0; j1 < j + halfwin; j1++) {
//neighbourhood average of pixels weighted by chrominance
// neighbourhood average of pixels weighted by chrominance
const float wt = fringe[i1 * width + j1];
atot += wt * lab->a[i1][j1];
btot += wt * lab->b[i1][j1];
@ -161,13 +161,13 @@ void ImProcFunctions::PF_correct_RT(LabImage * lab, double radius, int thresh)
for (; j < width - halfwin + 1; j++) {
//test for pixel darker than some fraction of neighbourhood ave, near an edge, more saturated than average
// test for pixel darker than some fraction of neighbourhood ave, near an edge, more saturated than average
if (fringe[i * width + j] < threshfactor) {
float atot = 0.f, btot = 0.f, norm = 0.f;
for (int i1 = std::max(0, i - halfwin + 1); i1 < std::min(height, i + halfwin); i1++)
for (int j1 = j - halfwin + 1; j1 < j + halfwin; j1++) {
//neighbourhood average of pixels weighted by chrominance
// neighbourhood average of pixels weighted by chrominance
const float wt = fringe[i1 * width + j1];
atot += wt * lab->a[i1][j1];
btot += wt * lab->b[i1][j1];
@ -181,13 +181,13 @@ void ImProcFunctions::PF_correct_RT(LabImage * lab, double radius, int thresh)
for (; j < width; j++) {
//test for pixel darker than some fraction of neighbourhood ave, near an edge, more saturated than average
// test for pixel darker than some fraction of neighbourhood ave, near an edge, more saturated than average
if (fringe[i * width + j] < threshfactor) {
float atot = 0.f, btot = 0.f, norm = 0.f;
for (int i1 = std::max(0, i - halfwin + 1); i1 < std::min(height, i + halfwin); i1++)
for (int j1 = j - halfwin + 1; j1 < width; j1++) {
//neighbourhood average of pixels weighted by chrominance
// neighbourhood average of pixels weighted by chrominance
const float wt = fringe[i1 * width + j1];
atot += wt * lab->a[i1][j1];
btot += wt * lab->b[i1][j1];
@ -198,10 +198,11 @@ void ImProcFunctions::PF_correct_RT(LabImage * lab, double radius, int thresh)
lab->b[i][j] = btot / norm;
}
}
}//end of ab channel averaging
} // end of ab channel averaging
}
}
// Defringe in CIECAM02 mode
void ImProcFunctions::PF_correct_RTcam(CieImage * ncie, double radius, int thresh)
{
BENCHFUN
@ -212,10 +213,9 @@ void ImProcFunctions::PF_correct_RTcam(CieImage * ncie, double radius, int thres
chCurve.reset(new FlatCurve(params->defringe.huecurve));
}
// local variables
const int width = ncie->W, height = ncie->H;
//temporary array to store chromaticity
// temporary array to store chromaticity
const std::unique_ptr<float[]> fringe(new float[width * height]);
float** const sraa = ncie->h_p; // we use the ncie->h_p buffer to avoid memory allocation/deallocation and reduce memory pressure
@ -322,13 +322,12 @@ void ImProcFunctions::PF_correct_RTcam(CieImage * ncie, double radius, int thres
const int halfwin = std::ceil(2 * radius) + 1;
// Issue 1674:
// often, CA isn't evenly distributed, e.g. a lot in contrasty regions and none in the sky.
// often, colour fringe is not evenly distributed, e.g. a lot in contrasty regions and none in the sky.
