Roon are listening attentively to everything that is going on in the Feature Requests forum. Please note though that they are also aware of (very) vocal minorities and idiosyncratic demands. By all means – keep this discussion going, but ‘pushing’ won’t help. Solid use cases (like the ones also presented here), preferably relevant to and supported by the broader community, do.
3 Likes
Thanks you made my day!
haha that’s scuppered it now… you’ll not get the same reaction 
And equally disgusting: getting a new computer with two drives and the tags with file paths don’t work anymore because they begin with D: instead if C:
EDIT: not referring to a Roon problem, Roon recognizes music if it gets physically moved. This was in reference to the suggestion of turning physical file paths into tags.
I agree. Roon is fairly tolerant of small mismatches in track timing, but if the metadata completely lacks track timing Roon gives up. This is not good. @mike
I know. I’m going to have to come up with another inflammatory title! 
1 Like
Essential contribution.
Thanks for the writeup on custom tags. It’s helped us to understand your use case a little bit better.
Part of the point here is: Roon’s Radio algorithm needs improvement (we know this, it’s being worked on).
BESTVERSION (maybe with PRIMARYVERSION as a synonym to match our terminology) is clearly something we should just do. Unfortunately your tagging scheme doesn’t group duplicates into sets–so the impact of that tag would be to influence who “wins” the set by default when Roon performs the grouping. There would still be some grooming to collapse things into groups properly, but once done, you could essentially toggle between “best version” and “all versions” by turning on/off Roon’s show-hidden-albums switch.
I don’t think we’d be averse to defining tags that allowed the grouping to be driven by that end, but I’d have to think through the implications further–how that mixes with the current duplicates determination–in order to be 100% sure.
The biggest issue I’m feeling is an uncertainty about how your tags are interpreted at the album vs track level (and if they are also interpreted in your world at artist level). BESTVERSION and SOURCEMEDIA seem to communicate album-level ideas, but what about the others?
I think I see how something like this could work for us. We have a big overhaul of focus coming, probably later this year. Might be the right point to fit it in. We’ll see.
We are displacing those products to an extent, but we are thinking bigger than that, too. We would not have much of a business if all we had were users of those three products you mentioned.
We are also not delusional–thinking that we will somehow displace Apple or Spotify or something like that–this is definitely still a product for a relatively small number of people who care more than the average person about music or audio.
People care about different things. Dedicated library organizers are a pretty small group–more people are doing their tweaking on the sound quality side–and even that group isn’t so large compared to sets of people like “anyone who has ever spent $300+ on a pair of headphones”, “anyone who has bought a HiFi component or AV receiver”, “anyone who acquires more than 10 albums per month”, “anyone who wants to DIY a whole-home-audio solution”–and we have something to offer to all of those people too.
The big struggle is keeping the product usable when adding power. We can’t promise whole home audio if the non-hobbyist spouse/kids can’t use it, or if a hifi dealer can’t sell it to their customers because it’s too complex to set up and use.
So in context–this complicated way of doing automatic DJ’ing that you’ve come up with is great for you–but even if we built this feature perfectly, and then you groomed the media perfectly to suit it, and set up all of the required stuff in Roon to support it perfectly, I could not take an iPad with your library and put it in front of another person and expect them to figure it out and use it.
4 Likes
Freud believed that an obsessively neat desk indicated a messy mind.
Sure, but that’s not the whole quote. The whole quote is “an obsessively neat desk indicates a messy mind…Darn, where is that album I just intended to listen to?” 
2 Likes
Thanks again Brian for such a thorough response. I do agree with this - I would not feel the need to take the reins on DJ’ing for myself to such a degree if Roon was better at surprising me in positive ways. What I’d really like to see here is the ability to choose whether to stick with “hits,” or “go deep” into an artist, genre, or other grouping, or to “go wide” to find related stuff, or deep and wide. A sort of adventurism slider type of thing, in multiple dimensions. Or at the very least, less predictability - Roon Radio sometimes feels like a top 40 station.
Just for completeness of response: MOODS and GROUPS are each recognized at the artist level, and SHUFFLE is at the album level. Of course, they are really all embedded in files at the track level, but uniform across those track files up to the foregoing levels, which then allows Foobar2000 filters to appropriate group them. .
That would be very helpful. Basically, if Focus could simply read and sort by any custom metatag field embedded in the file. This would be about an 80% solution to the needs discussed in this thread, which I view as pretty good, especially if the tweakers and librarians aren’t your core focus.
I’m sure that’s true - even when I have friends over and give a personal explanation, they almost always default to making playlists manually when I let them DJ. But it’s still handy for me to use these functions with guests because then I can focus on being social without worry that the music shuffle will go too far off the rails.
That said, it was definitely not my intent to propose that you make changes so that others could adopt my style of auto-DJing. My assumption is at a more general level, that there must be some significant subgroup of Roon users that in their prior media lives relied on custom embedded metadata for more advanced functions of their media software and that group would benefit from Roon having the ability to Focus on that custom data so as to be able to, in effect, “import” those functions into Roon by applying Roon tags as a substitute.
This is the other side of the coin from what I originally proposed above - that Roon export user-input media information so that users would not lose that effort, but really either is fine, because we can put that effort into the embedded tags and use the above process to import into Roon, thus preserving that scheme for use with other software as necessary (and that’s where we went down the rabbit hole debating the longevity of Roon, something I never intended). As long as it doesn’t exist only in Roon, I don’t care which direction it goes.
Hope I have been clear so as not to be seen pushing in a nonsensical direction (that’s a lob for those that want to quote me in their response…).
