[A6] Ordered A6, patch sets and data card

[Chad Gould]

On Wed, 16 Apr 2003, Chad Gould wrote:
> > That's what I thought - maybe not that fast, though. It's enough,
however,
> > to get clicking, which is exactly the speed you want. I haven't heard an
> > official measurement.
> > (Us synth kids... if the EG is speedy enough to "click", people complain
> > about the click. If the EG isn't speed enough to click, people complain
> > about the snappiness. I tell you... :P
> Yes but what causes the click? If the machine can sync the OSC and
> everything else and open the amp in less than 1 ms. Would it click?
> (If the OSC is free running then the click is close to inevitable unless
> the amp opens right when the OSC is passing 0 crossing.)

I'd love to know a more technical answer. I'm not quite sure what causes the
click, seemingly particularly on software EGs. Some AH posts blame "control
voltage feed-through" (distortion that is generated when CV voltage leaks
into the VCA) but I don't quite buy that - this sort of distortion is
supposed to be low in tone.

Here's my only guess: Recall the cycle for the low E bass note of a guitar
is 23.8ms (42Hz). That means, clearly, fast envelope decays and attacks
under this time are going to exceed even the value of 1/4 of the waveform
(5.95ms) required for any sort of natural sine slope and recognition to
happen (I think). Not that you need the exact slope always, but not having
the slope means, extra frequencies by causing the volume of the waveform to
be lowered / raised faster than the waveform slope actually  is rising and
falling. So this then would be a problem on all synths, but software
envelopes are "stepped" (square) so the distortion is much more noticeable,
especially on synths whose update cycle is real low. In the early days of
software envs, I'm sure the typically logarithmic analog envelopes (which
would help reduce distortion of this type I'd think) made the typically
linear software envelopes (with stepping it wouldn't even be close) look
even worse. :)

The A6's "normal" envelope I'm sure probably was designed to "step" as
little as possible, so you get something that sounds good. :) "Ultra-fast"
I'm sure pushes the capabilities beyond the A6's optimal cycle rate and
causes envelope stepping to the point where audible distortion can occur on
certain patches and cycles. This is just a guess... but surely you would
need a heck of a lot of CPU cycles for incremental, fast envelope updates
that minimize square distortion. :)

The only real solution is hardware if my theory is correct :), and even that
would click to a degree if pushed (since circuit designs have their own
slew, etc...). Alternately, you find other methods to produce a "faster"
sounding bass (e.g. compressors) I suppose. Or live with clicking as a
"feature". :) 

I suppose it would be easy on the A6 to adjust the ms slope up with time
(e.g. @ low E have a 6ms decay to avoid clicks, @ next octave go to your 2ms
decay...)... that's if my theory is right.
__________
CHAD GOULD