i...@kristoflauwers.domainepublic.net writes: > On Mon, March 14, 2011 6:57 pm, Bernardo Barros wrote: >> we have a decimal system and you want to represent a numeral system >> based on 12 or 24 like [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B]. >> You should propose a system base on 12 or 24 then. >> >> In computer science they use the hexadecimal system because it fits >> computer's bytes representation, if your object is the 12-tone scale, >> be consistent with your system :-) >> > > in practice, most computer programmers think in midi notes: 60 being > middle C, 72 the C above that, ... > as an extension, some software allows 'factional midi notes' (although > they are not in the midi standard, and hardware synths won't understand > them). so 62.33 is on third of a semitone higher then D. this is not as > far fetched as it may seem. i use it e.g. to present overtone scales in > just intonation..
This sort of linear-think is not necessarily helpful. Just right now I discussed how to put Werckmeister 3 into Roland's tuning tables. Roland has thought it a good idea to divide a half tone into 128 steps rather than 100 cents. Higher resolution. Unfortunately, a pure fifth is almost exactly 2cents off. Werckmeister has eight pure fifths, and four compensating fifths that are 4 cent off. Roland's "higher resolution" means that you can't really make anything reasonably close to Werckmeister tuning. You get intervals that are not pure, and the compensating fifths are not all of the same size. All a mess because computer scientists decided they'd do something clever. -- David Kastrup _______________________________________________ lilypond-user mailing list lilypond-user@gnu.org http://lists.gnu.org/mailman/listinfo/lilypond-user
