Re: Why the music is music and the noise is noise?

[Eckard Blumschein <Eckard.Blumschein@xxxxxxxxxxxxxxxxxxxxxxxxxx>]

At 11:46 25.04.2001 -0400, eliot handelman wrote:
>Yadong Wang wrote:
>
>> Can we build a system which can turn the noise to the music?
>
>I understand Yadong' question to mean, can we transform noise into
>something that attains an identity as music such as it exists in the
>world. If on the other hand you can find out how to hear noise as
>music then of course Yadong's question is trivial. So here are some
>considerations concerning a "conventional" musical apparatus.
>
>A way of thinking about this is to consider the symphonic storm --
>where you have representions of thunder and lightning, rain, things
>being blown around, and, in Strauss' case, the actual use of a wind
>machine. The storm is something like a "noise situation" with a
>coherent envelope -- the quiet before the storm, the accelerating
>raindrops, the cloudburst, all hell breaks loose, quiets down, etc.
>The question is, can you generalize so that any noise situation can be
>-- computationally, to boot -- rendered musically? For example, could
>you take the sound of shattering windows, or a collision, or a huge
>explosive chaos, and turn that into something that is musically
>coherent, in the sense of the storm scene?
>
>I believe this can be done. At present, though, this has to be mostly
>a work of the imagination, rather than of a system. I have a short
>demo up at my site which perhaps can convey a small aspect of this
>sort of transformation. I have in fact been attempting to generalize
>and render computational, as far as I can, a wide class of "noise"
>transforms with which one could build complex and yet musically --
>here I need a smiley or something -- "coherent"/"understandable"
>objects.
>
>And of course, as I understand from the experts on this list, we don't
>have the necessary scene analysis chops to be able to extract the
>necessary information from a huge explosive chaos in order to perform
>a musical rendering. But all that will be available in the future
>without any doubt.

Eliot Handelman opened my ears. "Noises" like sound from a roaring hart, a
smashed window, a cloudburst, the undertone of speech, and the sizzeling
sound of short arc welding are anything but merely a genuine random noise
that could be spectrally described and that is as unwelcome as the "musical
noise phenomenon" in restored recordings (cf. Cappé 1991). They convey a
kind of information having remarkable properties: It largely evades
spectral analysis. Nonetheless, it can contribute to perceptions being
relevant to emotions, subconcious stimulations, recalls of complex
experiences, extra-phonemic elements of speech, well paid skills, etc. It
is not understandable in terms of traditional theory of hearing and musical
notation even if such "noise-borne timbre-like information" obviously plays
an important and increasing role not only in musics. I would like to add my
guess that progress in understanding of neural processes from CN up to at
least A2 will make this kind of features more accessible to computer
technology. This might imply a lot. However, I do not agree with the idea
that information can be extracted from void noise, except it is tacitly put
in it by means of the transformation.    

Eckard Blumschein
http.//iesk.et.uni-magdeburg.de/~blumsche/AuditoryFunction.html