Dear all,
thanks for the explanations!
I understand the pragmatics behind the decision to have the dynamic
performer at the voice context, because of the possible usage for
parallel and independent instruments.
But I am still not convinced: as Maurits has pointed out, nothing in the
syntax suggests that parallel voices do _not_ inherit settings from outside.
If you look at the pdf output of the example, there is only a fortissimo
mark at the beginning of the measure and no other dynamic mark. Any
real performer (!) would not stop at the voices and say "oh, now there
are parallel voices, I should drop back to default volume". But
lilyponds MIDI output does.
Nevertheless I shall implement Aaron's suggestion: it works great for my
needs, thanks for your effort!
Best regards,
Thomas
