Measurement Mics – How Close Are They?
By Pat Brown
Pat Brown did some tests and compared some popular measurement mics and posted the data. Is the expensive mic worth it? This article will help you decide.
Recent discussions within the SynAudCon email group raised some questions regarding device tolerances for both loudspeakers and microphones. This was prompted by the article “Loudspeaker Pairs: How Close Are They?”.
Part 1 – A Random Comparison
I decided to follow-up on the “Loudspeaker Pairs” article with a comparison of some popular measurement microphones. While there are many ways to approach such a study, I used the following procedure:
- 1. Measure the response of a reference loudspeaker (a Bag End M6 reference monitor) at 1 meter with each microphone. Care was taken to assure that each mic had the capsule in the same position in space and that the mic was at the same 45-degree angle. The angle was chosen based on the common assumption that this produces the flattest response for the microphone, as it compensates for increased directivity at very high frequencies..
- 2. Designate one of the mics as “golden” and compare each of the other responses to it.
- 3. Normalize all to the same level to eliminate the differences in sensitivity.
- 4. Smooth each response to remove fine details that are due to reflections from cables, etc.
I chose the SoundFirst SF101 as the golden mic, since I have a calibration curve for its ACO-Pacific capsule. The curves in Figure 1 were made by dividing the SF101 response by the response of each microphone-under-test (MUT). What we are seeing is how each mic differs from the SF101. Note that the loudspeaker response cancels out altogether, since it is common to each measurement, and that I did not identify the individual microphones. As can be seen there is a lot of variance in the responses.
Part 2 – A Refined Comparison
I next decided to compensate for the angle of the microphone to the loudspeaker. Oddly enough the proper incidence angle is seldom specified for measurement microphones. To find the optimal angle for each I proceeded as follows:
- 1. Selected a mic for which I had a factory calibration curve, and that was flat in response.
- 2. Rotated the mic from 0deg to 90deg while observing the change in response. The DPA 4090 was the most consistent over the angles tested.
- 3. Used this as the “golden” mic at 0deg incidence to measure the reference loudspeaker, assuming that this is the “most correct” response.
- 4. Using that curve as my reference, I measured each mic again and rotated it until the least deviation from the reference response was attained. The angle was rounded to the nearest 45deg for practical reasons. The largest round-off was 10deg. I then noted that angle in Table 1. Figure 2 shows the overlaid difference curves for all of the microphones. It is essentially a repeat of Figure 1 but with each microphone positioned at the angle indicated in Table 1. I also smoothed the curves at 1/3-octave to minimize the effect of reflections from the cable and mic body.
– Once the angle and cable reflections are accounted for, the mics all fit within a 2dB window.
– The very fine comb-filtering in the responses show how the mic body/stand reflections of the MUT differ from those of the DPA 4090. This doesn’t necessarily mean that the 4090 has the smallest reflection signature, it just shows how the other mics differed from it.
– Many measurement platforms allow for microphone correction curves. A correction curve can be produced for each of these mics that would make it indistinguishable from the “golden” microphone. While mandatory for lab work, it is seldom done for field work and general system tuning, where the bottom line is what it sounds like rather than an absolutely correct response.
– The response deviations approached 10dB when the same angle was used for all microphones. This was reduced to 2dB when the correct angle was used for each. It’s important to know the recommended angle of incidence for your measurement microphone! – The price range of these mics vary by a factor of 10, yet all produce similar results when used properly.
– There are numerous other factors that distinguish one mic from another and can easily justify the price of a more expensive model. This article only considered the axial response. pb