Q&A: What Measurement Mic To Buy?

This post is part of a series based on questions received from readers and colleagues.

Questions may be edited for clarity, general applicability, or to protect the anonymity of the individual.

If you would like to ask a question, please email me via the Contact page.

Q: I want to put together a room tuning / SPL monitoring rig to help improve the consistency of my mixes. I was hoping you could help me out.

My main confusion is what mic to get (my current choice is the dbx RTA-M) and whether or not I need a calibration file, what a calibration file does, where do I get it and how do I use it.

Is the setup is simply just plugging the mic into my laptop via an interface and routing it into my measurement software (OpenSoundMeter) or is there's anything else I need?

The gear requirement for basic system tuning measurements is very minimal. You need a measurement mic (the dbx RTA-M is fine for basic tasks, although the build quality may mean that it’s not a lasting investment - for that you’ll need to spend more), an audio interface, and an audio analyzer program. Since modern sound system analysis is overwhelmingly done via realtime dual channel measurements, you will want a two-channel interface (two inputs and two outputs) at a minimum. The smallest and least expensive interface I use is the Focusrite Scarlett Solo. It has two inputs (mic and reference in) and two outputs (to system and to reference). You can currently purchase a Scarlett Solo for about $120 in the US. (I also like that the body of the interface is metal, which makes it more durable than some of the other commonly-used budget interfaces made of plastic.) Of course if you wish to take measurements with multiple microphones simultaneously you will need an interface with more microphone inputs.

(Although I would typically spec a four-microphone rig with wireless transmitters for a large venue touring package, virtually all of my fly-date / one-off work is done with a single measurement microphone. Fast and easy to set up when time is of the essence and doesn’t take up much room in my carry-on.)

People are often surprised to learn that for general system tuning work, most of the measurement mics on the market can be considered functionally interchangeable. The image below shows a direct comparison between 5 different measurement microphones, made by 4 different manufacturers, all in the same spot measuring the same loudspeaker. The microphones range in cost from $80 to $1300. Can you tell which is which?

You also mentioned SPL measurement, in which case your requirements flesh out a bit. First, assuming you plan to measure at typical live music concert levels, you will be exposing your measurement mic to significant pressure. My research has shown that as a general rule of thumb, we can expect to observe a Max Peak C/Z value approximately 32 dB above the average long-term A-Weighted levels. A typical concert at 99 dB LAeq would show a Max Peak C level somewhere in the low 130's. Thus, if we want our measurement system to maintain linearity, we need a mic with a max SPL rating of 135 - 140 dB, and most of the budget mics are out of the game long before that (you can’t blame them really). Typically this puts you at around the $350 price point for a measurement mic (I recommend something like the iSemCon EMX-7150 or the Audix TM1), or potentially substantially more if you're looking at models from Earthworks or DPA.

You also want to pay attention to the sensitivity of these microphones (at 135 dB SPL, what voltage is produced by the microphone?). I mentioned this idea in Chapter 12 of my book, but I’ll share more detail here: Most measurement microphones are relatively low sensitivity, producing about 1Vrms or less at such an SPL, and so are comfortably accommodated by most audio interface preamps.

As an example, according to the manufacturer's specification, the iSemCon EMX-7150 has a sensitivity of 6 mV / Pa. The “Pa” stands for Pascal, which is a unit of pressure equal to 94 dB SPL, and since 140 dB (the max SPL rating of this mic) is 200x more than 94 dB in linear terms, the output at 140 dB would be 200 times 6 mV, or 1.2Vrms (about +4 dBu). The Scarlett Solo has a stated max input level of +9.5 dBu, so this mic and this interface will work well together at concert levels.

If you were to use a high-sensitivity microphone such as the Earthworks M23, which is also rated for 140 dB SPL maximum, the sensitivity of 34 mV / Pa means that 140 dB gives an output voltage of 6.8Vrms, or almost + 19 dBu, which is substantially hotter than a typical board mix leaving a mixing console and will clip most audio interface mic preamps and convertors. So you either want a low-sensitivity measurement mic, or you want an audio interface with a high enough input capability to accommodate the high-sensitivity mic. 

Note that this isn't just about peak behavior. The now-discontinued PreSonus AudioBox USB used to be pretty popular for measurement applications, and its maximum input level of -3 dBu meant that clipping was a frequent occurrence at show level with a sensitive measurement mic. Since clipping the A/D converter invalidates all the data in your analyzer's Leq averaging buffers, this is a situation best avoided.

You will also need a sound level calibrator. In order to measure SPL, the analyzer needs to know what acoustic level caused the digital signal level that it's seeing.

Note that SPL is the only type of measurement that needs calibration. If you're not taking SPL measurements and only using your rig for system alignment work (transfer function measurements) or spectrum measurements (RTA) you do not need a calibrator. You just need your microphone and interface.

The sound level calibrator fits over your microphone and plays a known (lab-calibrated) reference level into the microphone, which you can then use to "train" your analyzer. The industry standard calibrator is the iSemCon SC-1. There are cheaper models available on Amazon, but make sure that whatever you're getting comes with the lab certification paperwork, otherwise there's no way to tell whether it's actually telling you the truth, which defeats the purpose. Calibration and data retention can become a much more complicated and expensive undertaking if you are working in situations where your data needs to be legally admissible (expert witness, consultation, enforcement, litigation, etc).

Finally, you asked about correction files. I am going to clarify the terminology I'm using here - when I say "calibration" that refers to the above-mentioned process of using a sound level calibrator to calibrate your measurement rig for SPL measurement. A correction file is something completely different - it's a file included with the microphone and generated by the manufacturer that describes the magnitude response of your microphone - how it deviates from "flat" - that you can load into your analyzer so it can offset the data at those frequencies and compensate. Although some people also call this a "calibration file" I always use the term “correction file” so as not to have it conflated with the SPL calibration process. This is a common source of confusion among people learning about measurement. 

The scoop on the correction file is that you probably don't need it. The top of the image below shows two measurements taken from the same position - with (red) and without (black) the correction file loaded. The bottom shows the change in the trace caused by moving the microphone one foot.

As you can see, the tiny changes made by the correction file are absolutely swamped by natural seat to seat variance in a real-world space. If you were working in loudspeaker design and R&D taking measurements in a laboratory environment, extremely accurate measurements are very important but for the work we do it's not worth my time worrying about. If I'm using multiple microphones, I start my work by putting them all in the same spot and making sure all the traces agree, then I don't have to think about it again.

The only time I've ever used correction files was to create custom ones designed to match the HF responses of several different models of wireless measurement transmitters provided to me by a vendor. But that's not an issue with the microphones, it's an issue with the wireless gear. 

However there is a benefit to purchasing a microphone that includes a correction file - it means the manufacturer measured and tested your microphone before they shipped it to you, so we know it's been checked. I like to say that if your microphone is of high enough quality to include a correction file, it's probably of high enough quality not to need it.

Note: Both of the graphics in this article can be found in Chapter 12 of my book, and are available on the Resources tab of my website.