[eu-gene] Psst...Wanna Buy an Algorithm? (Working Title)

Pierre Proske drmoth at tpg.com.au
Mon Oct 17 20:42:15 BST 2005


>Hello Pierre,
>
>Would you still consider the use of generative music in the dentist's
>surgery a condemnation if the music produced masked the sound of the
>dentist's drill and other disturbing sounds and also helped counteract
>the feelings of pain and anxiety that the patient might experience on
>his or her visit or if the music produced in the social security office
>gave the impression that time moved more quickly while waiting in the
>queue and also created music that could relieve the tension and stress
>associated with this sort of environment? 
>
>Paul
>  
>
Hi Paul,
I condemn this sort of thing in general, because it's an public 
implementation of a very subjective remedy that not all would agree to.
I have also yet to see any generative or interactive work to date that 
really tailors itself well to individuals on that sort of level. At best,
broad assumptions must be made regarding the target  group in order for 
these sorts of processes to have any noticeable affect. Or, it
needs to be a system that begins with no knowledge of the world, and 
then customises itself to a particular person's habits and trends.
In my mind (and I'm sure there are people that disagree) the existing 
music in dentist parlours, social security and the like tends to 
necessarily
fall into the "Brian Eno" category of insipid, soothing and harmless (I 
have nothing against Brian by the way) in order to cater for all. 
I can't quite see how a generative system would improve this, and I'd be 
interested to hear more concrete examples that you might have.
My gripe extends to other forms of unsolicited intervention/contact, 
such as telephone marketing and automation of daily procedures that only 
make
life more complicated.  Charlie Chaplin parodies this kind of automation 
in Modern Times (1936), a film which features the Mr. J. Widdecombe 
Billows Feeding Machine,
a machine that attempts to automate the dining experience (to great 
comic effect).
The major problem with your proposition, even though it sounds 
potentially exciting, is that we cannot predict the desires of people to 
such a fine grained degree.
This is directly related to the failure of Artificial Intelligence to 
create more than just reactive cognitive data stores.

Alex said:

>Well I didn't say generative music was flawed, I said the idea of
>constantly changing generative music is flawed.  Just because it's
>varying, it doesn't mean the underlying structure doesn't stay the same.
>We might get bored of it less quickly, but not much less quickly.
>  
>
I'm still reluctant to agree with the statement that "constantly 
changing generative music" is flawed. Minimalism, whether
generative or not, has always relied on very minimal, but constant 
change. Evolutionary and complex systems also
often display surprising behaviours which cannot be easily dismissed as 
facile, and they too constantly change. Perhaps I'm
straying outside of conventional music though.

>I suppose an artificially intelligent system with an understanding of
>music structure that somehow modified it's feedback loops based on the
>value judgements of its audience could produce novel music over time.  I
>don't think that would be a flawed approach, but I don't think we're
>quite there yet...  Are we?
>  
>
Perhaps we are not there in a traditional musical sense. But projects 
like fijuu http://www.fijuu.com, seems to
possibly fit your above description, for example.

>>> In my opinion generative tools for musicians using computers live will 
>>> become more and more de rigueur when attempting to add that 'live 
>>> improv' feel that can be difficult to recreate using software.  However, 
>>> these are still but tools, and should not depended upon entirely in a 
>>> performative context.
>>    
>>
>
>Well you might well be right with your prediction, but the software you
>describe do not sound as though they are "still but tools."  They sound
>like they are automatic music varying systems, with real creative input
>into your performance.  The author of this jamming software is having
>creative input into your live performance.  This goes far beyond the
>role of a simple tool.
>  
>
Perhaps this is a difference of perspectives, but to me the author of a 
piece of "music" software produces a tool,
as a smithy produces a hammer for a carpenter. Withough real creative 
input, that tool is useless. I can choose which
tools to use, and I set aside those I do not want. Is Max/Msp not a tool?
Of course, if the tool develops an element of autonomy, as opposed to 
randomness, such that the control is wrested unwillingly away from the 
artist, this becomes  a little more dicey...

>>> I stress again though that generative 
>>> does not have to imply everlasting performances. The process may be 
>>> everlasting, but it's use may be short lived.
>>    
>>
>
>Then its supposed infinite variation seems to be of no real consequence.
>  
>
But it's precisely the potential for the infinite that is fascinating, 
the mise en abîme.
However if your question is: "Are infinite music systems interesting", 
then I'm not so sure the answer is yes, and in fact the answer will 
never be yes as we would never be able to listen that long :)

>At times I think that the word 'generative' means all things to all
>people, and that we should really stop using it.
>
>
>alex
>  
>
I think that it's good to thrash out a definition, this is the 
generative list after all, but not everyone is obliged to agree with the 
same definition.
Its purely a socio-langual construct to help us better understand the 
"generative" practice.

Pierre.


More information about the eu-gene mailing list