Detect duplicate tracks by name (or other means) in playlists and bulk remove

This would be extremely useful!!!! Please add this feature!

Related to this post in Features & Feedback.

An option in the player where you can see if you added that song to any of your playlist. You will see the name of those playlist in there.

I love making playlist. I have a playlist where I put songs that I enjoy and would like to save in an specific playlist. When I know in what playlist belongs, I erase that song from that temporal playlist, and put it in the right place. The problem is, sometimes I can't remember witch song I already saved, I don't have a great memory. I could put those songs with a favorite, but I would fill my favorite limit too quickly.

I think no other steaming service have this feature, it would be really useful and attractive to users.

Deezer could use a feature to detect identical tracks in your favourites, so users could avoid cluttering their favourites by adding the exact same track multiple times albeit from different sources (‘best of’ albums / playlists etc). At the moment, the only way is to sort favourites alphabetically and manually delete duplicates.

Hey @alansheahan97 

Thanks a lot for your support. Moved your comment here, please vote for the idea above 

Hi @nemterest 

Moved your comment here too, these enhancements are looked into by the same teams and the good news is that they're looking for new features to add to playlists at the moment. Stay tuned and don't forget to vote above too. I've forwarded what you said 

I'm so glad this was raised. I have literally just joined Deezer having being with Spotify for over 4 years, I just cannot believe this feature isn't on the service, so far its let me download the exact same 3 songs onto a playlist, only difference...3 different albums, even the really old  versions of iTunes had this feature. I think I'll keep my Spotify going a while until I decide which is better, but to not have a simple feature like this...

Coming here from another thread and (mostly) pasting some thoughts I shared with @Rudi in private messages recently. Given the current status of this idea (“Not for now”), I have little hope to see anything done towards solving the problem anytime soon, but maybe the following will give someone some food for thoughts.

A proper automated system would detect duplicates many times faster than any human would be able to, so reporting problems on a case by case basis is a huge waste of time, both the users’ and the Deezer’s team, whose time would probably be better spent automating this thing.

While I don't know exactly what kind of metadata you keep about every song, the metadata that is visible in the user interface would be quite lacking for a recommender system. What I'd probably expect in a decent database based on my experiences with community-curated content is:

• linking all exact duplicates of a song, and selecting a preferred copy for the purposes of recommendations (nothing wrong with having multiple copies if they were published multiple times, but if they’re exact copies, the recommender engine should know about it),
• a list of strongly related songs (e.g. covers, remixes, jazz standards, live vs. studio vs. acoustic performance, etc., though storing the specific nature of a relation is probably not necessary—just stating the fact that two songs are similar would be enough); this list may refer to works outside of Deezer's catalogue where proper,
• linking all easily identifiable artists to each work (probably at least the composer and all vocalists in addition to the band name, and a music director for works like soundtracks—that’d do for popular music; pretty sure criteria for classical music would be somewhat different),
• maybe a list of related non-music works (e.g. a TV series or a game where the song was promoted), though that information can usually be derived from the album the song appeared in.

All of these information would be both useful to explore music (e.g. a user looking for songs with a specific vocalist or different renderings of a specific music theme), and to avoid boring recommendations (e.g. too many songs by a given composer within a single day). I’ve seen all of these working quite well for a community-curated internet streaming service, though at a smaller scale, where the engine had some nice tunable constraints, like “do not play the same song sooner than in 5 days, do not play any strongly related song sooner than in 2 days, do not play the same artist (not just “the same band”) sooner than in 6 hours, do not play more than three Touhou songs in a single hour (that one was extremely useful…)”, etc.

Given the scale of operation, Deezer is not really in position to fill in all that metadata without automation or coöperation with users that is much more streamlined than just a basic discussion forum.

The first and second bullets could be semi-automated using music similarity tooling (basic name matches will have more false positives and more false negatives). The third and fourth bullets could start from being managed by users as a form of a plain old web2.0 tag cloud, probably provided as an opt-in experimental feature to avoid it being visible to the more casual users, with the Deezer team stepping in to moderate/verify/publish them only when they amass.

I’ve been attacked by Kyu Sakamoto today again and again. Turns out, there’s *a lot* of duplicates:

I frankly wonder whether Deezer’s claim of so many millions of songs isn’t just false advertising falling under unfair competition laws.

This is sooo much needed!

Bumping it up!

Just used Tune My Music to merge all the services I’ve used in the past, and now have to clean a sh*tload of duplicates, slightly different versions, and “unavailable in your country” songs in my Favorites, would be awesome to have a feature to at least speed up some parts of this process 

Thank you @Lucas Otanari for letting us now!