Sigfrid Lundberg wrote:
>On Tue, 4 Jul 2000, Phil Taylor wrote:
>
>...
>
>> It looks like my idea of converting tunes to protein symbols in
>> order to use existing biological search and alignment routines won't
>> work very well. The problem is that the statistical distribution
>> of amino acids in proteins is very different from that of musical
>> intervals. There are twenty amino acids; the commonest accounts
>> for about 10% of an average protein, while the rarest accounts for
>> about 1%. In music, there is an indeterminate number of intervals
>> (they just get rarer as you get further away from unison); the
>> commonest (+- two semitones) accounts for about 25% of an average
>> tune, while the rarest is very very improbable. The algorithms
>> can deal with this, but it basically means re-writing the software
>> to deal with the extra symbols, and applying a powerful system
>> of weighting to the matches.
>
>Now, assume that we take a sample of, say, 5000 tunes and calculate the
>"spectrum" along the lines the way you mentioned in your other "Do-Re-Mi"
>mail.
>
>Do you feel that those note distributions/spectra could be used
>for calculating some similarity index between each pair of tunes in order
>to draw a phylogenetic tree similar to the one you wanted to to with amino
>acid sequence method? It does seem feasible to me.
Well, I'm using the method at the moment to recognise the mode of
tunes, and it seems well able to differentiate between different
modes. However, if you want to trace tunes phylogeny, you have to
take into account the fact that tunes often shift between modes
as they evolve. A shift of mode will make a big difference to
the observed spectrum of note useage, and the similarity of the
two variants will be concealed. If you know the tune "The
Princess Royal", it exists in both major and minor key variants.
To the human ear it's very obviously the same tune, even though
three of the seven notes have changed pitch. Perhaps you could do
something about that by grouping the variable notes together.
I suspect, though, that human listeners recognise similarity
between phrases in tunes, and that you cannot discard the information
contained in the note order without losing vital information on
similarity.
>The other interesting usage, to use this for tune searching seems far more
>difficult
>
It might be possible to use it to recognise genre: e.g. search this
database for all tunes which look like Irish reels.
Phil Taylor
To subscribe/unsubscribe, point your browser to: http://www.tullochgorm.com/lists.html
