How to push corrections done in Picard to MusicBrainz?

I am quite new to Picard and MusicBrainz. I am going through my CD collection and wanting to appropriately tag my files in Picard. But because many of my CDs are by obscure artists (at least to the general US public) they often are not in the MusicBrainz db or have errors based on my actual CDs.

Currently I’ve been submitting the CD ID to MusicBrainz to create (usually) a new release and trying to enter information as correctly as I’ve been able to nuke out. Then back in Picard I lookup the CD again to retrieve that information and then go through the matching process between my existing digital music files and what I have from MusicBrainz. I have yet to figure out the process for tagging while in MusicBrainz, but it’s trivial in Picard. But how to make these additions/corrections made in Picard and get them with as little effort as possible into MusicBrainz? Are these two pieces of software designed for seamlessly doing these sorts of edits back and forth?

Thanks in advance!

What Picard can help you with is initially submitting the release with metadata from your files. If you have clustered all files from one release on the left you can use “Submit cluster as release” in the context menu. See Submitting Cluster as a Release — MusicBrainz Picard v2.13.3 documentation

It’s not possible to submit other tag changes for existing data from within Picard. Generally MB allows editing only directly on the website.

3 Likes

That is quite helpful, thank you. So if I am understanding you correctly, I should:

  1. Starting point: I have already ripped my CD and used Picard to add/edit tags based on the information I have, which is mostly from the CD and its liner notes.

  2. Put the physical CD back into the drive and submit it to MB and see if it is already in the db or not.

  3. If it is in the db I can go through the process on my PC of correlating my inputted information with what is in MB. Often this will result in an information upgrade. But if I notice either mistakes or insufficiencies in the MB data then I would need to edit that entry directly in MB.

  4. But if there is no entry in MB and a new release needs to be entered then I can take advantage of the information I have already used Picard to properly tag and use the submit cluster process you have cited.

Is this the correct way to best make use of Picard and MB together?

Question: One thing I have encountered are “CD stubs” in MB. Will this clustering technique work with that situation?

Yeah, you have a good process there. Little things to note is that when you update data in MusicBrainz of a Release that already exists then data takes 7 days to go to a vote. (Sometimes quicker than that if get 3 unanimous votes)

If you add a new release then you have 24 hours to make edits to it to adjust and make it perfect. Handy to catch those typos.

CD stubs are old things that I don’t tend to worry about. Just focus on adding your own CD data. After you do add a CD, go back to the Release in MB and find your DiscID as you will need “Set Track Lengths” to make sure to update the times.

As you have already noticed MB can be pretty fussy with little details on a release. And difference in artwork is a separate release. Even if it is just a price code or slight change in copyright text.

When you do lookup a CD then watch out for the little green “TAGGER” button on the Release page. This lets you push your data back to Picard.

1 Like

Take care: If you use “Lookup CD”, you may get a false negative. It’s common for a release to be in the database, just lacking the metadata that identifies CDs.

If MusicBrainz has a release that matches yours (same barcode, label, catalogue number, artwork, packaging, etc.), but the “Disc IDs” tab above the tracklist in MusicBrainz has a zero in it, then what’s needed is not the submission of another release, but a submission of the disc ID.

3 Likes

Welcome to MusicBrainz, @Enigmatique !

It sounds like you are committed and diligent. That means you probably will be a fine editor here. Hooray!

I have a similar situation to you, in that I tend to buy CDs and music files from local artists. Often the artists themselves are not in MusicBrainz, let alone the releases.

My solution is to edit the MusicBrainz database first. I (almost) never edit metadata in Picard. I get it right in the MusicBrainz database first, and then have Picard transfer that metadata to my music files. The two things I like about this approach are that: 1) it improves the database for everyone, and 2) it stores my work in a database which will survive, even if the disk holding all my music files unexpectedly fails.

When I acquire a release, I first check quickly to see if it is in MusicBrainz in usable form. For a CD, having Picard look it up in MusicBrainz is a sufficient test. If it finds an entry, all I need to do is improve that entry. If it doesn’t find that entry, then I know I have some amount of data entry to do.

Where I need to do data entry, my first step is to enter or improve all the Artist entries for the artists mentioned on the release (especially the album artists). It is common to find Artist entries that have the same name as someone on my release, but are different people. So, entering disambiguation strings, and extra metadata like official website, for the existing Artist entry becomes part of the task.

Once I have the Artist entries, I can enter the Release. If I have a CD, I start the process from Picard, because Picard a) supplies a count of tracks and the track timings, and b) supplies a discID for the CD. But I do all the track name, track artist, Recording entity, etc. editing on the MusicBrainz web app.

Only then do I actually RIP the CD, break it up into per-track music files, and insert metadata into the music files.

You have music files with a lot of metadata already. My suggestion is that you open those music files in Picard, so that you can copy the metadata, and open the corresponding Release entry in the MusicBrainz web app. Then copy text from your music file metadata to the Release entry in the MusicBrainz web app. Continue until you are satisfied that the metadata in MusicBrainz could replace the metadata in your music files, and you would smile instead of weep.

I would characterise it as, these two pieces of software are absolutely designed to work together seamlessly, but not “back and forth”. Media-derived data like discIDs and ISRCs flow from Picard to the MusicBrainz database, and other metadata flows from the MusicBrainz database to Picard.

3 Likes

Thanks for the detailed feedback, as well as the others who have done the same. Currently I am trying to “clean up” what I’ve already uploaded into MB, and you may have noticed I am getting help in another thread for this. Once that I not only get a good edit completed on the CD I’m currently working on there, I will have to go back and reexamine the ones I previously uploaded. I’m sure I made many of the same errors on those. I’m saying all of this is that I will be coming back to this information once I am ready to create more new releases/edits from my large piles of CDs that I’ve recently ripped. Then I probably will have more questions, but you have given me much to chew over and process. Many thanks to all!

1 Like

This is a wise approach. Getting the metadata right is surprisingly complex, and takes a surprisingly long time. Think of it as a marathon, not a sprint. Take it in phases. Celebrate interim accomplishments.

Amoung the tags which Picard inserts into to music files are MusicBrainz Identifiers (MBIDs) for the Release as well as for the Album Artist(s), the Track, the Recording, and the Release Group. This means that in the future, Picard can open that music file and look up the related MusicBrainz database entry with 100% reliability. (Someone might conceivably have deleted the MusicBrainz entry, but the project tries to avoid that.) Thus you have access to the latest metadata. It is a simple and reliable operation to use Picard to update a tagged music file with the latest metadata. Thus you can think of your initial tagging of music files as a first draft, to be improved in future months and years as the database improves, via reliable Picard updates to music files.

It is always thus, even for an experienced editor like me who knows a lot about MusicBrainz as it applies to the music I work with. I see this project as always welcoming good questions in service of diligent contributions.