In the sample above, the link between the 'G#' in the 2nd measure and 'G' in the 3rd has been interpreted as a slur in one playback program and a tie in another. My interpretation is it's a slur (i.e. different notes). Is one or both of the programs incorrect or is there another explanation?
If it's a slur, the first note is G♯, the second is G♮. If it's a tie, they're both G♯.
Maybe there's some rudimentary notation program that doesn't know a slur from a tie, they're both just curved lines. But in all the ones I've ever used, YOU enter a slur, or YOU enter a tie. Playback interprets them appropriately.
Maybe you're talking about programs that attempt to scan a page of printed music? In that case, I can understand confusion over which was intended, slur or tie.
If you're scoring this piece, I strongly suggest a ♮ on the second G if a tie ISN'T intended. For the benefit of human players.
Assuming the original was notated by someone musically literate, all the signs point towards it being an extended G♯, a tie.
If I was reading those dots, I would hold the G♯ for its total four-and-a-half beats. As it's ostensibly tied across the barline, I'd expect it to stay as that G♯.
I agree it is confusing, although played in context ought to make it clear if it slurs to G♮. Accidentals do continue into the next bar when tied as such, though, meaning the barline doesn't do its normal cancelling job.
It would make good sense to write in a cautionary (courtesy) accidental probably bracketed, before that 'G' in the last bar, to remind players just which one is intended.
EDIT: in reality, the line is in key G, so ending on a G♯ would indeed be rather odd. The last part of the penultimate bar seems like a turn - mordent, perhaps, and would be better written with an A♭ as its last note. Problem solved!
Except - that wouldn't explain why there would need to be a line of any kind - slur/tie joining those two notes... a phrase mark under the last few notes might make sense, but.
This is a tie. The accidental does not need to be repeated across the bar line on a tie, and a natural would be needed to make it a slur.
The argument that the piece is in G and it's weird to end on a G# is assuming that this is the entirety of the piece, but I'm guessing that this is not only not the whole piece but quite possibly not the whole phrase either.
So yes, one of the programs is wrong. The one that thinks it's a slur.