How do you measure a microphone? In this short video, we demonstrate how to accurately measure a microphone’s frequency response using SoundCheck. You’ll see how to ensure an accurate microphone measurement by using two different techniques that separate the microphone’s response from the source speaker’s influence. We cover two key methods: the Equalized Speaker Method, which calibrates the speaker using a reference microphone for a flat response, and the Substitution Method, which compares the test microphone to a reference microphone to remove speaker effects.

Watch this Video Demo of Microphone Measurement Techniques

Learn More about How to Perform Microphone Measurements

Here are some additional resources:

• Free Microphone Frequency Response and Sensitivity test sequence: https://www.listeninc.com/products/test-sequences/free/microphone-frequency-response-and-sensitivity-test-sequence/
• Free Microphone Substitution Method sequence: https://www.listeninc.com/products/test-sequences/free/microphone-frequency-response-and-sensitivity-microphone-substitution-method/

Video Transcript: How to Measure Microphones

The challenge with measuring the frequency response of microphones is that usually, the microphone’s frequency response is better than the source speaker used to play the test signal to measure it. This means you need to use measurement techniques that ensure you’re measuring the microphone and not the speaker. SoundCheck offers two great ways to do this.

The first is what we call the Equalized Speaker Method. First we measure the source speaker with a calibrated reference omnidirectional microphone that has a flat frequency response. We apply the reciprocal as a equalization  curve for the speaker at a constant sound pressure level, then measure the microphone under test or DUT in the same position as the reference mic.

The second method is the substitution method. Here, we first measure the source speaker with a reference microphone, then the source speaker with the microphone under test in the same location. Then we subtract the microphone under test result from the reference microphone to show just the characteristics of our DUT. Unlike the Equalized Speaker Method, the substitution method is not performed at a constant SPL which can result in a poor signal to noise ratio where the source speaker response rolls off at low and high frequencies significantly.

Let’s take a look at these measurements

Pre-written sequences for both of these methods come with the SoundCheck installation, so we’re going to use those today.

I’ll start with the equalized speaker method. I’m using an SCM measurement microphone as my reference microphone, and I’m using the AmpConnect 621 audio interface because as well as microphone inputs, it has an integrated amplifier to power my source speaker – a single driver, wide-bandwidth, Cambridge Audio Minx. This makes it easier to repeatedly position the microphones by pointing them at the center of the speaker.

First we need to equalize the speaker. We’re going to use SoundCheck’s built-in calibration sequence to do this.

It plays our stepped sine signal, measures the frequency response… we have a few reflections here. We take the reciprocal of the response and equalize the  output of the signal generator and measure again. You can hear its boosting the low frequencies instead of it being + or – 20 db its + or -1 db or less…. Because the speaker’s non-linearity it will be slightly different at different levels so we can equalize the equalization to further refine the EQ curve.

Here’s our speaker and we can see the EQ curve and the residual correction after EQ… and you can also see the phase here.

Now we’re all calibrated, let’s make a measurement. We use the same stepped sine sweep played  through the speaker, capture the frequency response, calculate the sensitivity at 1kHz, and display our results on a graph.

We also have a couple of arbitrary limit steps here – you’d want to customize them based on your device. Here we go…

Here you see the frequency response and phase— as you can see, it is fairly flat. And down here, we have our frequency response aligned to zero dB at 1 kHz, which is the standard way of displaying microphone response, and we can see our measured sensitivity is about -5.5dBv/Pa.

Now, let’s look at our alternative method—the Microphone Substitution Method.

In this method, rather than using an equalized speaker we first play our stepped sine sweep and  measure with a calibrated reference microphone. We then substitute our microphone under test and measure again. Next we subtract out the measurement of the reference microphone to get the raw response of our microphone under test. This removes the effect of the speaker’s response on the measurement.

We need to make sure we have the test microphone and the reference microphone in exactly the same position to minimize variance in the measurement.

If we look at our test sequence, we can see that we have two  acquisition steps, one for the reference mic and one for the mic under test, and a curve division step using the division operation in linear mode.

Let’s run the sequence.

It reminds me to make sure my reference mic is calibrated, which it is, and it prompts me to put it in front of the mouth simulator… let’s do that. We’re using a source speaker instead, but you can use a mouth simulator if you prefer, they both fulfil the same function.

It’s acquired the data, and now it’s asking me to put the microphone under test in its place… I’ll try to get it as close to the same spot as possible.

Let’s continue. And it’s acquired the data, performed the division, and here is our response.

This lower graph here shows the responses of the two microphones – the blue line is the reference mic and the orange line is our mic under test. And this top graph here shows the normalized frequency response including phase after the division.

You can see the sensitivity at 1 kHz is -6 dB V/ Pa, which aligns well with our first measurement using the equalized speaker measurement.

So there you have it – two simple methods for measuring microphones that ensure any non-linearities in your speaker don’t affect the measurement. And of course these are some of the most basic microphone measurements that you can do in SoundCheck. We can also measure directional response, acoustic overload point, intermodulation distortion, signal-to-noise ratio, embedded mics, MEMS mics and more. If you have any questions, or would like to see more microphone measurements, please get in touch with us – we’d be happy to demonstrate!