Wiardgroup

Hi Matthieu
--- In wiardgroup@yahoogroups.com, "matthieuvandiepen" <matthieu@...>
wrote:
>
> Thank you all very much for the response! Especially Doc,still i'm 
> wondering one thing: both modules use analog and digital noise, why 
> is that? What is the difference? And why both use these different 
> noise sources for different applications?
>

i was trying to avoid this...
;'>
this is a very loaded subject, and it can go in about 50 directions at
once!
;'>
and because it is one of my fetishes, i could embark on at least 25 of
them! but i'd love to hear what some of the other eminent experts
especially Prof Richter, on this group have to say about this subject. 
Furthermore.... the Laurie Anderson restriction, "Writing about music
is like dancing about architecture" applies heavily to discussion
group  posts. 
   There are just some qualities that have to be demonstrated live in
order to be easily grasped. i guest-lecture at a local college and one
of my standard presentations covers random parameters as applied to
music. This effort started-out as one 2-hour class in the Electronic
Music Lab but the faculty expanded it to two 2-hour sessions because
the class consistently returned the following week with 2-hours worth
of follow-up questions. So now i cover the subject in TWO sessions.

Suffice to say, this will only scratch the surface and before i begin
i want to point you to some REAL authorities on the subject:

first,
http://www.musicsynthesizer.com/Circuitry/probability.htm

Grant put up a really nifty chart that sums up the differences between
 probablility distibutions in a column of blindingly practical and
succinct graphs from an old test equipment reference. God Bless our
parents generation of crusty old HAM radio veterans!!!!

next,
http://en.wikipedia.org/wiki/Information_entropy

Wiki put up a fairly decent (for once) overview and explanation of
Claude Shannon's definitive work on the subject  

In hard copy,
 
Thomas Henry put out a really nice practical book  with some great
insights into analog noise called the "The Noise Generator Cookbook"
but it's out of print

and finally, i personally, learned a hell of a lot from the very nuts
and bolts treatment of digital noise and shift registers in Don
Lancaster's wonderful "CMOS Cookbook", still in print, and still a
GREAT reference.

Ok... here goes...

To the EE faction, the difference is pretty basic. They tend to view
the phenomena in terms of the origin and the output. In their
shorthand, they told me that the difference is that analog noise is a
back-biased Zener junction, and digital noise is a just a big-ass
shift register.

to expand :
Analog noise originates as a sort of avalanche of electrons spilling
over a semiconductor barrier. It reminds me of a heavily loaded
Pachinko machine***1. In technical terms, the noise is generated by
the Zener breakdown phenomenon in an inversely polarized
base-collector junction. It exploits the shot noise of electrons at a
'choked' semiconducting gate. This physical phenomenon manifests
itself electrically as statistical fluctuations in current flow
present at the base of a reverse-biased bipolar transistor. (It is
also possible to use a reverse-biased Zener diode instead of a transistor)
The current fluctuations go VERY fast and when you amplify them a lot.
It makes a very fast wiggly line on the O'scope. To an EE, that's
analog noise.

Digital noise originates with a shift register that is clocked at a
very fast rate in the audio range. The old MM5437 and MM5837 chips
used as a one-chip noise source in many 80's era synths were
essentially big self-clocking shift registers on a single chip. If you
refer to my earlier post i explain how a shift-register works using a
conveyer-belt analogy. In the case of most conventional digital noise
sources, just consider a conveyer belt with room for many more people,
a steady and very fast clock, and no feedback. That's how the
randomness originates, and the output is not as wiggly as analog
noise. It jumps very fast from step to step rather than rising or
falling from point-to-point as analog noise does. Digital noise is
often called 'peseudorandom' because it's randomness is finite and
limited to the number of bits in the shift register. (spaces on the
conveyer belt imy little analogy) As a result at some point the shift
register method has to 'start over' and it will repeat itself. This
manifests itself in the output audio as the infamous digital noise
'thunk'. The larger the number of bits in the register the longer the
interval between thunks***2 

So...
so far we have Wiggly, point-to-point Analog Noise generated by a
semiconductor wired backwards, 
and Digital Noise, hissing along in fast discrete steps with a 'bump'
every time it runs out of bits and has to start over.

if you're listening at the output jack, the ear just kind of shrugs
and declares that it all just kind of sounds like a hiss.... so what's
the big deal? 

A sample hold is a sort of time-lapse photography it 'takes pictures'
at (ostensibly) a slower rate than the change-rate of the phenomenon
being sampled.....
Now at this point, i bid farewell to my respected EE friends and put
on my (rather silly and impractical) artiste hat. From here on ,my
declarations are PURELY subjective and not only defiantly unproven but
unprovABLE!!!!
  When i sample analog noise at some arbitrarily fixed rate and map
the resulting steps to pitch, my mind's melodic memory notes a
'character' this quality becomes clearer to me over time (and i am
notorious and oft-chided amongst the other residents of Mabuse Manor
for listening to this kind of crap for HOURS!!!).
(ahem)
But if i sample DIGITAL noise at the same clock rate my (perhaps
arguably addled and insane) mind registers not only a higher degree of
repetition (hmmmm THAT pitch sounded familiar!) but also a different 
melodic character.
Call me a head-case (and it wouldn't be the first time) but my
artistic conclusion is that both are 'good' and i need both in my
compositions.

That is an overall survey of 'la difference'
and Vive Ca Difference!!!!



...and to anybody still reading (what's WRONG with you???) i apologize
again, but that's about as small a nutshell into which i can put a
full answer to Mathieu's question.

-doc

***1 Pachinko is a kind of small pinball machine that was (is?)
popular in Japan. Unlike their American counterparts lots of balls
(50-100) are in play at once and there are no flippers (at least on
the ones i played) it's just a sort of barely controlled cascade of
little silver balls


***2 On the MM5437, the interval was minutes long, the MM5837 beat the
problem to a degree. Theres a good explanation at
http://www.vego.nl/8/08/03/08_08_03.htm
if you read German