Re: FW: Professional ???

[Juergen Reuter]

Hi, Christian, Jan, and all the others!

Please, calm down.  I think, you are both right:

Textual processing versus GUI:
- ----------------------------

This is an old controverse discussion.  A GUI is good when you need
actions that require entering coordinates, e.g. when placing objects on
screen.  This may include dragging & dropping notes on a score.  A
textual interface is good if you need flexible processing.  Typing a key
is usually much faster than moving the mouse cursor to a certain spot
just to activate some graphical button.  But moving the mouse cursor
somewhere to define a location is probably much faster than calculating
the coordinates on paper or a relationship to existing objects and then
typing them in.

Both concepts, graphical interface and textual interface have their
individual advantages and disadvantages.  Lily was designed to base on
a textual interface, and I think Lily should keep going like that.
Denemo shows that a graphical GUI can be built upon the textual
interface to combine both concepts.  However, I doubt that a graphical
interface for Lily ever will be able to provide all features that can
be done with the textual interface.  But you may still use a GUI to
enter notes and other symbols and then doing some fine tuning based on
the textual interface.  To better support a graphical interface, Lily
may some day be requested to add features for incremental processing,
e.g. re-compiling a single bar of a score that has been modified by a
user.  But this may require re-breaking lines of the score (and a lot
of other things).  Incremental compiling is a rather difficult job.

Contemporary music notation
- -------------------------

I agree that Lily currently does not support contemporary music
notation.  But what is contemporary?  There is a variety of
contemporary notation just as there is a variety of contemporary
musical styles.  And each composer and/or publisher seems to prefer
his own style of notation, including his own specific symbols.  There
is no standard for contemporary notation, as far as I know.  But there
are quite a lot of notational elements that are shared between many
composers and publishers.  So, what can we do here?  I think there are
some approaches:

* Collect elements of contemporary notation that appear not just in a
  single work, but tend to be used by quite a lot of composers and/or
  publishers, and implement support for these elements.  Note that
  there are differences between scores intended for performance and
  aural scores.  An aural score puts its emphasize on the over-all
  structure of a piece of music, so that a listener can follow it; it
  may drop details or show certain aspects inaccurately for the sake
  of readability.  Besides aural scores and scores intended for
  performers, there are also scores that the composer uses for
  construction of a piece that is produced in a studio rather than
  being performed; these scores are more technically oriented (e.g. a
  time/frequency diagram with exact scales).

* Keep on going to make Lily as flexible as possible.  Then the chance
  rises that a user can get some special feature just by setting some
  properties.  However, this requires good documentation so that users
  are encouraged to experiment with properties.

* Provide an interface for plug-in font symbols so that composers can
  easily add specific symbols that are not supported by Lily.
  Currently, if you want to add a new symbol (e.g. a notehead), you
  have to learn metafont, you have to learn how to embed your metafont
  code into Lily, you have to patch the sources
  (e.g. feta-bolletjes.mf) and you have to re-compile these sources.
  This is quite easy for programmers, but inacceptable for musicians.
  I heard a few times about Adobe Type Manager (atm), but I do not
  know nearly anything about this font design tool/language.  But if
  it is an open format just like postscript, Lily should consider to
  support font symbols created with external graphical tools (whether
  they be free or not) that generate atm files.  Then, even
  non-programmers would be able to graphically create new symbols.
  And, of course, it should be possible to integrate these symbols
  without needing to re-compile Lily.

BTW. I am planning to implement ligature notes engraver facilities and
to contribute them to Lily (during 2001).  I will start with a very
specific porrectus engraver with a shape as used in modal and early
mensural notation.  But with only slight modifications, this ligature
should also be suitable to engrave some of these "slurs" and "curves"
as you can see in contemporary works, e.g. in K. Stahmer's "Suesser
Tod" score intended for performance or in Wehinger's aural score for
Ligeti's "Artikulation".  Stahmer draws thick and thin shapes that
symbolize wide and narrow clusters of voices, respectively.  The y
coordinate still serves to denote pitches (though only roughly).
Wehinger associates the size of graphical symbols with loudness and
introduces different symbols and colors for different classes and
sub-classes of electronic sounds.  I also have an excerpt of a "score"
for some piece of Stockhausen.  This is, effectively, a (quite exact)
diagram (x coordinate: time; y coordinate: frequency), where different
sounds (e.g. sinus wave, square wave, ramp wave) are illustrated with
different colors.

Meaning/Interpretation:
- ---------------------

I think we all agree that music notation and music performance are
somehow related, but that you can not deduce one from the other.  The
limitations in converting MIDI <-> ly (both ways) demonstrate this.
Although mudela claims to be a music definition language rather than a
music notation language, its structure sticks to notation rather than
performance.  Or is there some way to define vibrato?  But I think,
that's perfectly ok, as long mudela is mostly used for typesetting
music.

The meaning of music notation probably was most exactly during the
19th century.  Composers and publishers put many performance symbols
into the score.  This was not always so.  If you look at some Urtext
edition of works of Bach or Mozart, you will see very few performance
symbols.  This is, because interpretation was quite clear from the
habits of that time, and the composer let the performer do the rest.
And often, a composer wrote a piece just to perform it by himself, so
that there was no need to annotate any performance hints.

Contemporary music tends to give back the performer some liberties.
Probably, jazz music has contributed to this tendency.  And often it
just makes no sense to exactly specify a piece of music.  Just think
of a cluster of human voices.  It makes no sense to define exactly
which frequencies should appear in the cluster, because you could not
perform it that exactly, anyway.  To summurize, inaccuracy is a
central aspect of music.  You should not try to fight against it.

Bye,
    Juergen