* Use target_include_directories to specify include paths
* Specify project root (parent of rtgui and rtengine) as include path
* Replace relative includes with normal includes
* Speed up preview when inpaint opposed enabled
Only reprocess from raw if the white balance is changed. Otherwise, a
cache from later in the pipeline can be used.
* Remove unused code
* Fix refresh map bit positions
* Make WB & inpaint opposed refresh less brittle
Co-authored-by: Hombre57 <natureh.510@gmail.com>
---------
Co-authored-by: Hombre57 <natureh.510@gmail.com>
* Essai HL
* Try Inpaint opposed
* Code improvment
* Add file
* Improvment to process inpaint opposed and color propagation
* Clean code
* Change Blend to Coloropp in Profile pp3
* Enable BENCHFUN hilite_recon
* Clean rtengine cmakelist
* Comment unused code
* Neutralise unused code
* Change bad Exposure in Pop2Lab.pp3
* Try to fix bug when Inpaint Opposed is used and White balance disabled
* Changes to refreshmap
* Change to improccoordinator M_RETINEX
* Clean unused commented code
* Force Inpaint-opposed in rawimagesouce if wb change
* Suppressed message in console
* Change events and limits to 1 the number of calls to inpaint-opposed
* Comment code
* Add gain theshold to inpaint opposed
* fixed typo in procparams.cc
* Change in option.cc itcwb_sort to true
* Change itcw sorted in options and rawimagesource.cc
* Change sampling read datas Itcwb
* Allow or not purple in WB itcwb
* Added option icwb.nopurple to bypass settings
* Added code comment Itcwb
* optimize Itcwb between green and student
* Formated code used by Itcwb with Astylert.bat
* Change color_match - thanks to Lawrence37
* Remove wrong text
The JSON file is called workingspaces.json, it can be either in the global iccprofiles directory, or in the user's ICC profiles dir (set in preferences).
The format is the following:
{"working_spaces": [
{
"name" : "ACES",
"file" : "/path/to/ACES.icc"
},
{
"name" : "ACEScg",
"matrix" : [0.7184354, 0.16578523, 0.09882643, 0.29728935, 0.66958117, 0.03571544, -0.00647622, 0.01469771, 0.66732561]
}
]}
if "matrix" is present, "file" is ignored. If only "file" is present, the matrix is extracted from the ICC profile. For this, we look only at the R, G, and B matrix columns and the white point set in the profile. Bradford adaptation is used to convert the matrix to D50. Anything else (LUT, TRC, ...) in the profile is ignored.
It is the user's responsibility to ensure that the profile is suitable to be used as a working space.
User jgschaefer [reported an error on pixls.us](https://discuss.pixls.us/t/rt-build-from-git-crash-on-launch-debian-testing-64-bit/1425)
which could be traced down to an empty basename for a HaldCLUT. The
original implementation did not throw an exception due to the use of
`std::string::substr()` instead of `std::string::erase()`, but silently
assigned the first working profile to `profile_name`.
Ingo has provided a solution for the strange Windows crash with
`_mm_cvtpu16_ps()`: It was not an alignment problem, but the use of
MMX instructions which led to the SEGV.
Now Ingo's solutions omits MMX instructions altogether and is
nevertheless faster than the `_mm_set_ps()` workaround.
Many thanks to @heckflosse!
Ingo had some cleanup suggestions in #3154 which I tried to realize with
this commit. Although switching to `vfloat2` is a clever idea, I can see
no further speedup.
Instead of using an `Image16`, which is organized in planes, store the
HaldCLUT in an `AlignedBuffer<std::uint16_t>` with sequential RGBx
values. This gives a speedup of roughly 23% here.
This commit adds a true LRU cache to `rtengine` which is used in the new
`CLUTStore` class. The code in `clutstore.*` was cleaned up with C++11
features and small optimizations taken from my `clutbench` project.
The `CLUTStore` class was converted to a true singleton.