Thanks again, Brian. Roon, and the Roon team, are both great. It’s interesting the varying interests represented by this community - fun to see!
3 Likes
Thanks again Brian for your comments. I wanted to come back to you earlier, but I was on a business trip.
Just to be clear - I am not saying the “old” way is the better way. I will just be sticking to the old way in parallel as good as I can 
There will be things that can’t be mapped to tags if you really like to have a proper database for (classical) music, that’s without doubt.
Having the vision to offer such an experience with zero effort is certainly the right one. I am just a little bit nagging about the path to get there. If you offer content in a piece of software like Roon, the first question people like me will ask is: can I modify it or add to it? At the moment I can’t properly do this. I understand that it was your design decision to start like this.
However, it is sometimes really painful to see data inconsistencies and being not able to correct them yourself without unidentifying albums. There are redundant album and track credits all over the place and the tool to correct them seems to me not very practical for mass adjustments.
Just as an example I am currently fighting with my Arthur Rubinstein albums which have Arthur Rubinstein and Artur Rubinstein both as credits with links but only Arthur is available in the Artist browser so I can’t merge Artur and Arthur into one… It seems to me that I’d need to manually delete all Artur Rubinstein entries from my album credits in order to fix this. Which would be especially funny, since according to Wikipedia and my soloist tag it is Artur and not Arthur, so at the end I would need to manually edit the Arthur I have in my artist list…
I do think that you got the basic data model right and I just hope that in the future you will open the Roon database fields to editing.
For my personal experience of Roon I it sometimes would be better to have a global “off” switch for Roon to identify albums and start with a library of unidentified albums where the database would only be populated based on my tags. I would loose a lot of information Roon could provide me, but I could have consistent data and wait for Roon to make progress on the classical metadata front.
I’m trusting that you will make progress there but at the moment I’m still not ready to give up my JRiver library completely, because with all limitations and ugliness of JRiver I still find my stuff with one or two clicks and can maintain all the (admittedly limited) data.
Still I’d really want Roon to succeed and therefore I’ll support it with my license fee even if I’m not using it every time.
All I can say is after using Pure Music and Amarra, Roon blows both away in how I listen, view and read (like the ol’ days when you’d get a new album and read liner notes while playing it) my collection.
Plus all the info they give you on your source connection, resolution and audio path your files are being played from.
Never tried JRiver but really appreciate all the tremendous work the people at Roon have put in for this great software!!
Another plus is the instant lyrics that are provided on almost every song that has vocals…The other guys never did that as well.
5 Likes
I don’t think anyone on this thread has any intention of bashing Roon. If we weren’t heavy Roon users we wouldn’t care. The concern that caused me to start this thread is that the more I like Roon, the more I want to invest time in making my collection work with it, and making it work with my collection. That, however, is a big time investment, and I would like to find a way to leverage my existing time investment or be able to get that information back out to use with other applications.
Roon says that their goal is to make it so that I don’t have to invest that time. I just feel that’s a long-shot. Some of us, by our nature, have our own ways we want our music to come to us and are willing to put in that effort. I just don’t want to lose the effort.
4 Likes
Well said James_I
1 Like
Not quite true: I recently recovered some code I wrote at university in 1981 from a 1/2 track DEC tape. I’d run it if I could find a free PL/I compiler… That code was originally written on punch cards and run on an IBM System/360. To my dismay, the music catalog program I wrote, and the metadata I put on punch cards, was not on this tape. If it was, I could honestly say that the music metadata I started to capture in 1979 or so was still available and usable today in its original form. In fact I still have it, as I’ve migrated that data from punch cards to a hierarchical database to a flat file manager to Oracle to Access and on.
On the other hand, the work the OP put into his metadata stands a much better chance of existing in a usable form 20 years from now than my stuff did nearly 40 years ago. As roon takes on new corporate form in the future, his data will live on in some AWS S3 backup such even after he’s gone too.
Like some others, I would like to preserve work I do in roon outside roon as far as possible. If tags are insufficiently expressive, roon could export non-licensed data to sidecar files at the album or group or some higher level. Or they could publish their datastore specs or provide an API to developers to extract it and migrate it. But as @brian said in his first post, export needs to be more complete. In his second post, he argued for getting this right as opposed to doing it soon. As he noted, most users do extremely little metadata editing. And those that do find it cumbersome, as I have. My strategy so far has been to go back to other tools to fix data at the source if roon can’t figure something out and then re-import. If only the editing wasn’t buried behind so many clicks. At this point in the product’s life, it shows that these are secondary considerations. If roon ever provides full and efficient metadata editing, the product will need a different UI for it. Expert mode? Arg, please shoot me for saying that.
Thanks,
- Eric
I have come to believe that the best and easiest path is not that Roon needs to export metadata we configure, but rather Roon would be greatly improved if more capable on IMPORTING. (Although both would be great.)
The short of it is, Roon’s database has all my custom embedded metadata. I can see it in the file info view. It just doesn’t have any feature to do anything with it. If Roon’s Focus features included non-standard fields and the ability to Focus on a given value in a given field, that would go a long way to being able to then apply a Roon tag as a functional equivalent.
Then Roon just needs to add the ability to make its tags work more like filters if desired - I.e. Use tags to find the intersection of 2 subsets rather than adding subsets from tags together, although the minus symbol offers similar logic.
By way of explanation, the focus of my original post on exporting was due to my assumption I’d need to redo and repeat my file grooming within Roon. I didn’t want to lose all that work if I or Roon ever moved on.
Roon not having great tools for this (which is fine, plenty of apps for this exist) I’ve since decided it is better to use more standard file metadata tools to maintain the library and then to hope Roon someday soon supports more robust importing as above.