Why unidentified compositions in unidentified albums?

My impression is that how this identification is done is not fixed (enough? yet?) and therefore the Roon team is reluctant to give out a “definitive guide”. The more probable identifiers a composition within an unID’d album has the higher the likelihood of a match (obviously) - Joel mentioned opus and work register as well as “canonical naming” - but that’s just working in the context of classical music. If opening up the system for composition and artist IDing is an option Roon should tell us - in the cases often discussed here it might help but there may be good reasons not to do it, too.

Yes, unfortunately none of us know what the identifiers are. But there is enough here to see that for unidentified compositions a corresponding match in an identified album seems to be some kind of “hard” stop. Maybe in the light of experience that hard stop should be relaxed in some way. I would like to see much higher rates of composition identifications in unidentified albums so it is only logical that will require some kind of modification of the “weightings” given to the current criterea. There may well be many ways of doing that and good reasons why some methods won’t work.

Late arrival here; caught the jist, but pardon if I missed something.

Tony, your above statement isn’t accurate. I have numerous un-IDed albums where all the compositions are linked.

There are THREE requirements for composition IDs:

  1. Canonical names a la Allmusic (and nothing else) are entered. I now habitually put them in the WORK tag with a third-party tagger such as Yate.
  2. There must be metadata to link to.
  3. There must be an “exemplar” album to triangulate on. I think it is this third requirement that trips me up most. Fortunately, TIDAL has such albums that, when added to my library, will complete the link-up to composition metadata.
  4. EVEN IF your album is IDed, there is no assurance that all or any compositions will be linked. If this happens, I go back to steps 1-3.

Third-party tagging is THE key, IMO. You can hope Roon will recognize your stuff, and Roon may deliver, but it is a hit-or-miss affair, and I got tired of asking myself, “why isn’t this composition showing up”. So I now always check the underlying tags first thing. And my success rate is about 98%.

The price for this success rate: time. About 5-10 minutes/album. Fortunately, I have gobs of time now; many don’t. For those, just repeat these words, mantra-style:


That post was some time a go. The context at the time was with an unidentified album there are two possibilities:

  1. Album 1 is unidentified and has a composition C. There exists at least one identified album 2 with the same composition C.

  2. Album 1 is unidentified and has a composition C. There does not exist at least one identified album 2 with the same composition C.

With scenario 1) it is possible to get a composition identification in an unidentified album. With scenario 2 it is not. I would like to see it possible to make composition identifications under scenario 2 as well. Maybe that’s changed? I don’t think so but I am away from home and cannot make a quick check with the limited library I have on my travel laptop.

Situation 1 is still correct as far as I have seen.

There is a bit of a practical workaround for scenario 2 that I have used. This is similar to what @John_V proposed, but doesn’t require Tidal

  1. Make sure the Target unidentified album has Rovi canonical names in work/part (Zarathustra, for the sake of argument)
  2. Choose a Sacrificial unidentified album in your collection (e.g. Greatest Hits of Æolian Nose Whistle)
  3. Deliberately mis-identify this Sacrificial album as an album that contains the desired Target composition in question (i.e. search for any Zarathustra album and choose one, preferably something with a famous conductor/performer that you guess might have good metadata)
  4. Your Target still-unidentified Zarathustra album should now have its composition linked to the counterpart in the sacrificial (purposely incorrectly) identified album. If not, you can go to the Composer page for that composition, and merge compositions with the newly (even though falsely) identified entry.
  5. You can now remove the incorrect identification from the Sacrificial album, and Roon will actually preserve the composition link for the still-unidentified original Target album in question.

Nice. I’ll give it a whirl. I don’t have a Tidal account.

I use a very similar trick all the time to create an equivalent performer temporarily so that the equivalents can be merged on the artist page. I seem to remember that was also your tip?

I’m also a Tidal holdout. The vast majority of my library comes from CD rips.

I’ve used this in the past for Artists, Composers, and other roles as well. I can’t remember if I suggested the method on the site previously or not.

I see your assertion and agree. There was talk of opening access to “exemplars” (as I call them) without having to own or import them into your library. But then, there’s been a lot of talk on this subject. So I’ll hush now. :slight_smile:

Hi, Jim @JWC ,

Thanks for this tip. Just tried it out and it works! It’s a lot of work, though, so if I’ll be using it much remains to be seen.
Much simpler solution would be for Roon to open up the possibility to choose a composition yourself, by way of a “Composition search screen” of some sort…

Great to see the old advice is still useful, if cumbersome.

I fully agree that having a way to Identify Compositions by search (similar to how Albums are identified) would be a great step forward.

Until then…