digiKam 8.3.0 is released

Dear digiKam fans and users,

After four months of active maintenance and long bugs triage, the digiKam team is proud to present version 8.3.0 of its open source digital photo manager.

This version arrives with the internal RAW decoder Libraw updated to the rolling-release snapshot 2024-02-02. Long time bugs present in older versions have been fixed and we spare a lot of time to contact users to validate changes in pre-release to confirm fixes before deploying the program in production.

The application internationalization has also been updated. digiKam and Showfoto are proposed with 61 different languages for the graphical interface. Go to Settings/Configure Languages dialog and change the localization as you want. Applications need to be restarted to apply changes. If you want to contribute to the internationalization of digiKam, please contact the translator teams, following the translation how-to. The statistics about translation states are available here.

Thanks to the translators who have worked on the online documentation internationalisations which is now available in 15 languages as French, German, Spanish, Italian, Japanese, Chinese, and more. You can read and search over the document here. You are welcome to contribute to application handbook translations following the coordination team instructions.

Windows Version Improvements and Stability

Since the 8.2.0, we left the cross-compiled version of the digiKam for Windows in favor of a native build using the open source Microsoft VCPKG tool-chain. Native Microsoft compiler is used to build the application which guarantees a better compatibility at run time.

In parallel we also switched the Windows version from Qt and KDE frameworks version 5 to the version 6. This big change has introduced some regressions and bugs which are now fixed with 8.3.0 for your pleasure. The fixes will gain the other platforms when the same migration will be done. This will guarantee a better stability of the project everywhere.

For the macOS and Linux bundles, they are still running with Qt version 5, but the migration is planned this year.

Internal Media Player Cleanup

With this release, we complete the transition to the QtMultimedia with Qt6 framework. For the Qt5 version, a new framework named QtAVPlayer is used. Both are based on FFMpeg backend to play video and audio contents, which prevents end users from adding closed-source extra codecs on the host operating system. The older unmaintained QtAV framework is completely dropped from the application. This has closed plenty of long-time bugs about the internal media player.

Note: QtAVPlayer has nothing to do with the older QtAV framework used in digiKam.

The Video-SlideShow plugin has also been improved. With the drop of QtAV previously used to encode and render video frames from images, the encoder have been ported to FFMpeg CLI tool and QtAVPlayer used to render the result.

New features arrive with this release to support audio tracks. Also, the tool allows you to encode time-lapse video, with frame luminosity equalization, and On Screen Display support to display metadata as date, comments, aperture, lens properties, etc.

Automatic Tagging Based on Contents

One plan that we have since a very long time is the way to perform auto-tags over collections by content parsing and topic detection. One student has worked since summer 2023 to implement this kind of feature based on a deep-learning engine with a pre-trained model. After the photo content analysis, the tool allows to detect forms, objects, places, animals, plants, monuments, scenes, and more. It generates a series of keywords parented with the Auto branch in the database. The user must review all new items pre-tagged by the engine to validate the contents detection.

Two new tools have been add in digiKam for these tasks: one in the Maintenance engine, and one other in Batch Queue Manager. Processing is done in the background on the available core of your computer. Detection is fully automatized, only the review of generated keywords must be done manually.

Tree deep learning models are provided to detect contents from images. They provide different results depending on the quantity of items to identify by image and the time consuming to parse the image with a result. More than one identification can be given for one image analyzed, In background, In the background, the C++ OpenCV Neural Network engine is used with the models.

You can read more details about the Auto-Tags project in the student blog.

Other Improvements and Add-ons

The Batch Queue Manager includes a new tool to import metadata from one image to a set of items. This kind of tool can be used to fill technical information or rights management to items where no data is present, as for example with scanned old photos. The source of data can be an image or a JSON file. The tool used in the background is ExifTool.

In the Setup Miscs view, new parameters have been added to configure a Network Proxy computer. These settings are used for all Network access to the Internet by the application, as for example to download map tiles from Open Street Maps or to push new data to the cloud web-services.

The Configuration dialog also gains a new entry to customize the Geolocation rules in the application. Tree tabs are available to tune:

    - The view contents, as the units, the quality, the cache, etc.    

    - The map plugins used to render contents and process data.    

    - The GoogleMaps API key for the credential rights to render the tiles without limitation from this web-service.    

Future Plans

Next maintenance version is planned to be published in february 2024 with more bug fixes and improvements.

We plan to port the Linux AppImage based on the last Qt and KDE frameworks version 6 as the Windows version.

The macOS version is planned to be ported to Apple Silicon. This is mostly done as all codes compile fine on this platform. Only the packaging section needs to be finalized.

Final Words

Thanks to all users for your support and donations, and to all contributors, students, testers who allowed us to improve this release.

digiKam 8.3.0 can be downloaded from this repository as:

• Source code tarball.
• Linux 64 bits AppImage bundles compatible with systems based on glibc >= 2.31.
• Windows 10 (or later) 64 bits installers or bundle archives.
• macOS Intel packages compatible with Apple Silicon CPU computers using Rosetta 2,

Rendez-vous in a few months for the next digiKam 8.4.0 maintenance release.

Have fun with digiKam…