Misidentification of some Bach works - BWV vs. BC

[Klaus_Kammerer1 (Klaus)]

I am observing a couple of odd misidentification of Bach works that seem to be based on some Roon magic ;-).

Case 1

Cantata BWV 198 (Lass Fürstin, lass noch einen Strahl)
Cantata BWV 199 (Mein Herze schwimmt im Blut)

For some recordings, like the Bach 2000 box set, the cantata BWV 198 is attributed to recordings of both BWV 198 and BWV 199.
What I find interesting, is that both works are attributed to BC A120. I have no access to a proper list for the Bach Compendium Catalogue, but it looks to me as if there might be some issues due to the same BC key used by two different works.

For the Harnoncourt/Leonhardt and Rilling recordings, both works are attributed to BWV 198

Case 2

BWV 542 vs. BWV 543, both having BC J42 as a shared key

Of course that’s only an outside observation without any claim to be the root cause for this, but maybe @support could look into this?

Thanks
Klaus

[Klaus_Kammerer1 (Klaus)]

@joel any possibility to get a reply on this? It seems I cannot get this corrected without any help.

[Klaus_Kammerer1 (Klaus)]

this has not been solved as of today…

for unidentified albums Roon insists of mixing up BWV 542 and BWV 543

That’s my metadata for an unidentified album (I even took the canonical names as AllMusic used them)

That’s how Roon interprets it:

I guess Roon parses the tags and decides based on some string that this is both the same - it doesn’t seem to take the BWV number in account…

@joel any help?

[Klaus_Kammerer1 (Klaus)]

@support

could I please get any reply on this one? It annoys the crap out of me…

it has been one year that I had reported it…

[mike (Mike Fass)]

Hey @Klaus_Kammerer1 – I can take a look at what’s going on here, but I should be clear that we are not going to manually edit specific albums or compositions – if there’s bad data at one of our sources we can report that back to them, and if the issue is happening in our metadata infrastructure, we can look at what’s going on and analyze what changes might be required, and what side effects those change might trigger.

This is pretty involved work, so I just want to be clear that depending on this issue, these changes may sometimes need to wait for other work, and we’re not always going to be able to go in and manually fix the error, unfortunately.

Anyway, to start, can you post some screenshots of the album pages for affected content?

It would also help if you could give us some background about what edits you’ve made to this content, meaning what import settings you’re using and whether you’ve made any edits to these albums in particular?

Thanks for the feedback, and for your patience Klaus!

[Klaus_Kammerer1 (Klaus)]

@mike

Thanks for coming back to me. The issue seems to me to be a mix of wrong metadata on provider side and some hickup in terms of Roons work identification process.

Library import settings are as follows

We begin with identified albums…

Let’s take Helmut Walchas complete recording of Bach organ works:

for the a minor prelude and fuge, BWV 543, I get two entries for Walcha. This does not correspond to the metadata on Allmusic:

Disc 1

Disc 3

The metadata on Allmusic seems to be correct.

Moving on to Lionel Roggs complete recording:

Disc 6:

Disc 12

here Allmusic is also correct in terms of metadata and Roon displays it accordingly on the album page. However, both tracks are listed on the composition page for BWV 543 in Roon just like the tracks from Walcha:

The composition link for BWV 542 leads me to the composition page of BWV 543

I have not done any edits on these two albums. Interestingly enough, although I have the library import settings for compositions set to “prefer file”, the album details in the album editor for both albums show as to prefer Roon.

Changing these settings to “prefer file” does not change anything. I also do not understand why a global import setting to prefer file is not shown on an individual album edit page.

Next funny thing: Peter Hurfords recording

I also see two entries on the composition page for BWV 543

Disc 2 in Roon

Disc 2 in Allmusic

File tags for these tracks:

Disc 15 in Roon

Disc 15 in Allmusic

File tags for Disc 15

This looks pretty inconsistent to me.

I get the same strange behaviour on unidentified albums:

Track 5 file tag:

Track 8 file tags

That’s why I suspected some Roon logic to interfere…

[tripleCrotchet (Tony)]

I have the same issue. In my case the BWV 542 by Zsigmond Szathmary goes to BWV 542 as expected. But the BWV 543 by Yvonne Lefébure also goes to BWV 542 and that is not expected.

In my recordings the movements are not split and so are being treated as single part recordings by roon. One consequence is the miss identification is less obvious on the album page. So on the Yvonne Lefébure page, for example, the track title of BWV 543 is correct but it links to BWV 542. When the work is multi-part, roon will overwrite file work titles and I think that is what @Klaus_Kammerer1 is seeing.

As a general comment, my roon experience is that I mostly see these sorts of composition mash-ups where a composer’s cataloging system has additional rules that someone in the metadata chain is not applying. In this case after a quick google there seems to be some general lack of standardization on how BWV and BC numbers are correlated. Difficult to say where in the metadata chain the exception handling should be done to compensate for this. I have also had a lot of problems with Debussy and Telemann for much the same reasons in that there are a lot of exceptions to general rules but few manual overrides to compensate.

[Klaus_Kammerer1 (Klaus)]

When the work is multi-part, roon will overwrite file work titles and I think that is what @Klaus_Kammerer1 is seeing.

@tripleCrotchet
The Lionel Rogg album has single tracks and is still misidentified. My guess would be that Roon’s algorithm to match to a work struggles with both works being JC42. I have used the exact names of the works used in allmusic for my file tags. Normally that does the trick matching it to the right “Roon composition”.

[Andre_Gosselin]

Next funny thing: Peter Hurfords recording

I also see two entries on the composition page for BWV 543

I have the same Hurford boxset in my library, identified by Roon. In my case, the composition page for BWV 543 lists only the correct Hurford performance (on disc 15), without the incorrect BWV 542 reference as in your case.Similarly, in my case, the composition page for BWV 542 lists only the correct Hurford performance (on disc 2).

I do not understand why the situation is correct in my case, and not in yours. If you want me to check something else, I will be happy to help.

Could it be that other albums in your library holding a BWV 542/543 perfomance (which I do not have) could exert an influence here ?

[Klaus_Kammerer1 (Klaus)]

Hi @Andre_Gosselin

Could it be that other albums in your library holding a BWV 542/543 perfomance (which I do not have) could exert an influence here ?

That would be my suspicion, too. When I identify the album, the works display correctly:

But that does not change the display on the album page

I do think that the error here is on Roons side and not on the metadata provider side. Something goes wrong when Roon does its black box composition magic…

I also do still not understand the logic about the “prefer Roon/prefer File” settings. To me it looks more like a “prefer file just when you don’t think you found a composition match…”. That’s not what I would expect from a “prefer file” setting.

[Klaus_Kammerer1 (Klaus)]

tagging @support here. I‘d really like to heat your view on these mis-attributions. It‘s terribly annoying and I do not see a manual workatound.

[Miguel_Rode]

Dare I reopen this thread 3 years later? I’m seeing the same issue as @Klaus_Kammerer1 reported initially where many instances of Bach’s Cantata 199 are mis-identified by Roon as Cantata 198. Not all of them, though - a couple of performances are correctly listed as 199, however these are from Qobuz. All versions in my local collection are mis-identified, even though the Work/Part tags are correct in all of the files.

Klaus, were you able to correct this somehow? @support, can we expect a fix from Roon?

[Klaus_Kammerer1 (Klaus)]

@Miguel_Rode
No improvements that I have noticed. Situation unchanged - I’ve resigned…