I've noticed that sometimes a different recording of the same piece will sound like it's being played in a different key. For example, I listened to Concerto No. 6 in B-Flat Major, BWV 1051: III. Allegro in a recording by Richard Egaar and then one recorded English Baroque Soloists. When I compared the two, the Egaar version sounded to me like it was performed a semi-tone higher than the EBS recording. If the piece is in B-Flat Major, shouldn't both performances sound like they are in the same key?
The difference, in short, is because one of the ensembles is using historical tuning practices.
The modern pitch standard is A440, meaning that A4 (the A above middle C) is 440 Hertz. Not everyone uses this; last I heard, the San Francisco Symphony uses an A a little higher (442, perhaps), and some push it down to, say 438. But A440 is nevertheless the international point of reference.
But the fact is that this wasn't always the case. Performers interested in historically accurate performance, especially with the use of period instruments, have settled on this A4 being a half step lower at 415 Hz.
As a slight digression, this is one upside of used fixed-do solfège without chromatic alterations in the syllables. If you're playing a piece in D major on modern instruments, that matches our "re" solfège syllable. And if you're playing it in historical tuning (a modern D♭), that's still "re"!
And speaking of D major, here are two samples for anyone unsure of what this difference sounds like: here is Bach's fifth Brandenburg concerto (in D) played with modern tuning; here it is with Baroque tuning.
Leaving aside the possible of different recording speeds, the half-step change in the pitch of the reference note will be noticeable to a lot of people. Modern example: Jimi Hendrix often (well, often enough) tuned his guitar down a half or whole step.