Hiya - first post here looking for a little help/advice.
Figured it was time I got an account and started adding data in. Tweaked a few bits and pieces and added in cute little 100-copy limited edition CD EP I have (Kyros / Four of Fear) - all up for voting at the moment if anyone’s interested.
I had a bit of trouble editing an edit I made to the Kyros Cloudburst CD single (got extra track name format wrong), so cancelled it and started again - hope that was all right - apologies if not. (My copy of the CD single has a bonus track not already in the database - also there’s a capitalisation query in that edit.)
So to the main problem I have. I’m using picard (currently from the development PPA as the stable one kept aborting on startup) and firefox (current) on Ubuntu 20.10, and I’ve worked out adding a new recordings from scratch and found documentation/help on adding a new release based on an existing version in the database, but I’m having trouble working out how best to merge some data direct from the CD data I have from picard with existing album release data. Is there a FAQ/tutorial on this that I’ve missed?
The Cloudburst CD single seemed to work okay as the existing track data matched, but the album I have trouble with now is Frequency, by IQ (https://musicbrainz.org/release-group/11e6fe10-ba69-3045-9e87-6fbfba3ac875.) I have the 2020 re-release and it is mostly equivalent to the original 2009 release which is already in the database. It has all the 2009 tracks, but it also has a bonus live track at the end not on the original and some of the timings are two or three seconds different from what’s in the database (I’m guessing it was reindexed when adding the bonus track.) If I go ahead with the edit as I currently have it mostly based on the 2009 record I’m worried that the track timings will appear to be different from reality.
Should I carry on like this, or create a new release from scratch? And if the latter, is there a work-flow for importing just the track titles from an existing album (or is it just back to repeated cut-n-paste?)
Hi - welcome to a complex world. You’ll soon get the hang of it.
Loads of ways of doing this. Here is a quick few notes from top of my head.
To get the 2020 CD into place, I would start in Picard and do a “Lookup CD”. When it fails to find a match, this then lets you hit the “Submit DiscID” button and upload the most exact track times. That will ultimately be how EXACT track times get into place. (Even if you add a discID afterwards you can use it to over-ride inexact times)
From the Submit CD button, when you get to the website, you’ll need to do a search for your Release and it will fail to match due to track count. So you hit the button to add a new Release. You search for the old Release Group and start adding your data.
When you step towards the track list it offer a match you can copy.
If not, you could open a SECOND tab and then go to that 2009 edition. Then use the Track Parser button to get at the track names. Then copy \ paste that block of text across.
At the Recordings page you can point to the 2009 version of the recordings.
You don’t need to get it all right in one go. You can go back into a release and keep poking.
I believe you learn more by making mistakes than being perfect
There is a newer set of Help Files around now and the editor of those will no doubt stop by this thread to say “Hi” and point you to a better set of help that I can give. (Pinging @rdswift… Is the Doctor in the House? )
Even if you manually added the release, once you use the Submit DiscID from Picard you can then attach your CD to the release. And then when clicking on the DiscID tab you can click a link that will “Set Track Lengths” which sets the Release to your exact track times.
Regarding the track timings: You can make a new release based on the old one but manually adjust the timings on the Tracklist tab. You can reuse the same recordings even with a 2-3 second difference if the difference is only silence or length of fade-out. See https://musicbrainz.org/doc/Style/Recording#Merging_recordings .
Listening to that IQ album on Tidal, I hear a number of cross-fades between tracks so I’m guessing that’s where the re-indexing came in. That issue was discussed here: Does cross-fade create a new recording or not? and the “consensus” seems to be they should be separate recordings. (I put consensus in quotes because only two editors actively participated in that thread; however, several people “liked” @sibilant’s reply which I take as agreement.)
In that case, you can still base the new release on the old one, and just use the “add a new recording” option on the Recordings tab for the tracks that have changed length.
You can do this using the ‘track parser’ on the Tracklist tab; edit the release you want to copy from, click the Track Parser button, copy the contents, cancel the edit, then on your new release use the track parser and paste the data from the other release. There’s an option within the track parser to use the times or not. [Added: in your IQ example, though, it may be simpler just to use the “based on existing release” option and tweak as needed.]
