Correcting metadata - it takes really way too long

No, I would’t go that far. It Just strikes me that there are still some basic functions that don’t work anywhere near what you would expect from premium software. I Just cannot believe someone considered the album identifying tool as being finished.

I’d be afraid about the Heisenberg uncertainty principle.

Don’t get me started…

The Fix Track Grouping and Track Identification (Let’s make sure everything is in order) tools are one of the weakest parts of Roon and something I have spent far too many hours painstakingly clicking on. I assume that the reason it hasn’t been replaced by something better is because of constraints on cross platform text handling in Roon’s graphic environment. Whatever the reason the current tools are extremely frustrating to use. Just replacing the click up/down buttons with grab bars would be a good start.


I’d rather it not be a battle. I think any sense of battle arises out of the “just listen to the music” crowd because it feels like being talked down to. It’s not with the Roon team.

I DO listen to the music. I’d match my collection and knowledge up to anyone. I want to slice and dice my playlists as I want so I can access the perfect track every time, and that requires some of these tools to be finished or to work better than they currently do. It’s not about the metadata per se.

As an example, do Radio and Shuffle work as well with unidentified albums? I suspect they do not and that the logic of the Radio and Shuffle algorithms is to some extent disrupted relative to unidentified albums. Yet the process of identifying them can be cumbersome to the point of practical unusuability.

Sure that would help. I also think a “make your best guess until I get around to correcting it” button would be helpful. If the title of the unidentified album is the same or very similar to an identified one by the same artists, then put that in as another version. Mark it as “partially identified” if the tracks still do not match up well enough.

Then also give us drop down lists of common items that cause ID confusion: vinyl rip, SACD or DVD-A Rip, Foreign release, bootleg, deluxe release, etc.

With those two options, actual track matching could be a second line of identification that only the real librarians might need to do.


It seems the trough of disillusionment has no end and the slope of enlightenment isn’t even a speck of light at the end of the trough.


hear, hear. I agree. Despite my griping, I have a lot of love and respect for the guys at Roon. I don’t like to be told what I can or can’t highlight by other Roon customers. Good call James.

really it needs a good brainstorm, maybe like the “your ideas for playlists are needed” thread which Roon posted a while back (did anything ever come from that, ahem? :wink: ). it’s a complex subject for sure, but like @andybob mentions, even grab-bars to begin with would be a big step in the right direction. I’ve said it before, but the process currently reminds me of this…

1 Like

I think the problem here is that the Roon concept is surely that you upload your Roon library and then Roon will autoidentify everything for you, and therefore manually editing data is not really needed. But of course the reality is that Roon will rarely identify everything so manual editing is required, but it is always going to be a secondary feature.

I’m familiar with this because the aim of my own SongKong tagger is to provide fully automated music tagging, but in reality with any library of any size there will always be at least 5% that doesnt get identified. So although there are many tools already available that do manual editing I have now added manual editing to SongKong, but it is not its primary function.

There are two main reasons why Roon will not identify your album:
1> Your album is not in the Music databases used by Roon
2> Your copy of the album is messed up/incomplete preventing a match

Considering 1>, there is alot to be said for adding the data directly into the database used by Roon. In particular MusicBrainz provides a very comprehensive database and editor and with a bit of practise you’ll find it quicker and easier to add a release into MusicBrainz then directly editing with Roon, then within a few days Roon should pick up the release. The beauty of this approach is that every release added to MusicBrainz can be ulitized by every Roon user with that album, so if all Roon users just added a few of their missing albums everybody would quickly get better matching rates. And the MusicBrainz database is really of the highest standard, there a few (mainly Classical) quirks but data depth is significant and there is rarely bad data, when there is bad data added because of the way data is monitored by other users it is corrected quickly.

2> Roon is not very good at identifying albums that are incomplete, or contain duplicate tracks ectera.
In this thread there is user request for SongKong to be added into Roon. But whatever tool you use the pragmatic approach is to process your metadata outside of Roon to give Roon the very best chance of getting matches, and to provide decent metadata that Roon can use to display/organize even if it cannot identify.

To summarize editing data within Roon is never going to be easy except for the most trivial of cases so get your data fixed outside of Roon, and then use Roon for controlling and playing the music.


The way I see it, the problem discussed in this thread is that Roon did add a feature to manually identify data, but it’s working subpar. Secondary feature or not, it is part of Roon and therefore should be on the same level as other Roon features, to make the app consistent.

(To me, by the way, it certainly isn’t a secondary feature, largely due to the fact that Roon is so weak when it comes to identifying classical music — properly tagged or not).


Sure you could argue it shoud be, but developers always have to prioritise and my feeling is this will never be a priority for Roon.

I couldn’t agree more :slight_smile:

Unfortunately, I think you’re right here too. And that’s exacty the point of this thread, IMO.

To add a bit to this: I think they’re making a mistake here. I don’t have any figures, but my guess is that since the audiophile community is a rather small niche, going forward Roon will attract more “ordinary” users, less tech savy and less occupied with tagging, making features like these more important. If these are the entrance to the Roon universe, the first impression will not quite be that of premium software…

1 Like

Good posts, and I agree on the fact that Roon is’t a full on editor and never will be I am fine with that and I am not e en asking for that either. All I am asking is that they look into the terrible way the identifying tool works. The track shuffeling is about the worst executed bit of UI I have ever seen. Why on earth would you make something so simpel so hard to use.

