[NetBehaviour] tangent jump and filtering behavior

Alan Sondheim sondheim at panix.com
Thu Feb 14 17:03:11 CET 2008

tangent jump and filtering behavior

1 tangent jump

shop- http://www.alansondheim.org/tangent.mp4 (old music 19th century
piccolo octaved doubled)

2 filtering behavior

Just as it's possible to construct digital image filters from mathematical
formula, and text filters using the program Jim Reith designed, it's also
possible to construct filters that act upon the recording/residue of human
behavior. In all of these situations, one might consider a unary filter
such that f(x) = x+0 or x*1, similar to the protective glass filter used
on camera lenses (but not the UV filter). With a camera for example, a
neutral density filter is x/n, where x is lumens; the filter for digital
images is similar, and the text filter requires that, instead of x equal-
ling, for example, a particular position of a word in a text, it equals a
position rounded to x/n, such that letters or words might tend to group or
repeat roughly n times. With behavior - the x/n applied to the output of
motion capture - the filter modifies, for example subtended node angles,
which are diminished in size.

Now with analog camera filters (silver photography), the maximum x/n is
x/1; with digital image filtering, the image can go to white (paralleling
an increase, say, in shutter speed in silver) as n itself goes fractional,
i.e. one has m*x where m > 1. In text, the result is skipping words (or,
more accurately, sequences); in behavior, the subtended angle, and there-
fore behavioral positioning, increases and the figure appears to move more
'wildly.' The increase of the angle is cyclical of course, on the order of
360 degrees; large m tend towards the synergetic appearance of _new,_ not
diminished or exaggerated, behaviors as a result.

With motion capture, digital imagery, and text, there is already filtering
in place; the first two are protocol-discrete, and text, on the order of
standardized graphemes, is discrete as well, most likely positioned within
unicode. Complex neural behavior (which is _not_ on/off binary) combines
discrete and analog filtering; the buzz of the world never appears,
although death sinks within it. So one might consider a number of fuzzy
layers - what passes for the buzz of the real; organic receptors, recep-
tions, and processings; reproductive technologies both analog and digital;
the assumption of coherent mappings between reproduction and the buzz;
filtering of coherent mappings which re/produces hidden detailing of the
buzz and/or uses the mappings as originary/catalyst for other produc-
tions (not to mention mappings which have no relation to the buzz as far
as the producer is concerned); and so forth - all of these characteriza-
tions messy, inhering, worlding within the true world.

(Abraham Moles, in Information Theory and Esthetic Perception, 1958/66,
already relates filtering and information; speaking of sonic experimenta-
tion, he states "In short, infinite clipping selectively destroys esthetic
information, and, from a practical point of view, leaves only semantic
information. Thus it 'filters' the two kinds of information." And "The two
methods, clipping and inversion, are essentially means of observing seman-
tic and eethetic information. They may be considered as _information fil-
ters._ Neither one produces a perfectly dichotomous, absolute filtering;
rather, each effects only a statistical filtering." The concept is rarely
used in the book, which predates attempts at a _phenomenology of filter-
ing,_ a phenomenology which (perhaps beginning with Pribram on retinal
knowledge) leads to considerations of organism and Umwelt in general.)


More information about the NetBehaviour mailing list