On Fri, 2005-05-20 at 10:59, Nick Collins wrote:
...
> well, to keep ramping this up, run an auditory model simulation informing a
> higher level musical model with both topdown and bottomup information
> passing with appropriate neural connections to motor cortex simulations etc
...
I've been thinking about this recently actually, having been playing
with some auditory models for college, like this one:
http://www.mrc-cbu.cam.ac.uk/cnbh/web2002/bodyframes/AIM.htm
I think implementing the whole mammalian auditory cortical-sensorimotor
loop might be a little beyond us at the moment. But perhaps we could
think about the auditory periphery.
Andy Wuensche's cellular automata software (www.ddlab.com) finds CA
rulesets with "interesting" patterns like gliders automatically, by
looking for patterns with high variance in the entropy of lookups to the
transition table. Not too random, not too static, but "just right",
goldilocks style.
> ftp://ftp.cogs.susx.ac.uk/pub/users/andywu/papers/cplex.pdf
Perhaps a system could be designed or evolved that maximizes the input
entropy variance of the signals in the auditory nerve, as predicted by a
model like Patterson's.
I reckon it'd probably sound like those ambulance sirens that keep
changing pattern to stop you habituating. Or maybe it would just sound
rubbish.
Ben
Received on Fri May 20 2005 - 17:06:46 BST