// so it's better to schedule dynamic and let every thread only process 16 rows, to avoid running big threads out of work
// Measured it and in fact gives better performance than without schedule(dynamic,16). Of course, there could be a better
// choice for the chunk_size than 16
// Issue 1972: Split this loop in three parts to avoid most of the min and max-operations
#ifdef _OPENMP
#pragma omp parallel for schedule(dynamic,16)
#endif
@ -340,7 +339,7 @@ void ImProcFunctions::PF_correct_RTcam(CieImage * ncie, double radius, int thres
float atot = 0.f, btot = 0.f, norm = 0.f;
for (int i1 = std::max(0, i - halfwin + 1); i1 < std::min(height, i + halfwin); i1++) {
for (int j1 = 0; j1 < j + halfwin; j1++) {
//neighbourhood average of pixels weighted by chrominance
// neighbourhood average of pixels weighted by chrominance
const float wt = fringe[i1 * width + j1];
atot += wt * sraa[i1][j1];
btot += wt * srbb[i1][j1];
@ -360,7 +359,7 @@ void ImProcFunctions::PF_correct_RTcam(CieImage * ncie, double radius, int thres
float atot = 0.f, btot = 0.f, norm = 0.f;
for (int i1 = std::max(0, i - halfwin + 1); i1 < std::min(height, i + halfwin); i1++) {
for (int j1 = j - halfwin + 1; j1 < j + halfwin; j1++) {
//neighbourhood average of pixels weighted by chrominance
// neighbourhood average of pixels weighted by chrominance
const float wt = fringe[i1 * width + j1];
atot += wt * sraa[i1][j1];
btot += wt * srbb[i1][j1];
@ -380,7 +379,7 @@ void ImProcFunctions::PF_correct_RTcam(CieImage * ncie, double radius, int thres
float atot = 0.f, btot = 0.f, norm = 0.f;
for (int i1 = std::max(0, i - halfwin + 1); i1 < std::min(height, i + halfwin); i1++) {
for (int j1 = j - halfwin + 1; j1 < width; j1++) {
//neighbourhood average of pixels weighted by chrominance
// neighbourhood average of pixels weighted by chrominance
const float wt = fringe[i1 * width + j1];
atot += wt * sraa[i1][j1];
btot += wt * srbb[i1][j1];
@ -410,10 +409,11 @@ void ImProcFunctions::PF_correct_RTcam(CieImage * ncie, double radius, int thres
ncie->h_p[i][j] = xatan2f(interb, intera) / RT_PI_F_180;
ncie->C_p[i][j] = sqrt(SQR(interb) + SQR(intera));
}
} //end of ab channel averaging
} // end of ab channel averaging
}
}
// CIECAM02 hot/bad pixel filter
void ImProcFunctions::Badpixelscam(CieImage * ncie, double radius, int thresh, int mode, float chrom, bool hotbad)
{
BENCHFUN
@ -435,11 +435,11 @@ void ImProcFunctions::Badpixelscam(CieImage * ncie, double radius, int thresh, i
#endif
{
//luma sh_p
gaussianBlur(ncie->sh_p, tmL, width, height, radius / 2.0);//low value to avoid artifacts
gaussianBlur(ncie->sh_p, tmL, width, height, radius / 2.0); // low value to avoid artifacts
}
//luma badpixels
constexpr float sh_thr = 4.5f; //low value for luma sh_p to avoid artifacts
constexpr float sh_thr = 4.5f; // low value for luma sh_p to avoid artifacts
constexpr float shthr = sh_thr / 24.0f; // divide by 24 because we are using a 5x5 grid and centre point is excluded from summation
#ifdef _OPENMP
@ -449,7 +449,7 @@ void ImProcFunctions::Badpixelscam(CieImage * ncie, double radius, int thresh, i
#ifdef __SSE2__
const vfloat shthrv = F2V(shthr);
const vfloat onev = F2V(1.f);
#endif // __SSE2__
