I think that was me -- I asked how many shift registers I would need to
make an audio delay line. I wasn't the least bit serious, though... that
would be a crazy way to emulate an old Reticon SAD chip!
In the same not-so-serious vein, while we're discussing analog shift
registers, here's a photo of a cool analog stitch register:
http://www.merribee.com/stitchregisters.jpg -----Original Message-----
From:
motm@yahoogroups.com [mailto:
motm@yahoogroups.com] On Behalf Of
Richard Brewster
Sent: Friday, December 30, 2005 9:37 AM
To: MOTM List
Subject: [motm] Clocking a sample and hold at audio rates -- Why?
Someone mentioned the ability to clock the sample & hold at audio
rates. People have talked about this with the CGS Analog Shift Register
too, and I have tried it. If the input is a sine wave and the clock and
input are both audio, what you get is a chopped up waveform on the
outputs. The clock is heavily mixed into the output. This is to be
expected. I seem to recall that people thought it would be some sort of
analog delay, where the output would be a replica of the input, but
delayed by one, two, or three clock cycles. This does not seem correct
to me, because of the clock feed-through. A true analog delay, such as
the Blacet Time Machine, goes to great lengths to remove the clock from
the output. This is done by using an ultrasonic clock, which requires
many stages (1024 is typical) to produce a long signal delay. The clock
signal is removed by a lowpass filter at 15Khz or so. I do not see how
clocking a sample and hold at audio rates can begin to simulate this.
As I said, I tried it with the CGS ASR. The result sounded a lot like
what you get by using an audio frequency square wave on the VC input of
a VCA. It is a ring-modulator-like product. Could be useful, but it
isn't a delay. Anyone have other thoughts on why you would clock a
sample and hold at audio frequencies?
-Richard Brewster