Re: Beams not connecting

[Hans Aberg]

I asked about this issue in the Usenet newsgroup rec.music.theory,  
were particularly Steve Latham helped out. It turns out that there  
are more than one styles in play, and Finale even allows one to  
choose (the meter typesetting style 3_/8 defined below is called "in  
one").

In one style, one tries to bring out the metric accent as much as  
possible in the beaming; I indicate this here as k!, as in 3!/8. At  
the other extreme, one ignores the metric accents, which I write as  
k_, as in 3_/8 (mnemonic: a dash as an absence of indicated accents).  
Lilypond typesets an intermediate form, which I write here as k?. In  
addition, beaming is used to indicate musical relationships, such as  
slurs and phrases, and to otherwise bring out the rhythm (as in  
Hindemith, "Elementary Training", p. 78-82).

I will use brackets "[" and "]" to indicate beaming. So |[e8] [e8]  
[e8]| is a measure of three 1/8 notes with flags and no beams, and | 
[e8] [e8 e8]| means that the two last ones are beamed together. And | 
[[e16] [e16]] [[e16 e16] [e16 e16]]| means that the two first 1/16  
notes have a single connecting beam with two flags, and the last  
four !/16 notes have a single beam, with an additional beam within  
the first and last pairs. This captures the beaming and subbeaming,  
except possibly the direction of lose flags in some cases.

So x = | e8. e16 e16 e16 |, is typeset as:
 (x1)  3!/8: | [e8. e16] [e16 e16] |
 (x2)  3?/8: | [[e8. e16] [e16 e16]] |
 (x3)  3_/8: | [e8. [e16 e16 e16]] |
One suggestion was that the form (x3) (which is what Issue 11 wanted)  
was used in CPP music typesetting for two reasons: the metric accents  
are always known (as CPP only admits a restricted set of meters), and  
it easier when drawing by hand to draw long beams than a series of  
broken ones or flags. So there is no reason to use form (x3), except  
for backwards compatibility with tradition: some are used to it, and  
it may therefore help reading. Steve felt that (x1) is the correct  
style, which is what Finale typesets if the "in one" switch is not  
set (and this "off", I think is the default).

Lilypond draws (x2) by default, it seems. Hindemith [loc.cit.] uses a  
similar beamings to bring out the rhythmic structure, so it seems me  
it isn't wrong, but should perhaps be used only when the user calls  
for it explicitly. This style could also be used with complex meters.  
Take for example 7/16, where 7 is decomposed as (2+1)+(2+2) (common  
Balkan meter), meaning primary accent on beat 1, subordinate accent  
on beat 4, and accents subordinate to that on beats 3, 6. Then one  
might typeset seven e16 as (if the user calls for it)
  | [[e16 e16] e16] [[e16 e16] [e16 e16]]|
(Actually, Hindemith's examples could also easily be typeset by the  
use of such matching braces "[" "]".)

Now, there some funny things happening:

Take the time reversed pattern y = | e16 e16 e16 e8. |. This gives  
the styles:
 (y1)  3!/8: | [e16 e16] [e16 e8.] |
 (y2)  3?/8: | [[e16 e16] [e16 e8.]] |
 (y3)  3_/8: | [[e16 e16 e16] e8.] |
Though this pattern is rare, even if one is using style 3_/8 ("in  
one"), it seems that one prefers the reversed form (y1), or even as
 (y4)        | [e16 e16] [e16] [e8.] |
(from Jørgen Jersild, "Lærebog i rytmelæsning", p. 21.)

So in the style 3_/8, one may depart from the unaccented pattern.

Also look at the PDF page 2 of:
  http://imslp.net/images/f/f6/BWV0872.pdf
In the first measure of the allegro (in 3_/8), the first 1/8 is drawn  
with separate flag, but in m. 5 and other places, they are all beamed  
together. The exception in m1. is perhaps due to it is a chord, or  
perhaps an emphasis of some sort. I am not sure this is a good  
example, but I think one might break up the unaccented beaming when  
it is so desired typographically or musically.

  Hans Åberg