> On 02.04.2015, at 15:28, D. Michael McIntyre <[email protected]>
> wrote:
>
> I encourage you to try as many complicated examples as you can produce.
> What happens when a magenta note occurs in the same place as a hairpin
> and a slur and a phrasing slur with some fingerings on a chord that has
> the stem direction set manually and all the notes have three marks?
>
> The other thing is potential LilyPond version issues. Their syntax
> changes constantly, and we haven't updated to support any newer versions
> in several years. Maybe the syntax is different in some of the versions
> we do support. Maybe it works with one version, but not another.
I tried to make complicated test cases and up to now I could not break it
(maybe it's just my limited imagination... 8-) )
It also works for all the currently supported lilypond versions.
So far, so good, but in course of making up these test cases I used Rosegarden
in ways I usually do not do. Now I ran into the following problem:
Normally I write music with the different voices in different staves to get a
better overview (SATB choir voices go to four staves, accompanying strings as
well and so on). Therefore each voice lives in its own segment. Every now and
then I put two segments into one stave, as this is the easiest way to see if
two voices cross, but still the notes live in their own segments. This makes
checking parallels easy, as it is perfectly clear which note belongs to which
voice, as, obviously, there is only one note per segment at any given time (or
a rest).
When making the test cases for lilypond export I played around with several
notes at the same time in one segment to check whether chords are exported ok.
If I allow for this it will become extremely difficult to check musical
transitions. What, e.g., if somebody notes three voices in one segment (just
quarter notes for the sake of simplicity) and at a certain time point one of
the voices has a rest? So you have three notes sounding, then on one beat only
two notes, then three again. You cannot note down the extra rest of the third
voice in between the notes of the other voices. How would the algorithm know
which of the voices has a rest, so how to figure out the musical progressions
of the other voices? Even if there is no rest, there still could be ambiguities
(ok, usually the voices to not cross, so by order of pitch one could have a
decent guess which is which, but is this always sure?)
Frankly spoken I do not really see that I could get this working well when
dealing with real life music...
So coming back to the remark of Michael about having (or not having) perfect
solutions (ok, here it's not about lilypond, but generally on the level of
perfection):
> You don't really have to achieve perfection, because it's impossible, so
> it's fine if the code still breaks in some edge cases. The LilyPond
> code will always break in some edge cases. It's enormously complicated,
> and music of the real world is full of so many edge cases it's ridiculous.
my question is:
Would it be ok if the parallels checker only deals with the situation that the
different voices are in different segments? Or would this be too much of a
restriction? For me that would not be a problem, but I have no idea how other
people use Rosegarden...
If it's ok, should there be a warning if the checker encounters multiple notes
a the same time in one segment ("could not check parallels in voice x at bar y
beat z") or would this be too blathering (chatty? talkative? not sure about the
correct word...)?
Of course I'd like to have a perfect solution, but at the moment I do not see
how. Any hints, tipps and tricks of course are welcome.
Thanks in advance
Armin
------------------------------------------------------------------------------
BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
Develop your own process in accordance with the BPMN 2 standard
Learn Process modeling best practices with Bonita BPM through live exercises
http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_
source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF
_______________________________________________
Rosegarden-devel mailing list
[email protected] - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-devel
