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

profile-based CA correction applied during 1:1 preview

Browse Source
This commit is contained in:
Alberto Griggio
2017-10-29 21:41:19 +01:00
parent 830fd6fdbb
commit 5a05b04e09
6 changed files with 29 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
rtengine/rtlensfun.cc
View File

@@ -76,7 +76,7 @@ bool LFModifier::isCACorrectionAvailable() const
}
void LFModifier::correctCA(double &x, double &y, int channel) const
void LFModifier::correctCA(double &x, double &y, int cx, int cy, int channel) const
{
assert(channel >= 0 && channel <= 2);
@@ -84,6 +84,8 @@ void LFModifier::correctCA(double &x, double &y, int channel) const
// lensfun applies it to all the three channels simultaneously. This means
// we do the work 3 times, because each time we discard 2 of the 3
// channels. We could consider caching the info to speed this up
x += cx;
y += cy;
float pos[6];
if (swap_xy_) {
@@ -95,6 +97,8 @@ void LFModifier::correctCA(double &x, double &y, int channel) const
if (swap_xy_) {
std::swap(x, y);
}
x -= cx;
y -= cy;
}