Re: VMs: MUSIC / Color Images

[Gabriel Landini <G.Landini@xxxxxxxxxx>]

On Wednesday 21 January 2004 10:01, elvogt@xxxxxxxxxxx wrote:
> What you certainly can do is "map" the VM text to a series of notes. IIRC,
> there is the "pentatonic scale" (ie, only the black keys on a piano). No
> matter how you hit them, simultaneously or in sequence, they'll never sound
> disharmonic.

There is a temporal difference between "harmonic" and "melodic".
Harmonic, (sounds played at the same time) yes, but melodic (in a sequence) 
probably not.

There was some interesting research done in the late 70's by Clarke and Voss 
on correlated fractal noises mapped into music. Sequences with no 
correlations (noises with 1/f^0 spectral power) produce unitellegible 
melodies, those with 1/f^2 (1D Brownian motion) produce "boring" melodies 
that wonder up and down, but those with 1/f spectrum are the most interesting 
ones, which tested listeners tended to identify with "music".

> So, what you could do is assign one of these keys arbitrarily to each letter
> of the VM, somehow encode note lengths etc., and you will get somewhat
>  listenable music.

The problem, here though is that mapping symbolic sequences (symbols are not 
quantities) to frequencies (which are quantities) is completely arbitrary. 
(I.e. what is the logic to map certain symbol to a particular frequency?). 
Mappings done this way (with an arbitrary look up table) introduce 
correlations and anti-correlations in the new sequence that do not exist in 
the original data.

Cheers,

Gabriel


______________________________________________________________________
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list