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
.lmsignoremore like.gitignore, seeREADMEfor more details #486 - OpenSubsonic API:
- Added sonicSimilarity extension support #301
- Added the
groupingsfield #846 - Added
tokenInfoendpoint #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