rcd@opus.UUCP (Dick Dunn) (08/24/84)
(Comment on an advertising series from Audiophile Systems, U.S. distributors
for Linn and Naim products.)
---
The second lie, as they state it, is:
All amplifiers sound alike.
Explaining, they say that "Some `experts' would have you believe that all
amplifiers that MEASURE alike SOUND alike..." and illustrate by pointing
out that an amplifier's power is generally measured by driving an 8-ohm
resistive load.
---
Well, the second lie, in its simplest form, is certainly a lie--what do you
expect from "reductio ad absurdum" misapplied? But they make a good point
in saying that standard power measurements can be deceptive. In fact,
speaker impedance tends to wander all over as a function of frequency, and
both the resistive and reactive components change.
So why is amplifier power measured into a resistive load? Is this some
hokum that the audio industry is trying to put over on us? Well, probably
not. If speakers all had some characteristic impedance behavior, it would
make sense to use that sort of impedance for a power test. They don't, so
you have to pick something to use as a basis for comparison, and a purely
resistive load makes as much sense as anything. It gives SOME indication.
And, to be sure, some of the problem of power ratings must be laid at the
feet of speaker manufacturers. After all, THEY are the ones making these
components with nasty impedance behavior. Amplifier designers have to do
the best they can to deal with wild impedance variations. Some do better
than others.
Still, it IS the case that you may be able to hear minor differences
between competently-designed amplifiers which seem the same in specs,
driving the same speakers. Some of the differences can be laid to
distortion products. That hardly means that you need to throw all specs
out the window. The "great lie" just sets up a straw man by indicating
that specs can be grossly misinterpreted.
Now, let's get down to how this relates to the products that Audiophile
Systems distributes. The claim for Naim Audio amplifiers (from the same
ad) is that they will deliver rated power with a musical signal to a real
loudspeaker while keeping "ALL distortion products to less than one
thousandth part of the required signal...", where ALL refers to THD, IM,
TIM, noise "or any other type of distortion you care to measure." This
would seem to answer the problem of non-real-worldliness of specs. It is
quite a bold claim.
Or is it? Let's look at what they're saying rather carefully. They are
actually offering a spec of a sort, the one-thousandth thing and all. OK,
the spec is made with a "musical signal". What's that? I can choose a
wide variety of musical sources with widely varying characteristics.
The spec is given for power into a "real loudspeaker". What's that? Can we
really expect that it covers standard acoustic-suspension, transmission
lines, ported, horn-loaded, electrostatics, and who-knows-what-else if all
they will say is a "real speaker"? Not so good; we're 0 for 2 on meaning.
So how about the "distortion products"? Well, THD is normally measured
with a sinusoidal signal so that there's a particular way to add up the
distortion products and get a repeatable value. There's no corresponding
way to assign proper weighting to harmonic distortion products in a
"musical signal". Similarly, IM distortion is measured with a particular
pair of frequencies. In fact, for some signals it would not be possible to
distinguish THD from IM distortion even if you could assign them repeatable
meanings. The final shot, "any other type of distortion you care to
measure" is so absurd it's hard to figure out how to reject it except with
a flame.
A different sort of difficulty with this non-measurement is that it gives
no regard to the fact that certain sorts of distortion (e.g., IM) are
audible at much lower levels than others (e.g., THD).
What we started with was (being kind) an observation that specs can be
misleading; in particular, they may fail to distinguish between two
amplifiers which are audibly different. What we finished with is a promise
that the amplifiers Audiophile Systems is selling can:
Take a musical signal (undefined) and use it to drive a real
speaker (unspecified) with distortion products (all types lumped
together with no consideration of their meaning) not exceeding 0.1%
(even though there are no standards for measuring them in the
described situation).
Nice work, no?
--
Dick Dunn {hao,ucbvax,allegra}!nbires!rcd (303)444-5710 x3086
...Are you making this up as you go along?