Besides that, what I would expect to see here is a link to the album location so I can open it in the OS file browser and from there edit the album in an external editor. As it is now it is a really hard workaround to get something corrected. Finding your album location is a hassle so any attempt to correct it in an external editor takes way too many steps.

I have many albums with some missing data like incorrect tracknumbers, should be easy to fix with the provided tool bit it is so hard to use to a point I completely gave up on getting the last 700+ albums recognized by Roon and that’a a real pitty.

The probleem is that even when you have corrected all your albums with an external editor there allways will be mismatches with the Roon database and you allways will have to come back to this tool as your only last chance.

So after all it is an essential part of Roon but is one that feels unfinisched and totally forgotten. It’s not o ly about functionality but about usability, and that is close to zero in my opinion.

As for the ‘just listen to the music guys’ According to Roon themselves the metadata is what you pay for, that’s what sets the high price of Roon. So why would 't I expect to able to make use of this metadata then? That what I signed up for in the first place. And I think it is more the reasonable to expect that it should not be this hard and overly complicatied to make use of this data. So I Just hope they first get rid of this terrible shuffeling tool because it simply does not work for many, I think even for most of us. At least I can not imagine there is anyone who likes working with it.

I have been looking at SonKong, looks like a very good tool but as long as it remains so overly complcated to get to the actual album directory fom within Roon it’s not of my interest. I am not interested to go through 5000+ album to manually find some that are recognized by Roon. I wan to be able to edit only the ones that need editing but I have to be able to get to them in a not so overly complicatie way.


For me this is not the key issue at hand. Missing albums are easily added to musicbrainz (if only they had an automated submission tool rather than that stupid web interface) and if you’re patient will eventually make its way into Roon unless Tivo gets there first.

The key issue at hand is that Roon’s primary metadata source (Tivo, which also feeds often serves crap metadata: typo’s, inconsistencies in track title and composer attribution across different versions of albums, often incomplete, incorrect or missing metadata for track title, composition and other credits, incorrect or no genre assignments, no discsubtitle for > 1CD albums and box sets etc…I could go on and on. As Roon insists on using track duration in its recipe for identifying albums Tivo’s shitty metadata causes albums to remain unidentified merely because Tivo have been too lazy to add track duration metadata. Same goes for track title. In turn Roon does nothing to present these mismatches in a way that you could easily dispense with them, no…everything is just one amorphous unidentified mass that you as the user must laboriously wade through one album at a time.

Then there’s the question of artist images, biographies which Tivo has also become increasingly poor at and Roon does little to improve this situation. I’d happily drop a bio in a text file in a Roon designated folder, named however Roon wants, which Roon can then pull and match to the corresponding artist where the artist has no bio - let’s face it, if you’re not talking the annals of American popular music Tivo isn’t exactly a rich source of artist information and it will probably never be. Rather than give users a way to enrich the metadata Roon seemingly prefers that we bumble along looking at grey screens with no artist bio or image. Don’t get me started on artist images, there are thousands of artists in my library with no artwork because Tivo doesn’t supply any and Roon hasn’t sourced any either. So, as a user, off I go again to re-perform the same thing over and over and over and over and over ad nauseam till I can’t take it anymore and quit. Again, a designated folder into which I can drop the artist images of interest, suitably named so that Roon knows who they are and can incorporate them would do the trick. But hey, why make it easy.

In short, whilst Roon relies on Tivo as its primary metadata provider the quality of the underlyling metadata is becoming poorer and making good takes way too much time and effort and in some instance e.g. artist bios we are simply shown the middle finger. This from a supposedly surfable magazine about our music. Roon then leverages this shitty metadata in preference to other sources to surface different performances of a composition, genre / style driven Radio etc. etc. and so the entire experience is diminished by its Achilles heel and its refusal to empower users to improve their experience.

One could easily extend that hype curve beyond 0 … the point at which customers lose interest. I hope for Roon’s sake that those customers who find themselves below the line iro expectations are lifers, otherwise the future may not look so rosy as it loses advocates and gains detractors which in turn makes prospective customers have second thoughts.

1 Like

A bit off topic:

If the album you’re going to add is available at another database (discogs for instance) or in one of the major web stores (like amazon, qobuz, bandcamp) chances are good that you can use a user script together with Greasemonkey or Tampermonkey which greatly helps adding new stuff to musicbrainz.

Also, there are user scripts which enhance the web interface to do things more easily.

No, it’s even more complicated . It’s like a combination of this shuffling puzzle and a memory game. Imagine you have to solve this puzzle but you can only see what’s on the tiles by flipping it one by one. That’s how the Roon tool works now. The big question is why does it have to be this hard?

Well maybe Roon thinks we’re such an old demographic these brain games will keep us mentally active :slight_smile:


Haha. I was trying to be kind.

Yep, they really should switch to making MusicBrainz primary source it is a much higher quality source and then spend a bit of money adding the albums that are in Tivo but not MusicBrainz.

I see your point that there seems to be no easy in Roon to get a list of folders not currently identified in Roon, but I would suggest you could just run SongKong on everything in one go. then if it does do anything that confuses Roon you could undo changes for those particular folders. However I realize that if most of your collection already looks okay in Roon you may be hesitant to do this.