roon identified it. but with the wrong version. the product and catalog numbers refer to another edition in 3 CD. Mine is the P.Glass edition in 4 CD.
ok, they are probably exactly the same recording… so no harm. but i am really concerned about the fact that roon identifications are not reliable at all.
one question to the developers.
is there any way to force roon association? i don’t know, is the tag UPC (or CATALOGNUMBER) effective in any way?
Niccolo, you could try splitting the discs into individual items. I’ve occasionally had a boxset that gets better identified as a collection of individual albums. Hope that makes sense.
You could also try running them through a music app first. MusicBrainz has free software that attempts to recognize discs and tracks and boxsets. I’ve started routinely using it prior to importing into Roon. It gives a fairly easy way to mass edit track tags that may result in better association.
Know this: mis-taggings abound – on original CDs, in “reference” DBs, and even the quality of the labels’ source data is variable. Roon is as good as any I’ve used to tease the correct titles out of the morass. But it is a dark art.
I think this is an area of emphasis by Roon, so stand by.
The problem with this is that it will make Roon lean towards MusicBrainz metadata rather than the fare superior Tivo/AllMusic data. Using the Allmusic artist/album fields in file tags has the opposite effect of pushing Roon towards the best possible metadata.
I’m sorry your MB-to-Roon’s ID process didn’t work for you, assuming you’ve tried it. My results have been quite good. Having well-articulated track titles with opus numbers, and track-level Composer(s), and track level artists, is a universal good. It has never led Roon astray in my experience. It has never done me harm.
John, I would generally agree that the MusicBrainz data is better than Tivo/AllMusic
But there is one issue, for Classical Music MusicBrainz stores the Track artist as the Composer, whereas I think it should be the conductor/main performer.orchestra/choir depending on the track credits on the album. Therefore when taking data from MusicBrainz the tool needs to recognize that it is tagging a Classical release and make adjustments, and instead of using the track artist make use of the tracks relationships.
We at Roon also disagree with this policy, but it’s more a problem for two specific reasons:
The metadata are not consistent: a) The classical style guidelines have evolved over time; b) data entry not conforming to the style guidelines does get past the style police.
Although one can make intelligent programmatic guesses, there is no flag (that I’m aware of) in the MusicBrainz metadata schema to indicate that the release is classical and that the metadata (therefore) follow different guidelines.
Exactly, and I spent some time trying to convince of the need for this. The main stumbling block was that there are some releases where it is rather subjective whether a release is classical or not, but since the same decision has to be made when deciding to use Classical style guidelines or not I don’t see this is such an issue. But I did raise an issue about simply marking if a release uses the Classical style guidelines and hopefully one day this will get done https://tickets.metabrainz.org/browse/MBS-9020
My way round it with SongKong was to develop a heuristic algorithm to identify if a release is Classical, with a way to override for a particular release for those cases where the algorithm doesn’t get it right.