Doepfer

Interesting description !
What I understood from the A149-1 description on Doepfer's site, 
(listening with a statistician's ear) is that the upper output from 
the upper section gives a Gaussian, "normal" distribution of voltages 
and the lower renders the voltages with a "uniform" distribution
For the lower section, this is the other way around, with an VC 
option for log-normal distribution of the lower out

btw: I was not aware of the max number of 256 realisations from the 
lower section's unniform distribution (when CV D =0).
If I understand it all well, transition from uniform to normal 
distribution narrows the resolution down from 256 to 2 ?

regards
Joost




--- In Doepfer_a100@yahoogroups.com, "(i think you can figure that 
out)" <petergrenader@h...> wrote:
> Hello to list.  Sorry I've been away so long.  
> 
> Yesterday I took delivery of the new 149 Quantized/Stored 
> Random Voltages and I thought I'd inflict you all with my first 
> impressions..
> 
> Interesting little bug indeed.
> 
> As always, I am the WORST when it comes to reading manuals 
> and spent about an hour with it before I downloaded Dieters 
> instructions.  Once I did, things that I was noticed were 'oddly 
> related'  made a lot more sense and I suggest to the lovers of all 
> things random out there that you do the download first, because 
> if you leave it to Don Buchla's typically cryptic graphics (which 
> Dieter also incorporated), like me you won't really have a clue 
> what is actually going on.
> 
> The top half first:
> 
> The top half of the 149 consists of two separate outputs of 
> quantized stepped random voltages.  Given that both run off the 
> same random source, they track each other in a remarkable way, 
> which again is modeled after the characteristics of the  Buchla 
> 266.  The top output  (n+1) is quantized in 1V increments and the 
> bottom one (n squared)  in 1/12 volt increments.  Connected to 
> the 1V/oct input of a VCO, this means that one is giving you 
> octaves and the other semitones - and because both run off the 
> same random source, they track one another.  Connecting two 
> VCOs (one to each of the two quantized outputs) will yield some 
> intriguing counterpoints which every once in a while resolve 
> themselves to unisons.
> 
> There is a pot which controls the depth of the spread which can 
> also be controlled by an external input. The level of this VC can 
> be attenuated manually with its level pot.
> 
> One thing to mention: the 149, at least mine, reacts to the 
> TRAILING EDGE of the incoming clock.  If you want your 
> envelope to to track these voltage changes, you're going to have 
> to invert its trigger (an 165 being the easiest solution) to get it 
in 
> sync with what's coming out of the 149.  It took me a while to 
> realize this as I thought at first the n squared out was frequency 
> doubled from n+1 ( the second reacting tto both rising and failing 
> edges).
> 
> Outside of the usual applications, my immediate thought was 
> using the N+1 out to control the 1V/oct VC input of an LFO used 
> to pace a sound event, thus rendering random musical rhythmic 
> intervals.  Being a big fan of this type of control, I have a 
pretty 
> good idea where the n+1 is going to get a lot of use.
> 
> The bottom half:
> 
> The bottom half behaves in a less predicable manner than the 
> quantized section.  It's outputs relfect a true random selection of 
> 256 available levels generated within the module, also in 
> accordance with the conventions of the Buchla 266.  So, yeah, it's 
> not true random in that there are only 256 possible resolutions, 
> but the net effect is pretty much the same.  I guess what I'm 
really 
> saying is that limitation is not perceivable to me.
> 
> The top output will give you a sampling from all 256 levels at any 
> given time.  The bottom output has an adjustable range (either a 
> pot or external VC) which in effect allows you to tune the 
> bandwidth of these fluctations.  Although I;m not 100% positive,  I 
> think it's reacting like a low Q bandpass filter, the peak of which 
> is the pointer of the pot. But instead of passing voltage, you're 
> passing and attenuating numbers within the range of 256. . Fully 
> CCW will attenuate the high numbers and give you emphasis on 
> the low numbers with a gradual taper into the mid range.  
> Centering the pot will emphasize the center with a graudual 
> taper to both the low and high ends, and fully CW will fully 
> attenuate the lower scale and give emphasis on the high 
> numbers with a gradual taper into the mid.
> 
> In any event, the result is very musical.  Atonal, but musical if 
that 
> makes any sense.
> 
> One thing to mention is the while this module is set up like and 
> behaves much in the manner of the Buchla 266, Dieter's method 
> to derive the random internally is not the same as Buchla's 
> approach.  Dieter opted for digital noise while Don Buchla used 
> a noise-modulated triangle wave as the source, much like Serge 
> and Wiard did, but their's use a saw wave.  Because this 
> difference (digital noise), I would be interested to hear the 149 
> agianst a 266.  I know of a 266 locally and wil be suggesting a 
> face off in the newar future and will report back to you all, 
> possibly even with audio examples.
> 
> In any event - if you're lover of random voltage as I am (my 
> system now has six  souces for stepped random)  - buy the 149, 
> it's great.
> 
> hope this helps,
> 
> Peter
> 
>  
> 
> 
> 
> 
> 
> I