It’s a mess indeed.
One should decide which tracklist or which edit history is best and remove the 2 others. Something like that…
It’s a mess indeed.
I’ll just add my opinion here, since the discussion has become very complicated imo.
I believe that recordings with different lengths of claps should be seen as separate recordings, because:
- Our schema does not define recordings as a record of “meaningful” (i.e. musical) audio, but any capture of an audio source. So we should not differentiate whether it is the performance or claps that caused different lengths in recordings. Bootleg recordings and official recordings should be linked with the same event (e.g. concert or session), not a merge.
- Having different duration of claps implies there were two different edits. This means different editor relationship with different dates, and possibly different editors. Merging recordings with different clap lengths would mess this up.
My point is just that if the differences are significant enough that the different track splits would necessitate different MB recordings (for example: if someone digitized the vinyl record and included two songs in one track, or accidentally split one song into two, then published that online), then it I think it is OK to add these as separate bootleg releases with distinct recordings, even though it is not an “authorized” release. If there is no difference then there is probably no point, I agree.
Because if someone at the record company does exactly the same thing and makes exactly the same error then the result would unquestionably be new recordings on MB.
Apart from that: there might be a technical problem. I was going to repair a “single medium Vinyl release”* (added a second medium) and tried to select the 18:13 recording. It wasn’t possible. Neither “show more” nor a direct search listed the recording. All shown search results had more than 19 minutes track times.
This is off topic, like what I said about the messed-up ripped vinyl releases.
But I don’t think ripped release uploaded to some pirate site belongs to be in MB.
Just like a homemade compilation is not a release, a rip is not a release either.
Back on topic, I agree with @silentbird, totally.
We should not merge recordings with different length of applause.
It’s not like silence.
You ended the last discussion with something like these words …
But as @yindesu pointed out, different lengths are explicitly allowed by the guidelines. That includes cutting or fading part of the performance, not just different lengths of silence at either end. The editing relationship would then have to be at the release level, not the recording level.
If I am reading the guidelines on fading correctly, it seems to be referring to adding fades that would not affect the length of the recording. It’s adding a fading effect on top of existing audio. That guideline ended with the words:
Where a fade is added to the first or last section of an existing recording, this is not an edit, as the section is not removed.
The bold part implies that it would be an edit if a section is removed or added. In the case of applause, certain applause is either removed (making the recording shorter) or added (making the recording longer). Therefore, the guideline on fades don’t apply to recordings with clapping. It only applies to faded recording and the original recording, where the recording length remains the same (i.e. no section is removed).
I think there’s some gray area here as trimming just a few seconds off the total length with a slightly-earlier fade is probably OK but trimming a large amount of audio not so much.
For a specifc example: two recordings with about 10s extra audio in one due to fadeout differences, I would argue they should not be merged: shorter (4:43) and longer (5:02). Or another pair from the same albums (where the difference is much more obvious when listening): shorter (4:53) and longer (5:13).
I think the same can be said about applause. Just a few seconds: sure, merge away. A full minute maybe not so much.
I’m not sure how the system picks which track length to assign to the recording. I thought at one point it averaged them, but that doesn’t seem to be it. In any case, there can be many track lengths associated with a recording that you don’t actually see when searching.
From what I see I believe the median length of associated tracks is used and if there is an even number of associated tracks the shorter of the two middle lengths is used.
It’s the median, so after the merge it will still be 8:13 for the recording as there are only 3 Vinyl versions with 19:19 and 14 with 18:13. So the value in the middle is clearly 18:13.
Usually it’s possible to find them. At the moment all associated tracks are 18:13. Maybe it will get better, after the recordings are merged.* Does anyone know which query results in the list?
*) if they will be merged! I still would prefer to have them merged as I think, it’s of no artistic relevance if the (uncredited) audience is faded out.
Thanks! Should have read this first
Have you tried copy & pasting the recording url in the search box? That’s what I usually resort to when I can’t find the one I want.
The name of the 19:19 recording is identical to the 18:13 recording. Only the 19:19 recording shows up. (of course, I could fake the track length for the search, probably)
This is my preferred way of dealing with these. If anyone tried uploading their own edits today they would be shouted at and voted down as a home made compilation\edit. You are not allowed to add personal CD-Rs, so these can’t be labelled as personal MP3 rips. So they stay as Vinyl with three different ways of listing the tracks. As there is no consensus on this, I have just left a note for now on the Release Group explaining the mess.
All three basically only have track lists copied out in different ways, the rest of the edit histories are me adding corrections. (The links are up there for anyone curious as to details)
Back to the OP and the clapping. Removal of 15 seconds of clapping seems like an early fade out to me. If it had taken out 15 seconds of music, then it would have been different and should be an edit as some of the artist’s work would have been lost. If I listened to either version of that recording I would get the same content from the artist. This site is called MUSICBrainz and is about the music isn’t it?
The 15 sec fade out is already the same recording (Part I). My pending Edit #73825662 - Merge recordings for Part IIb is about one minute of clapping. This might cause problems with the list of values and if it’s not possible to select the corresponding recording easily, I will revert it. Otherwise it would provoke users to add a new recording for a new Vinyl release.
I haven’t noticed - it’s no longer pending.
Result: the newly merged recording doesn’t show up at once (only 19+ minutes results with indexed search), but is found using direct search. I’m not sure, if that’s acceptable
IMHO, we should merge recordings with different length of applause as long as the musical performance at the heart of the recordings is the same.
I would completely ignore durations, e.g. if one concert encore recording starts with minutes of crowd noise waiting for artists to enter the stage and another one directly starts with the music I would still merge them.
After all, this is MusicBrainz, not ApplauseBrainz
Not sure if the person who defines split points between live tracks or cuts off crowd noise for compilation releases deserves an editor AR or if the date of this action is worth capturing. I think they don’t. And all other ARs for potential merge candidates should be the same - an additional argument for merging them.
Due to accoustIDs, etc. I’d keep them separate. I know (and the way ISRC Search, etc.) they are the same recording, but if one is tagging I’d say it’s more helpful to not have a wide range of times on a single recording of no more than 10 seconds. Exceptions of course would be silence or maybe, if you put in an annotation explaining it, applause that maybe faded out earlier than on other recordings. But tuning and crowd noise, etc at the beginning to me makes it a different recording, even if the music is exactly the same.
Agree on not merging the official ones but think bootleg releases could be merged together with disambiguation: “Bootleg rips”