Re: Polymetric tuplets look bad

[Erik Sandberg]

On Thursday 23 December 2004 00.34, Han-Wen Nienhuys wrote:
> address@hidden writes:
> > <<
> > \new Staff { \times 4/7 { a'4 a' a' a' a' a' a' } }
> > \new Staff { \times 4/5 { a' a' a' a' a' } }
> > \new Staff { \times 4/9 { a' a' a' a' a' a' a' a' a' } }
> >
> >
> > Why so much space for the first quarter note?  The 9:4 tuplets are the
> > shortest duration notes, so I would expect them to drive most of the
> > spacing, but they look very irregular.
>
> This is a dilemma; in single staff music, the (stretchable) space
> should be from right-edge of a symbol to the left-edge of the
> following. Unfortunately, this approach does not work well for
> polyphonic pieces, where the left-edges of the notes should follow a
> regular pattern. Currently, LilyPond uses a compromise between
> different staves; probably, we should switch off all delicate optical
> spacing for polyphonic parts, and use the left-edge of symbols exclusively.

This could however have a negative impact on simple polyphonic music, e.g. 
18th c. duos where the two parts often have the same rhythm.

Would a possible solution be to calculate both a polyphonic-style and 
monophonic-style spacing, calculate a measure of how much rhythmical 
difference there is between parts, and then do some kind of interpolation 
between the spacings, weighted by that measure?

Hopefully this could simulate quite well how a human engraver works. And, 
perhaps it could help improving cases like bugs/clef-rest.ly?

Erik