Hi all,
Chiming into this as a composer and performer who works constantly with
mixed-meter pieces and has a very strong "house style" that I've worked out
in years of practice, since there was a request for opinions:
>> If I understand you, you are suggesting that LilyPond's default
> >> treatment of (A + B)/D should be changed to group/beam as Gould
> >> describes for (A/D) + (B/D).
> >
> > Yes, beaming patterns for the two different, IMHO identical compound
> > meter notations should be unified.
>
> As a composer and arranger, and as a musical director who comes across
> various/curious time signatures, I agree.
>
> 1. (A+B)/D and (A/D) + (B/D) should be treated identically by default.
>
> 2. (A/C)(B/D), e.g. (6/8)(3/4) should not [!!]
>
I think my only hangup here is the specific case of 2/8 under
consideration. I would agree that (A+B)/D and (A/D)+(B/D) are generally
equivalent — I think of them both as "one bar that lasts A+B Dth notes and
is divided by sub-bar emphasis into the groups shown" (so 2+3/8 and 2/8+3/8
are both a bar of five 8th notes with a downbeat on the first eighth note
and a secondary accent on the third). That said, for a bar like this, the
2/8 part should absolutely be beamed together; I defer to others on what
the correct default for a standalone bar of 2/8 should be, but I feel very
strongly that a bar of 2+3/8 or 2/8+3/8 that prints the first two eighth
notes with unbeamed, individual flags would be a flagrant error of the sort
that would make me embarrassed to send a score out to a performer.
(I also think there's a use-case for (A/D)+(B/D) where the stress on the
second part is very minimal — like a bar of 4/4+1/4 where the last quarter
note is a kind of appended pickup to the next bar — but I don't think this
calls for different notational treatment; it's still one bar made by
stitching two smaller bars together.)
I'm in full agreement that (A/C)(B/D) is a different beast, signaling
independent bars that switch back and forth between the two time signatures
— ideally strictly, but I've definitely seen chaotic scores where the
alternation is irregular. (Not a practice I enjoy!)
> > In case my opinion is not shared by others – folks, please chime in! –
> > we have to improve the documentation, explicitly mentioning that
> > beaming patterns for (A + B)/D and (A/D) + (B/D) are handled
> > separately.
>
> For the particular case under discussion (#1 above): No explicit
> documentation is likely needed.
>
> For the additional case I brought up (#2 above): *IF* Lilypond
> automatically handles (A/C)(B/D), and handles it differently from
> (A/D)+(B/D) — which I believe it should — then it seems evident that
> explicit documentation is required.
>From how I read the above thread, it sounds like there's a proposed change
to have { \compoundMeter #'(2 3 8) c8 c c c c } as well as {
\compoundMeter #'((2 8) (3 8)) c8 c c c c } produce, with default settings,
a bar with five eighth notes where the first two are unbeamed and the last
three are beamed together. Is that correct? If so, I would certainly
appreciate a heads up in the documentation — I understand that I may be
outvoted on the preferred default notational behavior, but a change like
this would, from my perspective, catastrophically break basically every
score in my catalogue, and I'd hope for at least a brief warning about that
beforehand! (But if I'm misunderstanding an the proposal is to make both of
those produce a bar where the first two eighths are beamed together and the
last three beamed together, then I don't think we need a documentation
update would be needed, since that's what I have always understood the
behavior of these functions to be.)
Anyway, that's my 2¢!
Cheers,
brin
