Same recordings submit fingerprints from different AcoustIDs - why?

I like to merge (gone astray) recordings of Unfinished Sympathy (Paul Oakenfold mix). This was the first track of most of the single releases of this song →

Most recordings (to be merged) have contributed to a specific AcoustID (Track "4acbc83d-7e07-4ea1-93b0-c8e05b9852a8" | AcoustID). This AcoustID can also be found at the target recording (single version). But the target has another one attached: Track "b307d2d2-3792-4601-b0ca-03709fb948cc" | AcoustID. So I thought about disabling the other AcoustID from the recording…

I’ve got two versions ¹ of this recording - one from the original single, the other from the Singles 90-98 compilation. To verify that, I submitted fingerprints from both versions. I expected them to go to the same Acoust ID - they didn’t. The single version’s fingerprint matches one from the second AcoustID.

Fingerprints compared: Compare fingerprints #46460005 and #29429161 | AcoustID

Why are these separate AcoustIDs? And why is the fingerprint of the single version that much longer?
There are plenty of fingerprints with each AcoustID. What makes the difference?

¹) The two versions differ in length by 3 seconds (earlier fade) / one was ripped from vinyl, the other from CD

This sounds like it would/should create two different IDs, unless the pressing and the rip really were perfect.

In any case it looks like your test has confirmed that both acoustIDs correctly belong on this recording? :ok_hand: