fix cent offset for AEU accidentals

[rettinghaus]

see https://musescore.org/en/node/358194

[Jojo-Schmitz]

Just curious: where did you get those values from?

[Jojo-Schmitz]

BTW, changing

struct Acc {
    AccidentalVal offset;     // semitone offset
    double centOffset;
    SymId sym;
    Acc(AccidentalVal o, double o2, SymId s)
        : offset(o), centOffset(o2), sym(s) {}
};

to

struct Acc {
    double centOffset;
    SymId sym;
    AccidentalVal offset;     // semitone offset
    Acc(AccidentalVal o, double o2, SymId s)
        : centOffset(o2), sym(s), offset(o) {}
};

would save 8 bytes per entry, due to better padding, at the current 147 entries this makes up for 1176 bytes...

[rettinghaus]

Got them from here:

MuseScore/src/importexport/musicxml/internal/musicxml/export/exportmusicxml.cpp

Lines 2277 to 2285 in dd6be26

	// Arel-Ezgi-Uzdilek (AEU)
	case SymId::accidentalBuyukMucennebFlat: res = -0.89;
	break;
	case SymId::accidentalBakiyeFlat: res = -0.44;
	break;
	case SymId::accidentalKucukMucennebSharp: res = 0.56;
	break;
	case SymId::accidentalBuyukMucennebSharp: res = 0.89;
	break;

[Jojo-Schmitz]

Ah, I see, thanks!

Hmm, well... it was ypu adding them there ;-)

[rettinghaus]

Yes, this is only an approximation. To be precise you would take Holrian commas, but then you'd move away from the equally tempered (100 cent based) system.

The Wikipedia entry is a good start to dive into.

For a deeper dive you may go through:

https://www.ozanyarman.com/files/WeighingAgainstHistograms_SEARCH_FOR_THE_OPTIMAL_TONE-SYSTEM.pdf

Currently only four indiviual symbols from the Arel-Ezgi-Uzdilek range are supported by MuseScore, because the others look identical to those from other ranges. But actually they have a different meaning (i.e., different cent offsets) in this context.