A while ago I made a lot of speedups for tmos used in LHDR. My fork currently is 35 commits ahead of LuminanceHDR:master, most of the commits being speedups.
Here’s an overview, measured using a 24 MPixel file on my rusty AMD FX8350 with 32 GB Ram:
master my fork memory usage reduced by
Drago 11.8 s 0.3 s
Ferradans 90.0 s 72.0 s 80 * width * height byte
Reinhard02 scales disabled 4.4 s 0.2 s 16 * width * height byte
Reinhard02 scales enabled 53.5 s 9.2 s 16 * width * height byte
Pattanaik local tonemapping off 60.6 s 1.5 s
Pattanaik local tonemapping on 299.3 s 6.5 s
Ashikhmin Eqn 2 16.1 s 2.0 s 4 * width * height byte
Ashikhmin Eqn 4 20.4 s 2.0 s 4 * width * height byte
Reinhard05 23.2 s 0.6 s
Debevec hdr creator 9.8 s 0.8 s min(i, c) * 24 * width * height byte, where i is the number of input images and c is the number of cores
Thanks for information
I was surprised because I never noticed the requirement for cfitsio in my previous builds and I don’t remember having installed that in MSYS2
As @Hombre and I now got push access to the official Luminance HDR repository we merged a lot of stuff into the master branch.
As I don’t need my fork anymore, I will delete it soon. So please don’t build from it. All the speedups and other stuff now are in official master branch.
I have compiled an earlier version of LuminanceHDR, and I need to use the luminance-hdr-2.5.0-qtwebkit.patch patch file as version of linux does not have QtWebEngine.
This patch file does not not seem to work with your current version and does not seem to have been updated for 11 months. Or maybe I am running it incorrectly.
/test/LuminanceHDR-master$ patch -p1 < luminance-hdr-2.5.0-qtwebkit.patch
patching file CMakeLists.txt
Hunk #1 FAILED at 34.
Hunk #2 FAILED at 53.