Degrees of Freedom vs. Hollywood Control Freaks

Riad S. Wahby rsw at jfet.org
Fri Jun 7 19:08:50 PDT 2002


"Major Variola (ret)" <mv at cdc.gov> wrote:
> Yes, in the linear part of their operation.  But its the
> *distortion* (large signal behavior) which differs ---tubes distort
> differently when "overdriven".  I believe the difference when driven
> with a square wave is that tubes have a more RC-like output
> function, vs. a sharper (faster slew) transistor reproduction.

It's true that tubes distort differently, but not for the reason you
claim.

If you compare the transfer characteristics (Vce vs Ic with Vbe as a
parameter) of a tube to those of a transistor, you see that (using BJT
terminology) the transistor passes through saturation and reaches
linear operation at a relatively low Vce (triode/linear, saturation,
and Vds respectively, for you FETholes).  On the other hand, the tube
does not saturate quickly in this way.  If you draw a load line over
these transfer characteristics and grind through the math (or have
Matlab do the work), you'll see that BJTs and FETs produce mostly
odd-order harmonic distortion, whereas tubes produce more even-order
harmonic distortion.

The audible difference in this distortion is consistent with empirical
results from acoustic musical instruments: violins (and other stringed
acoustic instruments), which produce mostly even-order harmonics, have
a drastically different timbre from, say, a clarinet, which produces
mostly odd-order harmonics.

> One little known fact is that humans actually prefer a small amount
> of distortion in their listening.  The THD of amps with a lot of
> decimal-zeroes, is a good technical spec (easily attainable,
> cheaply, nowadays), but is totally a marketing scam.  First you
> can't hear the difference between .01 and .001 % THD, and second you
> prefer ~ .1%

While I'm sure you're right that most people can't tell the difference
between .01% and .001%, I'm not sure I believe that _everyone_ likes
distortion (or, if they do, that everyone likes distortion of the same
kind or level.)  It seems that when you _re_produce a recording you
may as well do so with as much fidelity as possible, leaving it to the
artist to introduce whatever distortion they intend when they record
it.

When I listen to Ormandy and Philadelphia perform Mahler 2 I want the
utmost fidelity; at the same time, I don't complain about the
distortion in the recording when I listen to the Smashing Pumpkins's
latest album.

--
Riad Wahby
rsw at jfet.org
MIT VI-2/A 2002





More information about the cypherpunks-legacy mailing list