Data AcQuisition And Real-Time AnalysisScope - Spectrum - Spectrogram - Signal Generator
Software for Windows
Science with your Sound Card!
Contact us about
Sound Card Microphone and Speaker Calibration
The complete process described here is a transfer calibration. This is the process of using a calibrated reference microphone to calibrate other microphones. You might want to do this if you only have one (expensive) reference unit but need multiple calibrated microphones. You might also use this if you have a calibrated microphone that must be used with a probe tube that alters the frequency response.
In the following discussion, "microphone" and "speaker" will be used to mean "input transducer" and "output driver". These don't need to be conventional microphones or speakers, and in fact don't even need to deal with sound waves.
The basic idea is to use the calibrated microphone to obtain the calibrated frequency response of a speaker, and then to use that calibrated speaker to calibrate the test microphone. If you don't need to calibrate a test microphone, you can stop after you obtain the speaker calibration.
You will need a way to hold the microphone in a fixed position relative to the speaker, such that you can switch between the reference and test microphones without changing the sound field. One way to this is to have a single microphone holder that is rigidly mounted, allowing you to remove one unit and replace it with another while keeping the exact spacing and orientation with the speaker.
In some situations you can use a small sealed chamber attached to the front of a driver, with a fitting to hold the microphone via threads, setscrew, or even press-fit into a rubber gasket. CAUTION: The chamber must also have a bleeder valve or port you can open when you are sliding the microphone into place, or the pressure build-up will deflect the microphone and driver diaphragms. This may permanently destroy powered condenser microphones. When the chamber is sealed, the driver response is extended essentially to 0 hertz, so this is excellent for low-frequency work with pressure-response microphones. However, free-field microphones will usually give erroneous responses in this situation.
Another way is to mount the reference and test microphones together at the same time, so you only have to switch them electrically. You can put them side by side and point them at the center of the speaker such that they are equidistant from the speaker axis. This relies on the sound field being symmetrical to insure that both experience identical sound. It is better at low frequencies, near the driver. At high frequencies or farther away there is a greater chance for room reflections to disturb the symmetry.
A final method where no fixed mounting structure is available it to just make careful distance and alignment measurements to insure that the units are in the same positions each time. You will still need some way to hold the microphone in position, but this can be an ordinary microphone stand as long as you can get the alignment the same each time. You might want to try taping the base of the stand to the floor, and taping or clamping a laser pointer on the head below the mic. The laser beam should hit the same spot on the speaker cone when everything is aligned. You'll still need to measure distance very carefully.
Whichever method you use, note that high frequencies are always more difficult to measure accurately, since they are much more sensitive to small alignment and distance changes.
Since (unless you use the sealed chamber approach) the microphone is exposed to room reflections as well as direct sound from the speaker, it is best to keep the mic fairly close to the speaker (a few inches). If that is not possible for some reason, try to insure that the surroundings are as acoustically absorbent as possible. Make sure the floor is carpeted, and consider making a tent of blankets to reduce wall and ceiling reflections. You may even want to do the calibration outside, away from buildings and over a grassy or rough dirt surface to provide an anechoic environment, as long as there isn't excessive background noise.
Once you have a suitable fixture or system, install the reference microphone and apply its Mic .CAL (or .FRD) file in the appropriate User Line dialog for your chosen Input line. The Output line should not have any .CAL or .FRD file applied at this time; that's what you are about to create.
Now you must obtain a frequency response curve of the system, which in this case will be the frequency response of the driver. For calibration purposes, it is probably best to use a slow stepped frequency sweep for the most reliable results.
When you have the complete response on the screen, go to the File menu and hit Save Y-log Trace as .CAL, .FRD, .CRV, or .LIM File. Save the response as a .CAL file (default) or .FRD file. (Do not save as a .CRV or .LIM file). This will be your speaker calibration file, so pick a descriptive name like RefSpkr.CAL.
Go to the output User Line dialog and load the file you just created under Load Speaker Cal File. If you want to verify that everything worked properly, you can re-run the same frequency response with both the Mic and Spkr .CAL files active in their respective User Lines. The response should be a perfectly flat line. (Don't save this response as a .CAL file... this was just for confirmation purposes.)
Replace the calibrated microphone with the test microphone, unless you are using the side-by-side method, in which case just change electrical connections to the input line.
Go to the microphone input User Line dialog and remove the calibrated Mic .CAL file by clicking on its button.
Now repeat the frequency response measurement procedure, only this time the .CAL file you create will be for the microphone.
It is important to note that the speaker .CAL file you have created is only for that particular setup; if you move the mic to another location, even with that same speaker, the response may vary. The sound from the speaker will of course be less as distance increases, but in addition there may be different room reflections. Also, the off-axis speaker response is typically quite different from on-axis.
So, if you are just interested in calibrating the speaker, you must calibrate it in exactly the situation you intend to use it. If the test subject will be at a certain distance and angle, you must put the microphone at that same location when you calibrate the speaker. Note that for living subjects, head and ear orientation can make a difference in loudness and frequency response.
For (say) exposing lab animals to simulated industrial noise, you may want to make response measurements at multiple locations in the cage. You won't be able to establish a precise noise level if the animal is free to move around, so it's best to have multiple response plots to demonstrate the range of exposure possible. Think twice before trying to come up with some sort of aggregate response, since the animal may be expected to orient itself to minimize discomfort.
On the other hand, if you are transfer-calibrating a microphone with this technique, it is not constrained to be used in a particular setup. Your newly-calibrated test microphone can be used in any application that the original reference microphone can be used in... within certain limitations. If the test microphone is the same make and model as the reference microphone, you can safely treat them as identical.
But different microphones have different off-axis responses, so if you have just calibrated with the test mic pointing directly at the speaker, you can't assume that when it is turned 90 degrees it will respond the same as the (unrelated) reference mic at 90 degrees. In particular, "unidirectional" microphones have quite drastic off-axis response changes compared to "omnidirectional"... always use the omnidirectional type for measurement purposes.
If you want to know how your test microphone behaves at other orientations, you can keep the same exact test setup and re-run the frequency response with the mic turned at various angles, but otherwise at the same location. (You no longer need the reference microphone; with the calibrated speaker, you know the "sound field" at that location.) You can save these response curves as .TXT files and use them with a spreadsheet program to produce polar response plots of sound level versus angle at specified frequencies.
What if you don't have a calibrated reference microphone? One inexpensive alternative for non-critical work is to purchase an omnidirectional electret microphone element for a few dollars. It is possible to find elements with very flat frequency responses at most audio frequencies.
You will need to solder the element to a stereo cable that will plug into your sound card Mic input. The Left channel is the signal to the sound card, and the Right is a DC voltage (about 3-5 volts) from the sound card to power the microphone. The element may have 2 or 3 terminals, and typically requires an added resistor and/or capacitor; see the documentation for the element.
Alternatively, if you are handy with electronics, you may choose to build your own preamp for use with the element, and feed its output to a Line input instead of Mic. That way you can avoid some of the compromises of the sound card's internal Mic preamp, which is often noisy and may also restrict the high frequency response. Again, see the element documentation for typical schematics, or browse the Web for complete microphone preamp designs.
Questions? Comments? Contact us!We respond to ALL inquiries, typically within 24 hrs.
Over 30 Years of Innovative Instrumentation
© Copyright 2007 - 2013 by Interstellar Research
All rights reserved