Announcement

Collapse
No announcement yet.

Measuring Microphone Frequency Response?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Measuring Microphone Frequency Response?

    Craig;

    Your two posts for measuring tape recorder responses were interesting to me. I wonder if you have any ideas about measuring microphone frequency responses using DC6, or at least a method to get an approximation to the actual response a microphone may have.

    I have considered, but not tried, clamping a high quality set of headphones to the mike head and reading back the test response, though the error of the headphones would be part of that response. I assume white noise would be the sources signal. Not really sure how microphones are really tested, and have always been skeptical of the advertised frequency response.

    I have a collection of mikes and it would be interesting to do some kind of test on them to see how they compare with some type of methodical test.

    - Phillip
    Last edited by Craig Maier; 05-23-2019, 11:15 AM.

  • #2
    Wow! You have chosen to ask a very complex question which has no simple answers to it. I do not know where to start with this one but to say what I just said!

    First off, microphones and their frequency response, sensitivity and distortion are very much dependent on the location of the mic with respect to the sound source and also the effect of the surrounding room environment in which it is tested and the SPL level introduced. The polar pattern of a mic comes directly into play with regard to sensitivity and frequency response and the relationship of the sound source to the mics central axis. Measurements made in the near field measurements will be different than those made farther away.

    In general, the problem is "what am I measuring here - - - the mic or the sound source." That is the first order problem. There is a physics basis for establishing what the SPL sensitivity of a mic will be based on a reciprocating piston (the Pistonphone) which produces a note (less than 1 KHz - often 250 Hz) under adiabatic compression conditions (no heat exchange occuring in the air - - - an assumption that is made using this process - - - P x V = Constant). But, this only reveals the sensitivity at relatively low frequencies.

    A method to consider (assuming that you have an anechoic chamber to work with or at least a very dead room) would involve the use of a swept sine wave or random noise (like those produced by the Diamond Cut Make Waves Generator) in conjunction with two mics. One mic could be the microphone under test while the other would be a calibrated reference microphone (like a Bruel and Kjaer). One instance of Diamond Cut could produce a signal into an uncalibrated loudspeaker in the room. This would be done while both microphones are used to measure the swept sine or the random noise signal at aproximately the same location in the chamber or room. A second instance of Diamond Cut could be used with its spectrum analyzer in dual trace mode with one trace displaying the reference microphone while the other is measuring the UUT (Unit Under Test). The algebraic difference between the two signals could be plotted out and that would represent the actual response of the microphone under test assuming tha the reference mic response is known to be flat. But, this test would need to be performed at various locations along the polar axis of the mic to present accurate data as to the total overall frequency response performance of the mic.

    Consider that their are also important paramaters associated with the mic like:

    Dynamic Range (wherein the SPL level effects the frequency response)

    Sensitivity

    S/N Ratio

    %THD (again wherein the SPL level effects the distortion)

    Off axis rejection

    etc.

    There is much more to this that can easily be addressed in this kind of a posting. At the end of the day, this is not an easy measurement to make in your home, unlike the measurement of a tape recorders performance. I hope that you appreciate the complexities. If you want me to write more on this, please advise and I will provide what information that I can on this topic, keeping in mind that I am not much of an expert on this topic.

    -----------------------------------------------

    ps - Radio Shack used to make an inexpensive Sound Level Meter, Realistic Cat. No. 33-2050. It had an advertised basic accuracy of +/- 2 dB @ 114 dB Sound Level. Its range went from 50 dB to 126 dB. Its specs show that it should be flat from 20 Hz to around 10 KHz +/- around 3 dB.

    Someone named Marek Roland-Mieszkowski, M.Sc., Ph.D. ran a test on this meter against a Bruel and Kjaer (B&K) model No. 2203 reference Sound Level Meter and published the results on the internet some time ago. The two were compared from 125 Hz through 20 KHz based on a test recording. The deviation for the B&K was Min 0.0 dB, Max 6.0 dB. The deviation for the Radio Shack was Min 0.5 dB, Max 5.5 dB. Obviously, some of that deviation is due to the test setup's frequency non-linearity. But, these two instruments match quite nicely, especially when you consider the price difference!

    On an absolute level basis, a calibrated tone was applied to both meters of 94 dB @ 1000 Hz from a B&K calibrator. The B&K meter registered 95.5 dB while the Radio Shack registered 94 dB!!!

    Nice work, Radio Shack. The RS cost me around $50.00 (the B&K is at least $1000.00). Maybe this is something that you should consider obtaining. I have had mine for around 15 years now with no trouble whatsoever with the unit (except that I often forget to turn it off and kill the battery).

    pss - I just checked on eBay, and you can pick up one of these meters used for around $10.00.
    Last edited by Craig Maier; 04-04-2006, 08:42 PM.
    "Who put orange juice in my orange juice?" - - - William Claude Dukenfield

    Comment

    Working...
    X