Often when browsing my collection I’ll come across what are clearly duplicates that I’d like to clean up but am unable to easily because these aren’t perfect tag matches (so they won’t appear in the duplicate scanner) and I don’t want to just archive/delete the duplicates from the main track browser as I don’t want to lose those tracks from any playlists that the duplicates were in.
Right now the only way to approach this would be to either:
Pick the version I want to keep, check the rest to see if they are in any playlists and then drag the version to keep to each of those playlists before archiving the duplicates
Edit all the title/artist tags to be perfect matches, then run the duplicate scanner again (This usually takes 20minutes to process) and then select the track to keep there
Neither is a particularly enjoyable workaround.
Two potential ways to improve this:
Create a way to replicate the same “replace all instances in playlists” functionality that the duplicate scanner does when removing duplicates in the main track browser. Only way I can imagine this in the current UI would be as a recipe where you select the set of duplicates, run the recipe, select the “master” track in the recipe options and then run it and any playlist that has any of the tracks in the recipe is updated and replaced with the master track
Have a way to run a mini duplicate scan on a smaller selection of tracks (specific playlist or selection could work) that has fuzzy search to allow it to find tag matches for close but not perfect matches like shown above. Such fuzzy matching wouldn’t work at all when dealing with a full library but when dealing with a small selection where there are known close matches you want to catch it should work pretty well.
There is a way to have the playlist replace behavior manually actually. Delete (or rename) the file on disk and then use Relocate and point it to a different file that is already inside Lexicon. Then Lexicon will ask you what you want to do. One option is to consolidate both tracks into one track, keeping your playlists intact. Good for single tracks but not very bulk friendly.
Fuzzy search is something I can add for sure.
I’m going to think about how I would make the scanner only use selected playlists/tracks. I understand the need and it has been asked before, especially useful for larger libraries. But I don’t really want to put it in the right click menu because then new users will think that is the normal way to scan for duplicates.
The 20 minutes for 64k tracks where the fingerprint is already cached is too long, I will try to speed that up by a few orders of magnitude I think if that is super fast then the order problem mostly disappears
I guess this is technically a workaround as well but to me, that sounds even less desirable than the two options I mentioned.
Yea I’d agree with not having it in the right-click menu for track selections. If it’s a playlist specific duplicate search though I don’t see too much issue with it going under the right-click menu on a playlist under “Tools”. At least in my own experience I use the right-click menu of a playlist very rarely, compared with the right-click menu on track selections.
Having the main duplicate scanner be super fast to get back to the results would certainly help reduce the friction there, though without the fuzzy matches (which I don’t think will be viable when scanning the whole library) there’s still some friction in having to manually change the artist/title tags to match so that the duplicate is detected by the primary scanner. And while I’d still do it, it doesn’t feel great having to edit tags on tracks that you’re ultimately looking to archive