Grouped tracks... what to do?

MusicBrainz
, , ,
1

I tried to set up a 10-CD box set using Picard. The set is already in the MusicBrainz database, here.

However, it won’t sync because the tracks in the MB database have been grouped (there is an annotation on the entry that the tracks were grouped when the database entry was imported from Discogs using a script).

So it seems this entry in the database is useless(?) for recognizing the CD, as the tracks are totally different.

How should I proceed?

Edit the existing entry? Create a new release?

If the latter, is there a way to copy everything over, so I don’t have to copy and paste everything?

Alternately, what about an improved script for copying stuff from Discogs?

2 Likes
2

Editing the existing entry would be the way to go. As it is, this release doesn’t reflect any real-world release as far as I can tell, and so is just a mess of bad data.

The way I would probably do it is to click on the Track Parser button for a medium, copy the contents into a text editor or word processor, make the changes and copy it back. Other editors might have a better idea, but that’s what jumps out at me.

If you take this on, I recommend posting a link in the Voting/Auto-editor request thread to see if you can get the edits approved quickly.

4 Likes
3

This is where the track parser is great.

If you can use foobar or another tool to copy over split tracks and track times it will make life much easier.
Or even add cluster as release from Picard, open the track parser from that new release, and then copy the contents over (without creating the new release obvs).

And then decide whether to reuse and move the changes to the old recordings.

If any of that doesn’t make sense let me know :ok_hand:

2 Likes
4

Oh, god, I hate these. The script really should be fixed for classical :confused:

Fixing it is the way to go, but it will take a while, and if you’re a relatively new user it will be a huge pain in the ass that will make you want to leave screaming and never come back. I’ll look into fixing this myself if I find the time :slight_smile:

Edit: Dealt with CDs 1-5, will look into the other 5 later today.
Edit 2: Should be done now. Still a lot to do (adding performers, etc), but eh, it’s a start.

6 Likes
5

It sounds quite similar to this bug:

Discogs tracklist seems difficult to parse.

1 Like
6

Thanks for the tip! Sorry to be such a noob, but what is the “Track Parser” button? Do you mean in Picard? Web site…?

And “foobar”… do you mean foorbar2000? (I’m on macOS.)

Also “Add Cluster as Release” in Picard? I can’t find this.

I have the discs, so I can help get the database info up to date (as long as it’s not a huge amount of work :pray:). I started adding disc IDs, and can do all of them if that would help. I notice the track names on the CD are longer and more explanatory than what was imported from discogs. The track times are slightly different too (off by 1 sec, usually).

If somebody could explain use of the track parser, maybe I could do that.

Finally: this is more of a project than I have time for right now, but in the future I could take a crack at building something that would parse the JSON from the Discogs web API and submit that as XML via the MusicBrainz web API (I have worked with web apps before). Would that be helpful, or are there already apps to do it, and which just need a little massaging?

EDIT: Picard let me add two disc IDs but seems unable to do #3, etc. What’s the correct procedure for this?

7

Hmm. Do your CDs have a different number of tracks? Maybe some of the tracks should not have been divided, or I messed something up :slight_smile:

That’s on the website - if you edit the release, on the Tracklist tab, you can use that button to copy-paste tracklists :slight_smile: The guide for adding releases is hopefully useful to understand the system :slight_smile:

That is a Picard plugin, it might not be active by default, in which case you could activate it from Preferences / Plugins.

3 Likes
8

Thanks. I enabled the plugin.

I’m confused about what I should do to make this sync up.

I drag my rip (using iTunes) of CD1 into Picard, then Cluster, and then Lookup. Nothing happens.

If I Scan the tracks, Picard suggests several other releases in the right-hand pane, but not this one.

If I insert the CD into the drive and click Lookup CD, a dialog window appears with one choice: the full set of all CDs. I click Load into Picard. It does that, but then what? How do I associate the music files with this release, I mean aside from dragging and dropping each track?

There are 55 tracks on CD1 of this release (i.e. 1-01 through 1-55). Same on the CD. But no joy.

What should I do?

9

If you right click a cluster, you can Add cluster as release.

1 Like
10

I see that menu pick now, but should I not try to fix the existing release to match what I’ve got?

If I add a new release, won’t it be necessary to copy over or fill in all the details?

EDIT: Reading the documentation, it seems like I would be adding a duplicate release, no? (I.e., the documentation says: “If the release you want to add is there… then you should have checked better before starting to add a new one!”)

Ergo, shouldn’t the next step be to fix the problems with the existing release in the server database? That’s what I’m stuck on, I guess.

