Re: [OT] Vivi, the Virtual Violinist, plays LilyPond music

[Graham Percival]

On Fri, Mar 18, 2011 at 11:35:37AM -0400, Kieren MacMillan wrote:
> Hi Trevor,
> 
> > Of course, we can't know about "good stuff" that vanished and has not been 
> > rediscovered :)
> 
> Have you ever heard Mozart's son's piano music? There are some
> pieces (especially the Mazurkas) which are clearly superior in
> construction and emotional depth to many of the more popular --
> and thus, by Graham's definition, "better" -- pieces of other
> composers.

Hmm.  Seeing it put that way ("what way?  you mean, accurately?"
"... yes."), I'd like to retract part of it.

IMO, the world would be a better place if we were more precise in
our musical judgements.  If you want to make a subjective
judgement (such as "clearly superior in construction and emotional
depth"), then that's fine; just make it clear that this is your
personal opinion.  If you don't specify that something is a
personal opinion, then "go objective or go home".

The easiest objective judgement is popularity -- or rather,
"amount of CDs sold", "amount of tracks downloaded from a legal
free music site", or even "amount of tracks downloaded from any
source, including quasi-legal (i.e. not legal) and
not-even-quasi-legal sources".

Judgements like "harmonic complexity" or "melodic construction"
can be objective, but you need to specify which algorithm you're
using to determine the harmonies (or melodic stuff).  And then use
that algorithm strictly.  Which, for practical purposes, means
using computer score analysis.


Since we don't even have widespread use of things as (relatively)
simple as harmonic analysis, let alone having a good way of
weighing individual components (like rhythm, melody, structure,
etc)., I think that the only objective judgement we can make is
popularity.  That's why I linked popularity to "quality" so
strongly.

However, I'm hoping that over the next 5-10 years, musicologists
will see the light and start working with tools like David Huron's
stuff, and then we'll see widespread use of automatic
melodic/harmonic/etc analysis.  Once that happens, then we really
might be able to say "Mozart's music is better than Madonna's
music because xyz", where xyz is rooted in completely objective
algorithms.
(or at least, in objective algorithms, using some constants that
were derived from collecting listening data from hundreds of
people in music psychology experiment -- that would be a good
balance between completely subjective judgements of musicologists,
and completely mathematical analyses)

Cheers,
- Graham