[sdiy] Re: Plan B Heisenberg Generator (and other new releases)

[Peter Grenader]

ross said (On Modular Synth)"

Posted: Fri Aug 04, 2006 10:57 pm    Post subject:

[The Model 24] looks like a Serge rip..i need more Buchla rips

The Plan B - Evil Twin sounds interesting. A harder filter than the
previous mod>>


The Heisenberg Generator is not a Serge rip, it's not a Buchla rip...it's a
completely unique circuit.  If it had to be likened to an existing module
however, it's behavior is closer to the Buchla SOU than anything else and
here's why:

While it can do other things, the Model 24 is primarily a dual mode random
source.  One side giving arbitrary (stepped) random voltages, the other side
producing interpolated (smooth) random.   This is nothing new.  Along with
Serge and Buchla, there's the Woggle Bug, the Doepfer A-149-1, the A-117,
the upcoming Livewire Chaos Computer series, among others.

The approach used to generate the pseudo-random on the M24 is unique
however.  The microp replicates the MM5837 noise chip used in the Buchla 266
via an algorithm, yet at much higher resolution (a pseudo-random selection
of 16,000 possible levels vs. 256 possibilities from the 5873 chip). In that
sense it 's Buchliciian. There are two of these generators working
independently which are summed together in the analog domain before they are
converted to noise, which decreases the possibility of patterns forming
significantly.  Don't be fooled tho - all real-world random generators
periodic to a certain degree, some more so than others.  Mine is less
periodic than the Buchla 266.  There are others which are less periodic than
mine.  But they are all just degrees of scale of pseudo-randomness.  My
focus was musicality and introducing a bit of personality into the
instrument, not in achieving atomic-clock type random accuracy.

One of the advantages of using a microp to do this was we were able to
temper (weigh) the outcomes by introducing conditional information into the
algorithm to improve distribution across the entire range to yield what's
known in statistical analysis as wide range random distribution.  This
eliminated the need for an internal noise-modulated LFO to derive the final
internal parent voltages (a triangle noise modulated triangle wave in the
Buchla 266 and ramp in the Serge Random Gen).

Bringing some of it's internal features to the faceplate and adding VC
inputs for all of it's tactile controls allow for some nifty
poly-functionality, some of which are also patchable on the Serge SSG.  (for
instance, it can be used as a slew generator or as a standard S+H of an
external voltage), but that's coincidental. Don't buy the M24 expecting to
get a Serge SSG!
  

I think also there's some similarities being formed now due to the
nomenclature used (outputs 'Stepped' and 'Smooth').   This is understandable
but unintentional on our part. We opted to be descriptive. There's been a
bit of a trend forming over the last couple of years toward function names
with attitude --> 'fist', 'ream', destroy',  'woggle'  that sort of thing.
It's kinda cool to go there, I'll admit it,  but I've wondered how well  it
will go over ten years from now. Will it survive or go the way of flamingo
silhouettes,  wavey lines and geometric shapes on one's rear window?  It may
stick, but i thought it best to play it safe.  Stepped Random, Smooth Random
- i get it.

Long story short, as far as it's operation the M24 more like the Buchla unit
and much much less like the Serge Stepped and Smooth.  In the weeks to come
I'll do a feature comparison grid on the page which I think will clear a lot
of the confusion.

- P
_________________
Peter Grenader
w: http://www.ear-group.net
e: peter at ear-group.net
p: 866 755-4468 (toll free)