SUMMARY *** NOTE: If you are reporting a crash, please try to attach a backtrace with debug symbols. See https://community.kde.org/Guidelines_and_HOWTOs/Debugging/How_to_create_useful_crash_reports *** STEPS TO REPRODUCE 1. Images on NAS 2. Open image from thumbnail 3. OBSERVED RESULT Loading of HEIC-images takes very long time EXPECTED RESULT I expect it to open close to instant. Video files .MOV opens instantly for playback. SOFTWARE/OS VERSIONS Windows: macOS: Linux/KDE Plasma: (available in About System) KDE Plasma Version: KDE Frameworks Version: Qt Version: ADDITIONAL INFORMATION
Which operating system did you use ? Note : HEIF decoding is delegate to libheif (https://github.com/strukturag/libheif) Can you share a sample file to test here ? Gilles Caulier
Well, if you look into libheif's bug reports and search for "slow", you'll find it. The loading time of digiKam does not differ from the implementations of other projects using libheif. We don't currently have an alternative to libheif on Linux. Maik
I was using windows 10, but tried on Linux and found it to be faster. Does DK use multi-thread or single-thread on windows?
DK use multi-thread everywhere. For HEIF loading, it use a separated thread to load image data (so to run the codec) Is the libheif/libde265 codec optimized under Windows ? Codec must use multi-thread to parallelize computations. Perhaps it's not the case under Windows. Again I suspect this problem located in libheif/libde265, not digiKam code. Gilles
In my win 11 installation 12MP HEIC images take up to 15s. I do not believe that this is better than on the Linux systems
@dajomu UPSTREAM bugs from libheif: https://github.com/strukturag/libheif/issues/645 https://github.com/strukturag/libheif/issues/552 https://github.com/strukturag/libheif/issues/472 https://github.com/strukturag/libheif/issues/133 Select right one and continue this discussion on github bugzilla from libheif. Nothing can be done in digiKam code to improve the performances. Gilles Caulier
*** Bug 441648 has been marked as a duplicate of this bug. ***
*** Bug 473356 has been marked as a duplicate of this bug. ***