Different recordings in first recorded vs. first released

Hello all. After some discussion in this error report (and a similar one here) sebcat suggested that I post my question here for further discussion--good suggestion! Sorry it's long, thread title is a fairly good TL;DR and if you don't want the context you can skip to the proposal.

Question at hand is when there is a first recording that is dated on the site, but the first release is not the same recording. As it happens, both cases here happen to be Nat King Cole, who recorded many tracks more than once, with his trio or with different orchestras (or both). And then there are many, many compilations out there, often with no indication which recording they include. I would imagine this is often the case though for artists with long careers and repertoire repetition.

In my case this is primarily a headache for musicbrainz edits, which is what made me into a SHS user in the first place--it has been very helpful matching recordings and works, and I link here as often as possible. Not that I'm trying to solve a MB problem with SHS (though imho everyone should use it as it's great for tagging), but all the cross-site collaboration in music history research is one of the reasons I love spending my time on it. I think the idea that as a non-participant geek listener you might wonder what recording you're listening to and be able to click through tags and versions and streams between MB, SHS, discogs with minimal effort is pretty great I find.

So, proposing: I think it would be useful to have some way to differentiate when the first-recording version doesn't match the first-release version (and, ideally, which one you're hearing if you play the linked stream, or include both if possible...) Again ideally, something stronger/clearer than free-text comments.

I understand that some policies are already in place, and not critiquing them or asking them to change. Just opening up the idea for discussion.

crappy could you share database links to the examples that you were discussing with SlimD ? These discussions are much easier with examples and links.

We do distinguish first recordings and first releases. Are you looking for us to do something different?

Copying some colleagues who may have a view too Oliver One microtherion artsinspired dany jojo camembert electrique Oldiesmann dudek Bastien

Are you looking for us to do something different?

Good question...

Quote from sebcat

crappy could you share database links to the examples that you were discussing with SlimD ? These discussions are much easier with examples and links.

Happy to illustrate however, but not sure what you mean by a "database link" in this case.

We do distinguish first recordings and first releases. Are you looking for us to do something different?

Can you show me an example where the first recording date does not match the recording date of the first release and that is clearly explained, i.e. where the first recording is not released first, or not at all? (Sorry, there are no good words here...) Then maybe I'll understand how you handle it well in some cases. (This is the case on both those "error" reports, but as I said in my comments there, they're not errors.)

Last edit: 2023-05-26 17:17:13 UTC by crappy

Quote from crappy

not sure what you mean by a "database link" in this case.

Database link is a www link to an entity in our database For example, a link to the song or performance you want to talk about.

Quote from dudek

Quote from crappy

not sure what you mean by a "database link" in this case.

Database link is a www link to an entity in our database For example, a link to the song or performance you want to talk about.

Nicely explained dudek . crappy I just meant pasting the URL to any of our database entities, like Nat King Cole.

Sorry really not trying to be obtuse (that is what I would have guessed), but do the entities on the error reports not count?

Quote from crappy

Can you show me an example where the first recording date does not match the recording date of the first release and that is clearly explained, i.e. where the first recording is not released first, or not at all?

If I get it right, you'd like us to have a separate entry for the earliest recording and for the earliest release in those cases where 1) they differ (i.e. it's not the same track) and 2) both are made by the same artist.

Quote from dudek

If I get it right, you'd like us to have a separate entry for the earliest recording and for the earliest release in those cases where 1) they differ (i.e. it's not the same track) and 2) both are made by the same artist.

Maybe not "separate entries" per se (depending on what the word "entries" means here), but some clarification.

Like with session identifiers, if artist A first records work B in take 12345, the assumption is that then recording 12345 is released. I'm just talking about some clear way of identifying the cases where take 12345+n is released instead. (And in the case of the streams, again, ideally, some way to identify whether what the user hears is 12345 or 12345+n.) If that helps...

The case that initiated this discussion concerns Lost April by Nat King Cole. The first recording (original) was on November 29, 1947. It was first released in 1952 and was credited to Nat "King" Cole, even though it was a "Trio" recording. The King Cole Trio recorded it again on December 20, 1947 and this version was released March 1948 on the 78 rpm single, Nature Boy .

Quote from SlimD

The first recording (original) was on November 29, 1947. It was first released in 1952 and was credited to Nat "King" Cole, even though it was a "Trio" recording. The King Cole Trio recorded it again on December 20, 1947 and this version was released March 1948 on the 78 rpm single, Nature Boy .

