Recording - May 2015 - 46

By Paul J. Stamler

Part 5: About Microphones 2
[NOTE: Some of the pictures in this article are also in my forthcoming book Inside
the Box (Stipes Publications); they do double duty here with that publisher's kind permission. The other charts are courtesy of AKG, Electro-Voice and Neumann.]
How a microphone sounds depends a lot on how it responds to the different frequencies that make up the audio spectrum. But how can you know that?
The most obvious way is to use a variety of microphones on a variety of sound sources
over a period of years, until you know the best combinations by heart. That's certainly the
best way, and all the mics I brought to the Ottoman Underground session (remember, you
can grab the sound clip I reference in this article at http://is.gd/RECFirstStepsClip1 to hear
what I'm discussing) are ones I've used for a long time; I know their sounds intimately. But
in the real world, often we have to make a ballpark guess about which microphone to put
up at the beginning of a session, knowing we might replace it later.
One route to that ballpark guess is to use the mic's published specifications. But here
be dragons.

The numbers game
I have on my desk a retail catalog of professional audio gear, and it has a page
devoted entirely to Neumann microphones. The usual suspects are here: U 87, TLM
102, TLM 49-that crowd.
Each mic has a frequency response specification which I suspect came from the
retailer's copywriters rather than from Neumann; the frequency responses are all
described as "20 Hz-20 kHz", covering the entire audible range. Every one of them.
If that specification were a real indicator of what the microphones sound like, you'd
logically conclude that they all sound the same. Well, I've used most of these mics -
even reviewed two of them for Recording-and I can tell you that they're excellent.
But they don't sound alike, at all.
The first and most obvious problem with a spec like "20 Hz-20 kHz" is that it leaves
important information out. Jumping ahead a little, take a look at Figure 1, which charts
the frequency response of an imaginary microphone. It responds from 20 Hz-20 kHz,

RECORDING May 2015

all right, but the responses at 20 Hz and
20 kHz are 20 dB lower than the response
at middle frequencies like 1 kHz-so low
that the mic might as well not be responding to those frequencies at all.
A more informative spec, and one used
more often in the audio world, gives the
range of frequencies between the points at
which the mic's response is down 3 dB
from its output at 1 kHz. By that criterion, a
mic with a response like Fig. 1 might carry
a realistic spec of "50 Hz-10 kHz, -3 dB".
This is a much more useful and informative specification than the naked "20
Hz-20 kHz", with no tolerances specified. A spec like "50 Hz-10 kHz, ± 3dB"
would be still more informative-but here
be dragons again.
The problem is this: a spec like "50
Hz-10 kHz, ±3 dB" would be equally
true for frequency responses like Figure 1
and Figure 2, though these two microphones would sound quite different-the
mic in Figure 2 would sound brighter. In
the end, the engineer in search of "ballpark" information about a microphone's
frequency response is much better off with
a picture than with numbers. Dragons lurk
there, too.
Every picture tells a story
We've already jumped into frequency
response charts, but I'd like to backtrack
a little and talk about what these graphs
mean and how to read them.
A frequency response chart describes
what signal levels the microphone would
generate when exposed to a soundfield
with equal sound levels at all frequencies
audible to the (healthy) human ear. These
graphs are produced in three ways. The
oldest way is to generate a swept sine
wave tone and feed it to a loudspeaker
calibrated for precise response (costly),
with the microphone in question placed
with the speaker inside an anechoic chamber (large and very costly). The mic's output is connected to a chart recorder, synchronized with the frequency sweep.
A newer way to generate frequency
response charts is to expose the mic to an
impulse, a very short pulse of sound.
(One way to generate such an impulse is
to pump a lot of electrical current through
a very fine wire, which will make it
explode with a bang; another is to fire a

http://www.is.gd/RECFirstStepsClip1

Recording - May 2015

Table of Contents for the Digital Edition of Recording - May 2015