JDS: Fair enough that we have to give the autobeamer a complete
set of instructions. However, it would be useful to mention in
the reference manual that *every* possible duration of notes is
handled independently, with the example ...
JN: Make that *can be handled* independently. One should (as is
done now for commonly used meters) set a meter-default, and
include all exceptions to that. If I hear a sane (and better)
solution, I'll change it.
JDS: We agree - it's just a matter of words. What I meant was that
the _autobeamer_ treats each possible duration independently.
That's why anyone modifying the default system _must_ set a
new default themselves then specify exceptions.
You know, Jan, you are making me think precisely about things
I've done instinctively for half a century - it's tough!
Composers almost invariably use beaming as an indication of
articulation, but editors usually use it to increase score
readability. The two can be quite different. Your current rules
are an expression of "readability" rules, and do pretty well at that.
And, I think "readability" rules are the best default for Lily.
But, I'm working from a composer's manuscript, and want to ensure
that every guide he gave us to the way he played his music is
preserved, which means that I want to preserve *his* beaming as
exactly as possible. That's why I'm looking for a complete
autobeaming specification that I can put in Scarlatti's pieces,
one where I can be certain that beaming won't be modified if it
is run using later versions of Lily. Otherwise I'll have to
proofread every single last autobeamed piece I have done every
time I upgrade Lily :( Anyone setting music with articulation or
rhythm conventions significantly different from European
classical-romantic music will want to do the same.
JN: Consider this:
time4_4beamAutoEnd_8 = 1/2
time4_4beamAutoEnd_16 = 1/4
time4_4beamAutoEnd_32 = 1/4
If you consider 8. to be in the 1/16 class, this
32 32 8.
will be beamed as:
[32 32] 8.
8.==3/16 -> stop at 1/4.
JDS: Let me try a better approach than my "consider 8. to be in
the 1/16 class".
Usually, 8. is associated with 1/16 articulation i.e. sharper
than 1/8. So, if your example continues as
32 32 8. 8 8 8 8 8 8 |
then normally it would be beamed as
[32 32 8.][ 8 8][ 8 8 8 8 ]|
(This is common in Scarlatti, by the way, as an accentuation of
pulse technique.)
But, if you have a romantic piece, you might have the equivalent of
<d'8 a8. f4 d2>
to indicate a fadeout to bass and _this_ 8. is obviously not to be
treated as a 1/16th.
So, it might be cleanest to say that a beam containing short and
long notes is usually best terminated according to the rule
applicable to the shortest.
That is, the autobeamer should look at everything currently in the
beam, rather than just at the duration currently being processed
(at the potential end of the beam, if you like).
Comments, anyone?
JN: We could change, and say:
time4_4beamAutoEnd1_8 = 1/2
means: only 1/8th
and
time4_4beamAutoEnd_8 = 1/2
means: 1/8 3/8 (not 2/8==1/4)
JDS: My personal feeling is that this is more complex than is
appropriate. Musical pulse is usually simple - we are trying to
determining solely from the notation what the pulse is, to
reflect it in the beaming. But, I'm speaking as a musician here,
not as a programmer or graphic artist.
John