Sorry for another novella here.

If this were correct, it would just be a weird kind of crossover, but it's not, and this is what my report tried to point out. I'll provide (external) context links below, but the 1947-11-29 recording was not released until much, much later.

As I understand it (but I'm really a NKC amateur, so what do I know), between late 1947 when both "takes" were recorded (both by the trio, the -11-29 just the trio, the -12-20 with a string arrangement) and 1951 when the single Unforgettable went bananas, the trio faded away, NKC became solo-famous, and then the album Unforgettable (not among SHS releases, as it was a bit of compilation perhaps) didn't mention the trio despite having plenty of tracks they'd recorded. Just "songs by Nat King Cole". (Irving Gordon wrote "Unforgettable", and NKC didn't write "Lost April' either.)

Anyway, as SlimD points out, the track "Lost April" was first released as the b-side to the single Nature Boy. The A label is credited to NKC, the B to the trio. But it's the same recording as the 1952 release, not another; they're both the 1947-12-20 take.

This is really neatly laid out on the excellent session/discography/fan site A Pile o' Cole, you'll see direct links in the context below. Their note about the Lost April take (2546) on the 1947-11-29 session:

a Trio-only version, originally used for radio transcription only while they decided to re-record it again with strings when possible for records.

All of this was in the original report, though with less text perhaps. But it's a very good example of what I'm talking about, and the NKC discography is full of this kind of thing, sometimes with a couple more thrown in for fun. (Another [probably less common here because they just get trashed, but easier to understand] is if the same session produces a release from take 27 but not takes 1-26. Technically, take 1 is the first recording...)

As for the recordings themselves, they run about the same length, but the difference is easy to spot: in November, after piano-vamp introduction, Oscar Moore's guitar kicks in. In December, the string arrangement takes its place.

This is a useful release: The Complete Capitol Recordings of the Nat King Cole Trio

1947-11-29 session

1947-11-29 version of Lost April

1947-12-20 session (the next one, as it happens, after 11-29)

1947-12-20 version of Lost April

Nature Boy b-side release (IA)

Unforgettable 1952 album release (YT official NKC channel)

The problem appears to be due to an error in https://www.jazzdisco.org/nat-king-cole/discography/ which shows the November 24, 1947 recording was released on the 1952 album, "Unforgettable." I will make the necessary corrections. Thank you for bringing up the topic.

Quote from SlimD

The problem appears to be due to an error in https://www.jazzdisco.org/nat-king-cole/discography/ which shows the November 24, 1947 recording was released on the 1952 album, "Unforgettable." I will make the necessary corrections. Thank you for bringing up the topic.

Thanks, was not rehashing the report though. It's good to fix a specific error (and there are mistakes everywhere, hard to be certain pre-digital [ultimately a good thing I'd argue, irrelevant I suppose], the general goal for all who spend/waste their time doing this is the same) (also btw the other one Error report #58058 is the same), but I am still asking the general question here on the forum:

How concretely does SHS differentiate when the first known recording does not match the recording featured on the first release? Please show me a good example, and then this whole thread is another one of those examples of time spent/wasted .

An example somewhat similar to Lost April/Nat Cole/NKC Trio is That'll Be the Day / Buddy Holly / The Crickets.

Quote from SlimD

An example somewhat similar to Lost April/Nat Cole/NKC Trio is That'll Be the Day / Buddy Holly / The Crickets.

Yes, that's close. Is there another example where both first recording and first release are the exact same artist though? This sort of reminds me of all the works that have a version by the King Cole Trio and later a version by NKC with maybe an orchestra

crappy, see Slowly by Webb Pierce (original recording) and Slowly by Webb Pierce (original release).

Beatles: Love Me Do (same artist name, different drummers), One After 909

Joy Division: All four tracks on Les Bains Douches 18 December 1979 plus some on The Peel Sessions [1] and The Peel Sessions [2]

Thanks to all for the examples, sorry the thread went stale.

Obviously the original feature request is already implemented. I guess the only other question I would ask is could the separate cases be better linked than in comments? (It's a nice-to-have, not essential, and some commenters are linking at least to the first released whole (e.g. Lost April is noted as appearing on Nature Boy).

Appreciate all the clarification.