11

I think the suggestion was that you could use that to fix the current release. But isn’t the current state of that release correct now? I thought I fixed it out based on the Discogs data :slight_smile:

2 Likes
12

Maybe I figured it out. I was having no luck with Lookup and Scan, but then I tried Search for Similar Albums…, chose the release from the list, followed by Load into Picard.

That seemed to do the trick :slight_smile:

I’ll try all the rest of the CDs in the box set before reporting complete success.

13

Another question: is it an issue if the song length in the MusicBrainz database is a second or two shorter than what I get by ripping the song using iTunes?

Is there a risk that playback might be truncated?

14

Personally I have found that a couple of seconds difference is usually okay. Although other eds will pick a song down to milliseconds. Many times using ITunes or EAC those results are 1-2 seconds off from the time listed on the medium or cover. Some of the differences may be due to opening and/or closing times added or subtracted due to production. If I am really wrong other eds will chime in. If more than a couple of seconds diff on a few recordings are found I always check for other releases.

15

Long post incoming. As @reosarevok has said, he has already fixed the release, so you don’t have to worry about that anymore, but this might come in handy in future.
Though usually you don’t have to do this much work on a release, let’s be clear!! MB is not always this complicated :wink:

In MusicBrainz, when you are editing a release, go to the track tab, and then the Track Parser will be at the bottom:

image

Clicking on it will display the tracks and their details in text form:

image

This, in combination with the options below it, means you can move information in from other programs (and other releases!) a bit easier than doing it track by track, if you can get the text format to line up. Change the text to match, and then hitting ‘parse tracks’ will update the track list.

My advice was to start adding a new release using ‘cluster as release’, grab the details from the track parser (will be filled in for you), copy them over to the existing release, and then don’t finish adding the new release :slight_smile:

This is not really made clear by Picard, here is how those features work.
Lookup: Uses your existing tags to try find the same release. So it will search MusicBrainz for the same release artist and same release title (or do its best to get something close). And then search for the same track title to match the tracks. If your tags don’t quite line up with the album details in MB, or aren’t filled out, it wont do well.
Scan: Uses AcoustID, where it scans the actual audio and tries to find the same one in the database. This relies on someone having submitted a ID, and is more likely to break your tracks up into different albums (because it’s not using the ‘album’ tag). I tend to use this one as a backup (after matching and tagging the release I always use it to submit the AcoustID).

I see this all the time, I’ve never worried about it haha. I’m sure your copy is fine. I wonder if it has something to do with gaps between track playback?
The track times you’ve put into MB by submitting the DiscID is the most accurate reflection of the time on the disc.

p.s. If you’re serious about your music collection I would move on from iTunes. Every now and again someone posts on here about how it messed up their library/tags. For ripping I would recommend using EAC to get FLAC’s, then you have a 1:1 copy of your CD. No pressure, just pointing it out because you’re clearly putting work in!

3 Likes
16

One second difference shown somewhere really doesn’t tell you much. For one the cover often has it a bit wrong. Maybe the cover was finalized before the final mixes where ready, or it was just a mistake, but things like this are not uncommon.

The other reason is how this gets rounded. Track times are not exactly multiple seconds long, and different tools round differently to full seconds. So the exact same track time of 30.4 seconds could be displayed as both 30 or 31 seconds depending on the tool.

MusicBrainz internally can store track times with millisecond precision, even though you cannot edit it that way. But e.g. if you apply the track times from a disc ID this will have millisecond precision.

But

1 Like
17

Thanks much for all this super-informative explanation.

Re: your last point about iTunes, yes, I don’t really trust it any more and am currently in the process of moving all my music to Kodi. In fact, it was somebody in the Kodi community that recommended Picard, and that’s how I found my way here!

1 Like
18

Digging up this thread since I’m the original importer and suffered first from the userscript’s bug - and, oh boy, thanks so much to @reosarevok for manually fixing that! - but I stumbled upon the beginning of a nice news: here’s a pull request that, according to its author, should fix the bug :

It’s been waiting to be merged for months now. I was wondering: is @murdos (or some other “code admin” - I don’t know how you call this in GitHub) still active regarding his great scripts?

1 Like
19

Hello dear importers,
If there are many Discogs importer users in this topic, could you please test this patched version with some releases that you know have the problematic track numbers?

  1. Click https://github.com/murdos/musicbrainz-userscripts/pull/281#issuecomment-1134139337
  2. Click the link
  3. Click the Raw button, to install this patched version
    On mobile, its: ••• menu, then View raw