[sdiy] MOTM-800 clicks. Arg!

[jhaible at debitel.net]

> This is a problem I worked on many years ago.  The concern is that the 
> *slope* discontinuity produced by the usual attack generator can cause a 
> pop, even though the total attack time is not all that short.  I built a 
> circuit that rounded off the the beginning of the attack curve and was able 
> to convince myself that the attack sounded more natural and that shorter 
> attack times could be used without getting pops.  Unfortunately, I don't 
> remember how I did the circuitry and I can't find any documentation.  As I 
> remember it was a bit complicated.  Anyway, the usual attack curve (step 
> voltage driving RC) seems unnatural.


Interesting! 
Rounding the beginning ... raised cosine comes to mind. (Not that
I'd propose using a raised cosine - just seeing the similarity
of your description and the known advantages of the similar 
raised cosine shape.)
I guess the perceived attack time would then be shorter than
the actual time, with the initial rounded part being more like
a little propagation delay than a rise time. Which brings me
to the next analogy:

This is exactly what modern logic ICs are doing, when they use
output edge control. Provide a fast slope (rise time), but get
rid of most of the higher harmonics.

Now what about this: Pre-processing the GATE signal with 
a fixed slew rate (linear integration), shaping this with
a nonlinear circuit, and then using this instead of the GATE 
signal for the asymptode voltage in an ordinary RC-envelope.
So anything longer than a few milliseconds would behave
like a normal RC contour, and very fast attacks would be rounded.

JH.

-------------------------------------------------
debitel.net Webmail