I am trying to sort out my music collection so I decided to use MusicBrainz Picard. For most of the job it does well, but there are few songs that for some reason Picard gets wrong tags. I mean completely wrong author and title and everything else.
I ski that data and leave manually entered proper info but each time, file gets processes, Picard gets wrong tags.
It seems tome that someone entered wrong info into MusicBrainz. How to report such issues to someone who can fix?
Hello, welcome to the community - we’re here to try and help
Firstly some questions
Are you using Look-up or Scan function in Picard (they are two different methods and can return different results)
Have you adjusted your Preferred Release Types setting under Options at all?
and most importantly, can you give us an example (screenshots are sometimes best, of Picard with the file that has picked up the wrong details) - you can paste directly into the reply window here without needing to upload with a third party provider
An educated guess right now would be that you’re using Scan, which uses the AcoustID community database (which works closely with Musicbrainz but is not a Metabrainz product/service). This tool that is built into Picard will effectively “listen” to your song, create a fingerprint, query AcoustID’s database and if there is a MusicBrainz link in their database, load that information into Picard.
The problem is that there are many instances when AcoustID fingerprints are incorrectly associated with MusicBrainz recording entities, causing this behaviour.
Well it is actually hard for me to say what actually happens.
Usually I load folder into left panel and then it automatically starts doing something. Some album groups are created in right panel and files are moved there with offered updated tags without me clicking anything. As it happens at once for number of files I get completely lost.
I would expect to load files in left panel and then scanning, lookuping and updating or whatever could would occur only when I run that option and only on file or files I select.
In right panel, not only album groups are created but new songs emerge (I suppose songs that belongs to that albums) even if I do not have those files at all. It is quite confusing and hard even to find file that vanished from the left panel.
Lots of files in right panel got icons changed indicating that something is changed and needs attention. In most cases, cause is that duration of the song in file differs from what is collected from external database (that happens again and against duration of a song is not something I can actually change).
But in some files offered changes are significant, or even wrong. So i have to deal with each single file as I cannot tell what changes are offered until I take a look. Due to wrong tags offered I cannot trust external database to just confirm updates, and it is tedious to check each file again and again.
Another issue is that external tags are full of advertisements. Those are mostly in comment tags but other tags too. If I remove them, they get back on next update.
I do frequent updates on whole database as I am still experimenting on organizing files and I need to run file rename and move options from Picard. To do that I have to load files into Picard and when I do that all tagging havoc starts all over again.
That said, I tested again one of the files I knew there is an issue. and now, it did not behave the same. It did not automatically got wrong tags, not even when I used Lookup or Scan option.
If this behaviour is happening then it means that file already has Musicbrainz tag data stored somewhere in the music file, unless the option for “Automatically scan all new files” is ticked under Options > General
To stop this behaviour, you would want to tick “Ignore MBIDs when loading new files” - this will mean that if your files do have any Musicbrainz tag data included Picard will leave the files on the left-side of the program.
Correct this is how it should work with files that have never been processed by Picard
I’m not quite sure what you mean here, however if your albums are not clustered on the left-side of the program before they are processed, Picard can have a bad habit of loading in loads of different albums if that track has appeared on them. This can be resolved by making sure you use the cluster function beforehand, or automatically if you enable “Automatically cluster all new files” under Options > General
Without knowing what icons appeared it is hard to explain what the issue may be. If durations mismatch that is likely due to the wrong MBID’s being stored in the original file or an error in the AcoustID database. Some icons however may change simply advising that the data in MusicBrainz has changed since this file was last processed. Any fields that are to be changed should appear in green, red or orange font in the details pane at the bottom of the screen.
This can be largely resolved by making use of the cluster function, otherwise Picard just see’s a load of digital files and will match it on “best case” scenario. Assuming you have your music already partially organized into folders/directories, the cluster function will help a lot.
This isn’t any data that is tracked by MusicBrainz, and as far as I know Picard doesn’t fill the comments ID3 tag field by default. Unless a userscript is setup to fill this field, it is likely this content is a hold-over from the original provider of your music. Back in the day many Torrent and music blog sites would fill this field with assorted rubbish, you can set Picard to clear this field immediately by going to Options > Tags > Clear existing tags - CAUTION this will remove ALL ID3 tags except those being loaded by Picard, use the Preserve box to prevent particular tags from being cleared
Do you mean you are making changes in MBz to then have Picard organize your files? That is fine, however there are limits to how many files Picard can handle at any point, and for large libraries it is not recommended to load all files at once.
That was the first I checked: if some automatic processing is turned on. I made sure it is not, except automatic check for updates of Picard itself.
Now I turned on “Ignore MBIDs when loading new files” and will check if that makes a difference.
Clustering option works but only on those files there were not automatically moved to left panel. I cannot do clustering before that. Maybe there should be option to prevent any processing automatically?
Option to mark files that obviously collect wrong tags not to update would also help.
I thought about automating clearing comment tags but, then, there are some files that do have useful info in comments Also they do not put advertisements only in comments but other fields which should have other purpose and are are mostly used for that purpose.
My collection is rather small - just for personal use. I do not intend to reload it in whole in the future. I am doing it now until I set things up.
You’re correct that Picard by default is primarily album oriented when you look up your files. You can change this behavior by adjusting the Preferred Release Types settings. If you want Picard to suggest singles or compilations rather than albums, You can try moving the sliders for Singles and Compilations all the way to the right and all others all the way to the left. You may have to experiment with this a bit to get the lookups that you want. You might also want to have a look at the Lookup Files and Scan Files sections of the User Guide to see how they differ in the way that they retrieve the information from the MusicBrainz database.
Files on the left, unmatched. You can press ‘save’ on these to apply Picard file renaming settings etc using existing tags.
Items on the right, MB database items/albums/songs. The main purpose of Picard is to match your files from the left to the right, so that you can ‘apply’ the data from the matching item in the database to your tracks.
Basically, without moving a song ‘to the right’ there’s no new data to be added. If it’s not being moved to the right after ‘lookup’, you won’t be changing any tags (but you may still be getting useful file changes upon hitting save).
If you want tags updated or added there’s no way around having to match items to the DB. But you can adjust your script to strip out ‘album’ and ‘album artist’ tags, for instance.
What I do is load directory to left panel, click on file, edit tags, save and it works fine. Everything is saved.
The only issue I have is that if I use Lookup to get tags from database instead of entering them manually, then it is processed to the right panel.
Just an checkbox in settings to turn of moving to the right panel would do the job.
What I have to do now is to clear right panel, Lookup on file in left panel, then it is moved to right panel, I have to look for it there, review tags, fix if necessary and save. Then I clear right panel and proceed with next file in left panel. Bit tedious but it works. It would be great to have an option to avoid fiddling with right panel.
There’s no other way for Picard to know what database item you want to take tags from, it has to match to the right to grab DB tags from an entity.
A checkbox or similar to hide un-matched database items on the right could be useful for you though… you can create a ticket, but I suspect you will be looong done with tagging by the time it’s looked at