--Frank Wells
Over the past few decades, audio technology has been
refined even while it is being diversified. Analog electronics can be as clean and wideband as the
designer’s tastes and goals dictate. Digital technology has moved from theory
to stumbling first attempts to a high degree of effectiveness and quality
performance. Audio software is increasingly transparent, where appropriate,
and, for many end users, digital emulations of analog circuits can be
indistinguishable from the circuits they are modeling. Microphone design can
now be considered an exercise of tailoring character as opposed to wrangling
with physics.
An underlying theme in those statements is that we can
achieve largely transparent gear designs when we want to. And often, we don’t
want to. An exception to that exception is in the tail end of the audio chain.
Loudspeaker technology has improved over the years, to be sure. New materials
and the advanced application of the various sciences involved in moving large
volumes of air are getting us closer to the ability to achieve transparency in
sound reproduction. That said, when measured, the transparency we can achieve
with loudspeakers would appall us if we were forced to live within those
standards in earlier stages of the signal path, and ultimate transparency may
remain elusive till we discover some new way to move air.
This is especially true with headphones. Accurately
reproducing the full range of audio in a package strapped to our heads is a
daunting challenge. Giving the performance specifications for headphones a
gander should give first warning—the frequency response specs are typically not
qualified as falling within a range of levels, for instance. We know
loudspeaker technology isn’t perfect, but at least expect that the frequency
response rating for a given loudspeaker will fall within a window of, say +/- 3
dB SPL, a range specified in the sales literature. Not so with headphones.
If you were handed literature spec’ing a loudspeaker with
a frequency response of 30 Hz to 20 kHz, +/- 8 dB, you probably would reject it
from consideration without even listening. Headphones, based on our
measurements in our April 2010 Session Trial bench test evaluation of
headphones, routinely have that sort of deviation from flat performance—and the
response curves are jagged, with huge peaks and valleys, and often some
apparently deliberate LF boosts of massive proportion. Then there are
additional considerations like distortion performance and consistency in
performance between left and right ears.
Headphone evaluation is going to remain highly subjective
in practice for some time to come. The test data is interesting and somewhat
instructive, but in the end, even “pro” headphone performance varies so wildly
from the kind of standards we apply to other gear that objective measurement
has limited value. The questions then become, “What do I like?” and “What can I
work with?” We hope you’ll at least find the April bench test exercise in
objectivity as interesting as we did.