Track lengths changed inexplicably


#1

Hi,

I’m looking at release https://musicbrainz.org/release/a0f2cecc-6cc6-4863-9977-6d61de8ee3f5 and don’t understand how the track lengths for tracks 12-20 came to be.

Looking at the editing history, the last time the track lengths were changed, they were changed from [2:48, 3:09, 3:31, 2:37, 3:52, 2:41, 4:20, 3:36, 3:45] to [2:48, 3:09, 3:31, 2:40, 3:52, 2:43, 4:53, 3:46, 3:45] – that edit was done in late 2017 by me, fixing incorrect track times.

However, revisiting the release now, the track times are [4:45, 3:57, 3:22, 3:40, 3:28, 3:01, 3:55, 2:39, 4:16]. Not only are these completely wrong, they have no basis in the edit history. They match none of the previous values, nor any edit made after my edit. The associated records are shared by many other tracks, none of which have these incorrect lengths.

This isn’t due to disc IDs, either. The only disc ID that was ever added to the release is mine, and contains the correct track lengths.

Am I missing some way in which track lengths can change that isn’t accounted for by the editing history? I am hesitant to change them back before understanding what happened here.


#2

Usually a later edit should fail if an earlier edit makes changes to it’s target, but here the opposite happened.
Your edit applied, but an edit created within the 7 days of your edit being opened applied later:

https://musicbrainz.org/edit/49859492

That alone still doesn’t explain it though as the lengths now are different than the results of that edit.


#3

I can’t explain what has happened, but I think the track lengths of your discid are probably correct.

The cover art (at discogs) has similar track lengths, except for tracks 18 & 19.
The acoustid fingerprint lengths are usually close, too, also for track 19.
Track 18 has no acoustid, but there is only one recording of this song, but among


there is only one recording with a length ~4:20 (the length on cover)
and the fingerprints of this recording have lengths 4:48-4:56,
so I think the length on the cover is wrong in this case.

2:50 I’ll Be Home This Christmas (cover 2:48, acoustid fingerprints 2:45-2:51)
3:10 Blue Christmas (cover 3:09, most fingerprints 3:08-3:12)
3:31 All I Want for Christmas Is My Girl (cover 3:31, fingerprints 3:30-3:31)
2:40 Frosty the Snowman (cover 2:37, most fingerprints 2:35-2:49)
3:54 This One’s for the Children (cover 3:52, most fingerprints 3:48-4:00)
2:43 The Christmas Song (cover 2:41, most fingerprints 2:40-2:46)
4:53 Another Lonely Christmas (cover 4:20, no fingerprints on this recording, but the only recording
with recording length ~4:20 is https://musicbrainz.org/recording/eb8f7164-94bd-4817-b715-c57eed41afd6
and has fingerprint lengths 4:48-4:56)
3:46 White Dreams (cover 3:36, fingerprint lengths 3:40-3:45)
3:45 The Perfect Year (cover 3:45, most fingerprint lengths 3:41-3:52)

(most=some dubious acoustids excluded, e.g. submissions with very few sources or different artists)


#4

Right, the track lengths on the cover are somewhat incorrect, that was the topic of https://musicbrainz.org/edit/49802618 where I took the track lengths from the physical CD (the same my discid is from, and the same with the wrong lengths on the cover.)


#5

In that case the track lengths should be updated using the disc ID to have the correct lengths set.


#6

I would still like to understand how the track lengths changed in the first place.


#7

I also wonder that. For some tracks, e.g. 12 and 16, it came with the merge in https://musicbrainz.org/edit/57661589 . But for example for track 13 I don’t understand it either.

Anyway, the source of this mix up seems to be the wrong track times on the back cover, and setting it to the disc ID provided times will fix that.