Evaluating Audio Equipment
Maybe they think it's reassuring or impressive, but when
hi-fi magazines publish equipment directories with something like "6500 Components
Listed!" on the cover, many potential audio buyers just shudder. How can you possibly
make a sensible choice from such a wealth of equipment?
You can't check it all out, so you have to rely on somebody
to select a few reasonable choices for you to consider. That may be a friendly audio
dealer; he wants your business, of course, but that doesn't mean he can't be fairly
impartial, at least when it comes to choices among things he sells. For many people,
however, the place to start is the equipment review section of a stereo magazine.
But evaluating audio equipment is no simple matter. I've
been doing it for a lot of years, and it never gets easier. Sure, it's possible to get a
few ears together for an informal listening session, as we detailed in "Features
& Articles" in August, but most good
equipment reviewing requires a more structured approach.
In many ways, hi-fi is a numbers game. All manufacturers --
most, anyway -- rely on "specifications" to sell products. These are nothing
more than claims or intentions -- a specification describes how a component is supposed
to behave, not how it actually behaves. If an independent lab performs measurements on a
piece of equipment and it lives up to its designer's intention, it's said to be "on
spec." But generally, measurements are useful mainly to determine how competent (or
honest) a stereo company is. Otherwise, most don't mean very much: it may be impressive to
see that this amplifier boasts a distortion spec of 0.001%, while that one claims only
0.01%, but neither level is audible, so the measurements are pretty much irrelevant.
Some are more important: frequency response, for instance,
or noise levels, but most of today's equipment has little trouble achieving excellent
results in these areas. So while numbers may tell us something about stereo equipment, in
the end we have to listen to it. Nobody would fork out thousands of dollars for an audio
system without spending quite a lot of time finding out what it sounds like. For people
who evaluate audio components or processes for a living, listening is a critical part of
the job.
Not everybody agrees on how it should be done, however. One
camp maintains that you learn nothing unless you live with a piece of gear for a very long
time so you can really get to know its quirks and foibles. There is something to this
notion: there are lots of annoying (or pleasing) things about equipment that take a fair
while to show themselves.
The main proponents of the long-term listening method tend
to decry the other system, in which things are compared side-by-side by switching back and
forth between them, straining to hear any differences and interpret them. In truth, this
may be a long way from a normal musical listening experience, but there are some aspects
of audio that can only be appreciated in this sort of direct A-B comparison.
Components that exhibit relatively large differences
probably benefit from both sorts of tests. Things that are supposed to sound the same need
the microscopic treatment.
And sometimes it's possible to imagine differences that are
either not as significant as we think or not there at all. Visual clues are common here --
bigger seems better, or a favorite brand edges out another -- and it's important to filter
these extraneous influences out. The usual way is a double-blind test, in which neither
the listeners nor the test operator know at any given moment which of the possible
equipment choices is being listened to.
Even if the full double-blind technique is not required, as
in our comparison of digital recorders in August, it's very important that the source
material be identical and that things like equalization and level be the same. A minute
difference in loudness may not be heard as such, but rather as a difference in quality.
In those listening sessions, we used some single-blind
techniques, and were very careful about levels and so forth. But in the end we concluded
that being unable to hear differences at all in many cases made it unnecessary to go to
the lengths of trying to interpret such differences. Especially since we concluded that if
there were in fact sonic variations we missed even in direct comparison, they couldn't be
very important.
Still, there were undoubtedly electrical differences
between the components that would look quite gross on paper, even if we couldn't hear
them. Their importance is mostly to alert us to areas where we should be at least trying
to hear something.
In the end, most proper equipment evaluations combine
formal measurements in a lab with extensive listening tests. With luck, the conclusions
from both should reinforce each other.
...Ian G. Masters
ian@mastersonaudio.com
|
