After a few weeks of cleaning up the tags on my files, Roon is now able to identify a good number of the albums.
With pop and jazz, Roon is doing a fine job.
However with classical music, I am finding that Roon’s metadata is quite inconsistent, when it comes to artist, conductor, orchestra, leading to major issues, how the albums are sorted in the different views.
For example, my albums by the German conductor Günter Wand:
Listed under Günter Wand:
Beethoven Symphony Nr. 9
Bruckner Symphony Nr 4
Bruckner Symphony Nr 6
Bruckner Symphony Nr 7 [Japan]
Brahms Symphonies Nr 1-4
Listed under Günter Wand / Berlin Philharmonic Orchestra
Bruckner, Symphony No 8
Listed under Bruckner; NR Elbphilharmonie Orchestra, Günter Wand
Bruckner Sinfonie Nr 8
Listed under Bruckner; Berlin Philharmonic Orchestra, Günter Wand
Bruckner, Symphony no. 5
Sometimes dates and locations are given for the performances, sometimes not.
Could this type of inconsistency (which is rampant with almost all conductors / orchestras be corrected?
It is really standing in the way of making Roon a serious tool for the classical music listener.
I am sure there is an object oriented database behind the displayed data that properly and consistently lists
work
movement
conductor
orchestra
soloist(s)
recording date
remastering date
for each object. From this data it should be easy to construct a consistent display in the user interface.
I was wondering, whether anyone of the Roon staff has been able to look at this?
I have quite a lot of metadata issues that I would like to report back to Roon. Is there any way to ‘bulk’ communicate these, or do I need to open a new topic for each issue?
Thanks a lot
Hey @jacobacci – thanks for checking in here. We’re absolutely committed to getting classical right, but it’s a big job and we’re a pretty small team, so be aware that we’re working on this class of a problems in a few different ways.
In the near term, we do sometimes report errors from this site back to our metadata providers. This happens when there are clean cut cases that demonstrate the provider’s data is wrong, or inconsistent with their data in general.
However, there are times when these providers disagree about how data should be structured, or where the error is happening in our internal metadata infrastructure. In these cases, we may not reply to every single metadata report, although it is always helpful to have these issues tracked here on the community site.
We are working on some longer-term changes to the structure of our metadata cloud that should allow for lots more flexibility in the future. This could include reporting errors or making edits directly to Roon’s metadata cloud, or crowd-sourcing corrections from users automatically. These are all options we’re working on or considering for the future, but I don’t have firm timelines to share at the moment.
There’s so much we want to do in this area, so we need to balance the time we invest in individual errors against the time we spend building architecture and fixes that eliminate lots of errors. Just don’t get discouraged if we don’t respond to every report, since it usually just means we know the issue, and are confident it will be resolved as part of a bigger change. For now, just know that we absolutely intend to get this right long term.
Thanks Mike
I absolutely agree with the approach you are taking. Problems need to be fixed at the source. Not much use fixing individuals symptoms (individual metadata errors), if there is a structural problem (i.e. with the database structure).
Over the last few weeks I have cleaned up the tags of my library, so that Roon can identify as many albums as possible. Cleaning up my own source data has also enabled me to find other errors, such as missing tracks.
Despite the errors I am finding in Roon’s tags, I am over the top that something like Roon exists. It has the potential to become what no one else has achieved in the past - an information platform for music enthusiasts that offers a structured map of the universe of performed music.
In my view there are two preconditions to make something like Roon work in the long term:
a well thought out object oriented database structure that reflects the information structure users think in. I do not believe this structure is very different for Jazz, Pop or Classical. the elements are the same. ID3 tags have not been able to offer a map to this universe. Roon has a chance to solve the riddle. BTW, users in the US and in Europe look at the database objects in a slightly different way. US tend to sort names by first name, Europeans by last name).
Secondly, there needs to be a way to deal with legacy within this data structure. Roon will never be able to catalogue all the music in the universe. I am dealing with Furtwängler at the moment. There are umpteen editions of the recording from different labels that I would like to be able to link to the same metadata node, even if Roon does not have the particular edition (e,g, the excellent remasters by Eduardo Chibas) in its database. I have made a suggestion how this could be achieved here.
THANKS a million for the great work!