The History of Uncertainty

via Reed:
"A panoply of not-quite-randomness from 1970 to the present. On the left is the Model 265 with its mesmerizing lamps and undisputibly the best sounding noise generator of the series. At center is the pinnacle of late-1970's Buchla, the model 266, a 6-in-one box of wholesome goodness with some extra not-so-uncertain features as a duophonic sample & hold and a voltage-controlled integrator (glide). On the right is the 266e from 2004. Although it looks a lot like its predecessor, the e-series uncertainty module is really its own animal. It does equal divisions of the voltage sweep in the "quantized" section. The original 266 does some sort of wacky binary equation that yields irregular but western-friendly relationships. For example, if you set an oscillator's mod range to an octave, setting the range to 4 steps will give you a major second, perfect fifth, dominant 7th and octave. If you do the same setting from the 266e "quantized" section, you'll get a diminished arpeggio (4 equal divisions). And then there are the stored random voltages....the original 266 has a "time correlation" control, which is kind of a hat-tip to the last-note "correlation" control on the 265. The 266e gives you vc skew, which sort of lets you approximate a s&h sometimes if you've had enough scotch. And by that time it's a good idea just to call it a night."