Re: voiceOne dynamics should go above the staff

[David Kastrup]

"Trevor Daniels" <address@hidden> writes:

> Mark Polesky wrote Sunday, September 19, 2010 2:37 AM
>
>> Gardner Read, ch.14, "NOTATIONAL PRACTICES", p.253:
>>  "The general rule is, of course, altered should there be
>>  inadequate room because of elements [...] related to the
>>  staff just below, or when different dynamic markings
>>  affect two voices written on one staff..."
>
> We have to be careful to interpret this correctly.
> None of these writers were familiar with the use of
> "voice" in the computer engraving sense.  By "voice"
> these writers mean parts that are on one staff but
> are to be played or sung by independent musicians.

I disagree.  Baroque polyphony is commonly executed on single
instruments, as composed.  The principal selling point of the pianoforte
(hammer piano) was that it allowed playing several dynamics at once
without being confined to working with registers.  More than a single
dynamic is even needed for executing Bach violin solo pieces, for
multiple voices executed sequentially (via multi-string bowing patterns)
or in parallel (via skewed bow pressure on multiple strings).

And of course, the principal polyphonic string instrument, the lute, was
also employed with voices differentiated in articulation and dynamics.

> But it makes no sense to separate the dynamics of individual voices in
> music that is intended to be played by a single musician, such as
> guitar or piano music[1].

You can even differentiate dynamics of different voices on an accordion
(where every reed sounds off the same bellow, the principal control of
dynamics) by working with articulation, variations of button depression,
psychoacoustical masking (a loud onset masks volume variations of a
continued note) and using registration (where available, to
differentiate between left and right hand).

> Indeed, in piano music LilyPond provides facilities for combining the
> dynamics from two staves.

Not necessarily desired.  In pieces or passages where common dynamics
are desired, it might be convenient to enter them just in one voice (and
have them propagate across all voices, also in Midi).  But the dynamics
would then not be specified in varying locations.  In general, I think
that dynamics should be present in every voice entry, just like
articulations.  And it would usually be the task of the engraver to
merge the dynamics when identical.

One result of the current mishmash is that the dynamics performer just
fiddles with global volume rather than key velocity, propagating
dynamics across voices in that manner.  More or less accidentally, like
we currently do visually.

This bug report is just one more outcome of Lilypond having no clean
concept of dynamics related to voices rather than merely to current
time.

-- 
David Kastrup