DIY Sampler

[Elliot Inman]

As a person new to this list and new to this adventure, let me preface my
comments with the fact that these are the observations of a true beginner
[def. Beginner (noun, person): Novice who still has unrealistic
expectations].  But I may have something to add on the DIY Sampler.  Last
Christmas, to learn something about IC's, I asked Santa for Radio Shack's
Digital Voice Recording and Sound Lab (one of the Science Fair kits).  The
kit is one of those spring-loaded breadboard contraptions with the little
wires that must be connected from #23 to #45, #12 to #67, etc.  Laugh if
you must.

The board, a cheap $30, is built around an IC, the UM5583, a chip that
allows for a few seconds of playback  -- including some other integrated
functions you would expect: an amplifier, a low pass filter, AD/DA
converter, and a timing mechanism.  The unit allows you to record and
playback a "sample," create a robot voice, speed up / slow down record and
playback, loop the sample, and a variety of other interesting functions.

Did it help me to learn about IC's and digital sampling?  Yes (although the
manual is long on lists of wiring connections and short on explanations).
Did it give me that Dr. Frankenstein "My Monster is Alive!" feeling?  Yes!
Was the "sampler" musically useful?  No, for the following reasons.

Bandwidth aside, the usability of one of these IC's as a sampler is going
to depend on the degree of control you have over the recording and playback
of the sample.  Timing is critical.  You either need to be able to
instantly begin recording at the precise beginning of the sample you want
to record... or have some way of truncating the recording so that the start
of the recording is the start of the sample.  

These voice recording chips may not be the answer.  No one cares if Barbie
hesitates a second before she says, "Math is hard" or if an answering
machine delays three seconds before the first message is played back.  But
in most music, a second of dead air between the triggering of the sample
and the actual sound would make it very difficult to play in real time.  Of
course, you want no dead air at the other end of the recording, either, if
you want to loop the sample.  Samplers, especially sampling software, i.e.
Cool Edit, have made this easy.  I have no idea how easy it would be to do
this with IC's... or if some chips have built-in truncation (like a simple
"eatwhite" formatting C++ programming command that eats null space before
the first valid character entered).

Also, just sticking to the fundamental difficulties, you also will probably
want to be able to stop the sample in mid-playback and have a fast "rewind"
or retriggering speed to allow you to control the length of the playback
(note duration) or do multiple triggering (kuh-kuh-kuh-kuh-cool). The first
samplers (was it an E-mu or a 360 Systems unit?) could not.  Once you
triggered the sample, you could not stop it.  Today, kids' toys have IC's
that allow this (check out the VTech stuff at a toy store), so there must
be cheap chips somewhere. 
    
Someone with a better knowledge of IC's probably has simple solutions all
of these problems.  After all, 12-bit AD converters are apparently very
cheap.  Hook up a 256K EPROM to a 12 bit AD/DA converter, trigger it with a
555 Timer controlled by a single POT to control loop speed, add a mic in
(through a 386 op-amp) and audio out (through another 386 op-amp), and you
have the world's cheapest sampler.  (Possible??)  Now, that's
entertainment!  I look forward hearing what others have to say.
 
Elliot