Which measurement microphone?

The question is almost always “which model”. That is the second one. The first is: do you know your microphone’s deviation — as a file? A microphone with a known 6 dB rise is better for measuring than a flat one whose curve nobody has.

Why the file matters more than the model

A measurement compares what the loudspeaker emits with what arrives at the microphone. Everything the microphone itself bends is in the result — and the software cannot tell it apart from what the room did.

If the software knows the microphone’s curve, it subtracts it. If it does not, the microphone’s rise becomes an apparent rise of the room — and the auto EQ cuts it in the PA. You then equalise your microphone into your system, and the room ends up sounding like exactly the fault you set out to fix.

What a measurement microphone has to do

  • Pick up omnidirectionally. A cardioid weighs reflections differently from an ear in the room and turns every rotation into a different measurement. A vocal mic is unsuitable, however good it otherwise is.
  • Come with a calibration file. A UMIK-1 brings one along, measured individually for its serial number — that is the best case, better than any generic profile for the type, because two units of the same model differ from each other.
  • Enough signal-to-noise for the room you work in. In an empty hall almost anything will do; with the system quiet and the ventilation running, the noise floor decides whether anything down low is still measurable. More sensitivity helps there more than a flatter curve.
  • What you can skip: a class-1 microphone with a calibration certificate, as long as you do not need documentation for an authority or a promoter. For setting up a system, the known curve decides, not the certificate.

And if you have no calibration file?

Then more remains usable than you would think. Arrival times, polarity and path differences do not depend on the amplitude curve — so time alignment works uncalibrated too. The same goes for comparing two measurements from the same microphone: whatever changes between them is real.

Only the absolute frequency response becomes unreliable — and with it everything that derives a target curve from it. Anyone measuring uncalibrated should therefore read the auto EQ as a suggestion and, in doubt, allow less rather than more.

What AURIX does at this point

AURIX reads calibration files in REW format — one line per frequency, as supplied with a UMIK-1. Two profiles are built in (Pioneer APM7008 and MCACC), in case you have one of those receiver microphones lying around.

Which calibration is currently active is shown identically everywhere in the software — in System and in the RTA view. That sounds like a detail, but it is the one statement you use to check whether your measurement was corrected at all.

Try AURIX for 7 days

No credit card. Your own calibration files can be loaded during the trial too.

Read on