Updates:

• Recommendations can now use a new audio acoustic similarity-based engine (using MusicNN embeddings). The tag-based engine is still the default one, so you have to pick it in the scan settings:

Embedding extraction throughput depends heavily on both the number of scanner threads and the speed of your storage. Typical reported speeds range from ~1k tracks/hour on a Raspberry Pi 4 (3 threads) to ~25k tracks/hour on an Intel Core i5-13500 (10 threads). To speed up extraction, consider tweaking scanner-thread-count in lms.conf (defaults to half the number of logical CPUs) #301

• Associate user-provided images to playlists #833
• Made .lmsignore more like .gitignore, see README for more details #486
• OpenSubsonic API:
  • Added sonicSimilarity extension support #301
  • Added the groupings field #846
  • Added tokenInfo endpoint #838
  • Isolated the API key generation widget, added a copy button #796

Bugfixes:

• Fixed XSS when filtering collections with a maliciously forged tag #844
• Fixed unexpected termination when a podcast updates its image URL

Note to package maintainers: new optional dependency to the onnxruntime library. Beware of the new option musicnn-model-path in lms.conf. This dependency is mandatory for the audio acoustic similarity engine to work.

Updates

• Added an optional role filter for artist appearances #824

• Added basic checks to ensure directories set in lms.conf exist #828
• Relaxed the constraint on the ListenBrainz token format (no longer enforces UUID) #825
• Subsonic API:
  • Can now search for albums and artists using MBID #826
  • Added coverArt in DiscTitle responses

Bugfixes

• Fixed unwanted process termination when the client sends a bad request

Updates:

• Subsonic API:
  • Added Jukebox support #20
  • Added getNowPlaying support #792
  • Added readonly field in playlist response
• Added copyright information for each track #793
• Bumped package versions in the official docker image

Bugfixes:

• Fixed a regression causing DSub to stop working
• Fixed a regression related to the album artist rework #818

Note to package maintainers: A new optional dependency on the PulseAudio client library is required for the jukebox service to work.

Updates:

• Handle artist names as credited. The same artist can now be referenced using different names (useful for artists who have changed their name). #731
  Example:
  Example showing multiple credited artist names
• Subsonic API: the stream endpoint now accepts the FLAC format #797
• Improved accessibility for screen reader users
• Updated Chinese (zh) translations

Bugfixes:

• Fixed macOS compilation issues #801 #802

Note: you must rescan your library after updating.

Updates:

• OpenSubsonic API
  • Added support for the new Transcoding extension. #787
    Currently supported output transcoding formats: MP3, FLAC, Ogg/Vorbis, Ogg/Opus.
  • Added support for the Index-based Queue extension.
• UI
  • You can now apply codec as a global filter.
  • Global filters are no longer applied within album and artist views.
  • Added a new configuration parameter ui-allow-downloads in lms.conf. #776
• Other changes
  • Removed WebM/Vorbis and Matroska/Opus transcoding outputs.
  • Added support for parsing and playing Dolby AC-3 and Dolby Digital Plus (E-AC3) files. #781

Note: A library rescan is required for the codec-filtering feature to function correctly.

Bugfixes:

• Subsonic API: really fixed MusicAssistant compatibility #759

Updates

• Scanner: scan for metadata and embedded images in a single pass for each audio file
• Scanner: when an artist is referenced using multiple names, the scanner now prioritizes the name defined in the artist-info file, falling back to the most recent album artist link (#731). A full rescan is required to correct any artist-naming issues
• Subsonic API: always return http errors for media retrieval endpoints
• The lms-metadata CLI tool has been replaced by lms-audioinfo, a lower-level tool
• Fixed build with wt 4.12.1 #771

Bugfixes

• UI: fixed bad fallback when hitting only a secondary release type as used by picard #766
• UI: fixed audio sometimes not starting on mobile
• Scanner: fixed large label tags parsing (> 512 chars) #765
• Subsonic API: fixed MusicAssistant compatibility #759

Updates

• Updated installation instructions for Debian Trixie.
• Fixed artist.nfo support to work with the musicbrainzartistid tag (case-insensitive). #750
• Added support for matching fanart image files located next to artist.nfo files. #750
• Introduced a new setting, artist-info-image-file-names, to specify which artist image files to look up alongside artist.nfo files. #750
• Restored parsing support for .dsf and .dsffile formats. #748
• Improved XML output performance in the Subsonic API.

Updates

• Podcast support: currently using the Subsonic API, along with the getPodcastEpisode OS extension.
  See the new global settings in lms.conf to configure the podcast refresh period and episode auto-download parameters. #746
• Added support for the custom tag ARTISTSSORT.

Bugfixes

• OS API: fixed a typo in the explicitStatus field. #743
• Fixed an issue where embedded artwork was sometimes not detected. #718

Note to package maintainers: new dependency to pugixml

Updates

• Added the ability to associate external images with discs. See medium-image-file-names in lms.conf for the default file name used #699
• Added wildcard support for image file names set in lms.conf #697
• Added an "About" window to display the LMS version #428
• Added a button to star/unstar tracks #687
• Logging is now handled by LMS itself and no longer relies on WT (the web framework used). The access-log-file option in lms.conf has been removed

Important note: you must force scan your files after this upgrade:
image