One final bit of advice: There is a test server at https://test.musicbrainz.org/ where you can practice editing without fear of messing up any real data.
Many thanks for all the help/suggestions. I think I know where to start looking now. I suspect my main error in just poking around to start with was in glossing over the “track parser” button. I’d initially got the impression it was related to some kind of track audio parsing (something like moodbar or fingerprinting etc.) rather than track name text + data parsing etc…
You have started in the right way. Working on a release of a band you know where you can now tweak and learn without being shouted at. We are always here to answer questions. There are no daft questions as we are all learning.
And that track parser is REALLY useful. I use it along side Notepad++ to assemble text from all kinds of places.
On the cross-fade front, I’ve just read that thread and I’m in two minds as to whether it applies here or not.
Firstly, though, I don’t believe the original album has been remixed or anything. There’s certainly no explicit indication/credits to that extent on the sleeve notes, just the copyright “2009/2020” references this year. I’m happy to be corrected on this as I don’t have the original 2009 version though, but my working assumption is that tracks 1-7 are exactly the same audio as on the original.
So, the problem here is that I wouldn’t describe what’s happening on this album as “cross fading”. E.g. Tracks 3+4+5 on the album are one continuous piece of audio. It’s just that whoever mastered the 2020 version picked a slightly different point to define where one track ends and the next one starts. There’s no actual fading involved.
On the other hand, though, problems with AcoustID mismatches may still exist, and on a track by track basis the content does change, but the sum of the tracks does not. If anything, it’s more akin to how much silence is embedded at the start/end. (Which I think also affects the timings of some of the other tracks too.)
This multi-track suite effect is quite common on the albums I have in my collection (Prog often doesn’t lend itself easily to shuffle-play) so it’d be good to work out the best policy for this in advance. I’m happy to be told either way, but my gut feeling at the moment is that the common tracks between these 2009/2020 releases should be the same recording… although will the database take two instances of the same thing with different timings?
That would make me create separate recordings, and then add in detailed annotations to say that these are split in a different way. I also have releases like that. (Pink Floyd’s The Wall being a good example)
I generally look at it as: If I had the two different versions of this CD, and they were both ripped as FLAC files, could I swap tracks 4 over between the two releases and notice a difference? If yes, then they are separate recordings.
If the only changes are moving around silence from one track to another, or an earlier fade, then they remain the same recordings.
Okay - that makes sense. I’ve submitted it as a whole new release and recording and then merged it into the existing group for that album. Hope that works okay.
With a release like that I add an annotation note to point at the different editing of that track 3+4+5. If there are a lot of Releases of the album, there will be a point in time where this new split takes over in a remaster and is then the “correct” one from that point onwards.
It is interesting when you look at Pink Floyd’s The Wall. When that was first put onto CD by the Japanese they got the split so wrong with Young Lust that they actually label a different track with that title. https://musicbrainz.org/release/8eb1532a-bccb-4cc4-b590-5be19386a6f7
Notice tracks 8 \ 9 \ 10. The disambiguration comments explain the splits. Empty Spaces actually includes Young Lust in the same track, and “Young Lust” is really the beginning of One of My Turns. Leaving One of My Turns with only the second half of the track
It is easy to happen when it is the manufacturer cutting up a track and not the artist. This is corrected on later releases.
Have just improved the annotation for those recordings. Need to also add notes to other tracks on the same release that are also a different length. The confusion caused by a concept album.
Yeah, I think I remember some similar errors in the indexing of early releases of The Lamb Lies Down on Broadway / Genesis. (Must dig my copy out of the attic sometime to check.)
Anyway - I’ll add a bit like that to the annotation for Frequency 2020…
…when I can work out how. I suspect the database has that field locked until my current edit passes its voting week; is that right?
You can add and adjust an Annotation at any time, no votes needed on that. There is also a quick access to it with “Edit Annotation” on the right hand side of the Release page.
)
