On 27 Feb 2007, at 22:17, Graham Percival wrote:
again and produce a minimal example? We don't have the resources to
investigate non-minimal examples.
The example with the meter 25/16 is the most important to me to right
now, as somebody wanted to have a copy of the music. (And it was notated
and printed at least as early as in the 1970s.)
Since it worked in an earlier LilyPond version, perhaps you inserted a
patch to fix the sub-beaming in another situation.
But beaming and sub-beaming is used to indicate metric and logical
primary and subordinate accents. It seems LilyPond already has that
capability, but the interface is complicated. Perhaps a human interface
is needed where this information can be easily communicated to LilyPond.
(The idea might below might perhaps also be used for the n-tuplets
problem and the \times construct, where others seem to have a similar
problem. One then also has the problem that sub-beaming may be used to
indicate logical accents, which may vary with the musical context. The
problem is similar to the meter problem, in view of that complex meters
may sometimes be rewritten using n-tuplets.)
Take for example the meter 7/16, and assume a measure has seven 1/16
notes, that should primarily be beamed as 3 + 4. Then the two groups of
3 and 4 1/16 notes can be grouped in different patterns, according to
how one perceives the metric subordinate accents. For example, the 3
might be grouped without sub-beaming breaks, or a 1+2 or 2+1
sub-beaming, which in some circumstances may be viewed as different meters.
So, one idea is if one could indicate the meter as
\time 7/16 [ = (2+1)/16 + (2+2)/16 ]
If the braces "[" and "]" are present, the idea is that LilyPond will
only use the information to produce the correct beaming and sub-beaming.
If the braces are not present, then LilyPond might write out in the
meter sign (see below).
The example I gave is then using a level above this beaming and
sub-beaming: a compound meter, with dotted bar lines within the measure.
One might then perhaps write:
\time 25/16 = 7/16 | 7/16 | 11/16
[ = (2+1)/16 + (2+2)/16 | (2+1)/16 + (2+2)/16 | (2+2)/16 + (2+1)/16
+ (2+2)/16 ]
The idea is then the time signature also writes out the "= 7/16 | 7/16 |
11/16" part, which traditionally is written as "= 7/16 + 7/16 + 11/16",
I think.
Even if one does not implement such a thing, I suspect that LilyPond may
be in the need pf an efficient interface to indicate the beaming and
sub-beaming that the musical context requires, and that LilyPond cannot
guess this by some general formula. So that might lead to some ideas,
like perhaps the one above.
Hans Åberg