Value of keyFifthsFlipAt

[rpatters1]

Congratulations on making significant headway with transposition and percussion.

I could use clarification on the language describing keyFifthsFlipAt. It would be very helpful to have specific examples. As it reads, the the most common use cases would seem to be either +/-7. That is, it flips when the number of sharps/flats equals or exceeds 7 in magnitude either way.

I'm also not sure the value of the sign. It seems like this should be inferred from the key signature rather than specified. But either way, presumably a transposition that adds sharps (e.g., Bb) would be +7 and a transposition that adds flats (e.g., A) would be -7.

[mscuthbert]

(for context: https://w3c.github.io/mnx/docs/mnx-reference/objects/part-transposition/ )

Many instruments flip "earlier" or "later" that you might expect. There is an older tradition with transposing band instruments to never have 6 sharps, and to make sure that 6 flats is used instead. Some traditions even prefer 7 flats to 5 sharps, so when a piece is in D, an Eb instrument might be in Cb rather than B.

I am also glad you brought up this discussion @rpatters1 because there is something that @adrianholovaty and @dspreadbury and I discussed but did not make it into the spec, and I now think that we got it backwards: We had decided that the "keyFifthsFlipAt" would be encoded based on the concert pitch of the piece --so that if an Eb alto sax wanted to flip at 6 sharps, it would be encoded as 3.

I think that this decision was a mistake and that the encoding should specify at what key signature the flip takes place from the performer/transposed perspective. First, it's easier for humans to compute and verify -- a large band score where every transposing instrument has keyFifthsFlipAt: 6 would make good sense, as opposed to the jumble of 5s (horns), 4s (Bb instruments), and 3s (Eb instruments) that would be needed to have the same effect. And, families of instruments tend to have the same keyFifthsFlipAt preference[*], so that in a musical, the person doubling on Eb Alto, Bb Tenor, and even C Melody Saxophone traditionally would have the same avoidance of, say, 5 sharps, so they all could be coded as keyFifthsFlipAt: 5 where even the C instrument might be playing enharmonically differently than the guitars and violins. Clarinetists often have similar preferences for whether they'd rather read E-E#-F# or Fb-F-Gb whether on Bb or A clarinet, so the keyFifthsFlipAt would be the same for both.

In any case, whichever interpretation becomes canonical we need to put it in the spec.

[*] -- okay, okay! I know no one ever knew they had a "keyFifthsFlipAt" preference before, but you know what I mean.