As I already said this is really a bit of a conundrum.... I think a good analogy would be imagine writing in a word processor (e.g. Writer) or typesetting system (e.g. Latex) VS. trying to feed what you wrote to a Text To Speech (TTS) engine/system able of rendering the text you write as if it were a professional actress.
Feed plain text to a text to speech and it will sound like... well text to speech, if you want to add articulations etc. you have to 'break' the plain text with some sort of semantics (markup), or invest a lot in some sort of interpretation engine which can also do linguistic analysis. That kind of thing gets magnitudes more complicated with music.. But after all, notation is really made for humans, while the matrix editor (or even the event list editor) is much more similar (if still high-level) to what you would feed to a machine (by the way the matrix editor is called "piano roll" in some applications for a reason, so 'machine' is not necessarily a computer). Well well always very interesting discussions on the RG mailing list :) On 08/04/2016 15:19, D. Michael McIntyre wrote: > On 04/07/2016 07:04 PM, Silas Mortimer wrote: > >> If I might ask, because I've been wondering about this, what makes >> doing notation so difficult? > > I think the root of it is because notation is a very analog, > infinitely-variable kind of thing that is difficult to represent and > manipulate in an orderly digital world. > > I have probably over 1,000 pages of commercially published sheet music > for various instruments sitting around in my house, and it probably > wouldn't take me 30 seconds to find a score that Rosegarden can't be > used to reproduce. It would probably take me more on the order of 30 > minutes to find a score that Rosegarden CAN reproduce exactly like the > original, with no compromises. I would probably have to pull that out > of some basic band method book too. > > Notation is difficult, because of the amount of effort that would be > required to address any random one of a hundred different scenarios I > could come up with that Rosegarden doesn't know how to handle. > > Kneed beams. How the hell would we ever make kneed beams work without > seriously rethinking everything from the ground up? I have utterly no idea. > > Anacrusis is something I've banged on off and on for years, and we still > can't really handle it probably, or get it exported to LilyPond > properly. Close, but not really a cigar. I have a trumpet method book > with 1,000 pages of stuff Rosegarden can't handle. It's basic, common > stuff that's hard to work out how to achieve in a notation editor > grafted onto a MIDI sequencer. > > After 15 years of this, I could go on for days, Silas. Doing notation > on top of a sequencer is borderline insanity, but it's a crazy kind of > fun to challenge the limits of what is possible, even if it isn't smart > or practical. > > The true notation editors like MusE Score and Finale (they work directly > with notes and lines and staffs instead of MIDI) have an easier time > with a great many of these problems, but they face their own nightmares. > Those things are especially weak when it comes to rendering imperfect > human performances on a page. I've seen absolutely nothing on any > platform in close to 30 years of computer music that could produce a > playable sheet of music without a considerable amount of fiddling around > to tweak all the glitches. > > I'm pretty sure if that magic button could be written, it would be on > the market by now, and would probably cost $10,000 a copy. > ------------------------------------------------------------------------------ _______________________________________________ Rosegarden-user mailing list Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe https://lists.sourceforge.net/lists/listinfo/rosegarden-user
