Can Shazam identify a whole DJ set?

The short answer: not on its own. Shazam names one recording at a time, from a few seconds of clean audio. In a DJ set, records overlap, are pitched up or down and run through filters, and an hour holds a dozen or more of them. Shazam recognises part of that, but it also names records that were never played.

What Traklyst's own scans show

Traklyst asks Shazam about short fragments of a set, from the start to the end. Across the library that is 66,350 fragments so far:

• Shazam recognised something in 59% of them.
• An average set of 75 minutes took 222 fragments.
• In 12 of 288 sets it recognised less than one fragment in five, mostly sets full of unreleased music.
• Per set, Shazam named on average 43 different records; the finished tracklists hold 22 on average. Part of that gap is one record under two spellings, the rest records it heard in a blend that were not played.

That is why one match proves little. A track only goes on a Traklyst tracklist when the recording was heard at least twice with the same result, or once at the point where a published tracklist names it.

Why Shazam struggles in a mix

• Two records at once. In a blend, the fragment holds both, and Shazam picks one, or a third that sounds like the two together.
• Pitch and tempo. A record played faster or slower than it was released matches its fingerprint less well.
• Edits and unreleased music. Promos, dubplates, bootlegs and the DJ's own edits are in no recognition database. They stay unnamed, however carefully you listen. See what “ID” means in a tracklist.
• The room. In a live recording, crowd noise and the sound system get in the way.

Shazam on your phone, in the club

It works for the one record you hear right now, if it is released and played cleanly enough. It does not give you the set: you would have to catch every record while it plays, and the ones in a long blend are often the ones it misses.

What works for a whole set

1. Look for a published tracklist first: the description, the chapters, the comments, 1001Tracklists or MixesDB. In the Traklyst library, 168 of 319 sets had one somewhere; for the rest, the audio was the only source.
2. Check that list against the audio. Published lists can be incomplete, in the wrong order, or wrong.
3. Listen to the whole recording, not one moment. Fragments from start to end, and a track only counts once it was heard twice or where a list names it.

That is what Traklyst does when you paste a set: it reads what was published, listens to the recording, and marks every track as confirmed by the audio or taken from a list. How it works has the details, and the guide to finding a DJ set tracklist compares every method.