#endif
#ifdef _OPENMP
#pragma omp for
#endif
@ -587,9 +587,7 @@ void ImProcFunctions::Badpixelscam(CieImage * ncie, double radius, int thresh, i
}
}
}
}
// end luma badpixels
} // end luma badpixels
if (hotbad) {
JaggedArray<float> sraa(width, height);
@ -602,7 +600,7 @@ void ImProcFunctions::Badpixelscam(CieImage * ncie, double radius, int thresh, i
#ifdef __SSE2__
const vfloat piDiv180v = F2V(RT_PI_F_180);
#endif // __SSE2__
#endif
#ifdef _OPENMP
#pragma omp for
#endif
@ -628,7 +626,7 @@ void ImProcFunctions::Badpixelscam(CieImage * ncie, double radius, int thresh, i
float** const tmaa = tmL; // reuse tmL buffer
JaggedArray<float> tmbb(width, height);
if (mode == 2) { //choice of gaussian blur
if (mode == 2) { // choice of gaussian blur
#ifdef _OPENMP
#pragma omp parallel
#endif
@ -638,13 +636,13 @@ void ImProcFunctions::Badpixelscam(CieImage * ncie, double radius, int thresh, i
gaussianBlur(srbb, tmbb, width, height, radius);
}
} else if (mode == 1) { //choice of median
} else if (mode == 1) { // choice of median
#ifdef _OPENMP
#pragma omp parallel
#endif
{
#ifdef _OPENMP
#pragma omp for nowait //nowait because next loop inside this parallel region is independent on this one
#pragma omp for nowait // nowait because next loop inside this parallel region is independent on this one
#endif
for (int i = 0; i < height; i++) {
@ -843,6 +841,7 @@ void ImProcFunctions::Badpixelscam(CieImage * ncie, double radius, int thresh, i
}
}
// CbDL reduce artifacts
void ImProcFunctions::BadpixelsLab(LabImage * lab, double radius, int thresh, float chrom)
{
BENCHFUN
@ -862,26 +861,21 @@ void ImProcFunctions::BadpixelsLab(LabImage * lab, double radius, int thresh, fl
const std::unique_ptr<float[]> badpix(new float[width * height]);
if (radius >= 0.5) { // for gauss sigma less than 0.25 gaussianblur() just calls memcpy => nothing to do here
#ifdef _OPENMP
#pragma omp parallel
#endif
{
// blur L channel
gaussianBlur(lab->L, tmL, width, height, radius / 2.0);//low value to avoid artifacts
}
//luma badpixels
constexpr float sh_thr = 4.5f; //low value for luma sh_p to avoid artifacts
// for bad pixels in L channel we need 0 / != 0 information. Use 1 byte per pixel instead of 4 to reduce memory pressure
uint8_t *badpixb = reinterpret_cast<uint8_t*>(badpix.get());
constexpr float sh_thr = 4.5f; // low value for luma sh_p to avoid artifacts
constexpr float shthr = sh_thr / 24.0f; // divide by 24 because we are using a 5x5 grid and centre point is excluded from summation
#ifdef _OPENMP
#pragma omp parallel
#endif
{
// blur L channel
gaussianBlur(lab->L, tmL, width, height, radius / 2.0); // low value to avoid artifacts
#ifdef __SSE2__
const vfloat shthrv = F2V(shthr);
const vfloat onev = F2V(1.f);
#endif
#ifdef _OPENMP
#pragma omp for
@ -898,7 +892,7 @@ void ImProcFunctions::BadpixelsLab(LabImage * lab, double radius, int thresh, fl
shmed += std::fabs(lab->L[i1][j1] - tmL[i1][j1]);
}
}
badpix[i * width + j] = shfabs > ((shmed - shfabs) * shthr);
badpixb[i * width + j] = shfabs > ((shmed - shfabs) * shthr);
}
#ifdef __SSE2__
@ -912,7 +906,11 @@ void ImProcFunctions::BadpixelsLab(LabImage * lab, double radius, int thresh, fl
shmedv += vabsf(LVFU(lab->L[i1][j1]) - LVFU(tmL[i1][j1]));
}
}
STVFU(badpix[i * width + j], vselfzero(vmaskf_gt(shfabsv, (shmedv - shfabsv) * shthrv), onev));
uint8_t mask = _mm_movemask_ps((vfloat)vmaskf_gt(shfabsv, (shmedv - shfabsv) * shthrv));
badpixb[i * width + j] = mask & 1;
badpixb[i * width + j + 1] = mask & 2;
badpixb[i * width + j + 2] = mask & 4;
badpixb[i * width + j + 3] = mask & 8;
}
#endif
for (; j < width - 2; j++) {
@ -924,7 +922,7 @@ void ImProcFunctions::BadpixelsLab(LabImage * lab, double radius, int thresh, fl
shmed += std::fabs(lab->L[i1][j1] - tmL[i1][j1]);
}
}
badpix[i * width + j] = shfabs > ((shmed - shfabs) * shthr);
badpixb[i * width + j] = shfabs > ((shmed - shfabs) * shthr);
}
for (; j < width; j++) {
@ -936,7 +934,7 @@ void ImProcFunctions::BadpixelsLab(LabImage * lab, double radius, int thresh, fl
shmed += std::fabs(lab->L[i1][j1] - tmL[i1][j1]);
}
}
badpix[i * width + j] = shfabs > ((shmed - shfabs) * shthr);
badpixb[i * width + j] = shfabs > ((shmed - shfabs) * shthr);
}
}
}
@ -948,12 +946,12 @@ void ImProcFunctions::BadpixelsLab(LabImage * lab, double radius, int thresh, fl
for (int i = 0; i < height; i++) {
int j = 0;
for (; j < 2; j++) {
if (badpix[i * width + j]) {
if (badpixb[i * width + j]) {
float norm = 0.f, shsum = 0.f, sum = 0.f, tot = 0.f;
for (int i1 = std::max(0, i - 2); i1 <= std::min(i + 2, height - 1); i1++) {
for (int j1 = 0; j1 <= j + 2; j1++) {
if (!badpix[i1 * width + j1]) {
if (!badpixb[i1 * width + j1]) {
sum += lab->L[i1][j1];
tot += 1.f;
const float dirsh = 1.f / (SQR(lab->L[i1][j1] - lab->L[i][j]) + eps);
@ -971,12 +969,12 @@ void ImProcFunctions::BadpixelsLab(LabImage * lab, double radius, int thresh, fl
}
for (; j < width - 2; j++) {
if (badpix[i * width + j]) {
if (badpixb[i * width + j]) {
float norm = 0.f, shsum = 0.f, sum = 0.f, tot = 0.f;
for (int i1 = std::max(0, i - 2); i1 <= std::min(i + 2, height - 1); i1++) {
for (int j1 = j - 2; j1 <= j + 2; j1++) {
if (!badpix[i1 * width + j1]) {
if (!badpixb[i1 * width + j1]) {
sum += lab->L[i1][j1];
tot += 1.f;
const float dirsh = 1.f / (SQR(lab->L[i1][j1] - lab->L[i][j]) + eps);
@ -994,12 +992,12 @@ void ImProcFunctions::BadpixelsLab(LabImage * lab, double radius, int thresh, fl
}
for (; j < width; j++) {
if (badpix[i * width + j]) {
if (badpixb[i * width + j]) {
float norm = 0.f, shsum = 0.f, sum = 0.f, tot = 0.f;
for (int i1 = std::max(0, i - 2); i1 <= std::min(i + 2, height - 1); i1++) {
for (int j1 = j - 2; j1 < width; j1++) {
if (!badpix[i1 * width + j1]) {
if (!badpixb[i1 * width + j1]) {
sum += lab->L[i1][j1];
tot += 1.f;
const float dirsh = 1.f / (SQR(lab->L[i1][j1] - lab->L[i][j]) + eps);
@ -1016,9 +1014,7 @@ void ImProcFunctions::BadpixelsLab(LabImage * lab, double radius, int thresh, fl
}
}
}
}
// end luma badpixels
} // end luma badpixels
float** const tmaa = tmL; // reuse tmL buffer
JaggedArray<float> tmbb(width, height);
@ -1032,7 +1028,7 @@ void ImProcFunctions::BadpixelsLab(LabImage * lab, double radius, int thresh, fl
gaussianBlur(lab->b, tmbb, width, height, radius);
}
// begin chroma badpixels
// begin chroma badpixels
double chrommed = 0.0; // use double precision for large summations
#ifdef _OPENMP