Topic: Is it possible to get plots/graphs of mic output?

[timP]

I am interested in seeing really how close my mics/ caps are to each other and am wondering if I could take/send them somewhere to get plots produced, ala: the ones that come with matched pairs of mics.........




just curious......
thanks

[rokpunk]

I am interested in seeing really how close my mics/ caps are to each other and am wondering if I could take/send them somewhere to get plots produced, ala: the ones that come with matched pairs of mics.........




just curious......
thanks

i have a copy of Smaart 6.0 we could play around with.
somehow we can produce plots with it...but i'm not 100% versed on the software yet, so we'd have to do some playing around.
maybe chris church can instruct us how this is done. i'd imagine it's not much more involved than pumping tones through a speaker and taking readings with each mic, then plotting them out. i know the software can do it...it's just a matter of figuring it out.

[Gutbucket]

I'd love to be able to do this since I've been doing the mess around with baffles, surfaces and diffractive shapes on my omnis.  Generally you are correct about a tone source and a turntable of some sort so you can measure the angle of the mic to the source.  The big problem in the real world is that you'd end up picking up so much garbage from other directions from reflections and room modes that the measurements would likely be meaningless.  Big name mic companies do this in an anecoic room (or in specialized, sound sealed boxes for frequency response plots only I think, not polar).  

Maybe you could rig up a speaker and your test mic outside at night, at the top of some tall festival mic stands, far away from sleeping neighbors.  That would be the poor man's anecoic room.  Stan Linkwitz tests the directionality of  the drivers and baffles used in his speaker designs that way.  Details are somewhere on his site, a great resouce for understanding the priciples of sound reproduciton thourougly investigated by a rigourous engineering mind.  I'd love to build some of his speakers someday.  

[DSatz]

timP, there are labs that can do this type of measurement for you, but you'll have to pay at least in the low three figures to get this service done, and it probably won't tell you much about the low frequencies.

You can try to do a "null test" if you have the equipment for it--a preamp or recorder that lets you invert the signal polarity of one microphone (or maybe an inline adapter that does the same thing--flipping pins 2 and 3 on an XLR), plus a way to mix the two outputs and fine-tune the gain so that they mostly cancel. Then you can listen to the difference signal and/or run a spectrum analysis on it.

Alternatively you can make up a set of repeatable test tones (third-octave-band-limited pink noise or whatever), play them through a loudspeaker (preferably without damaging the speaker, your ears, or your relationships with those who live with and around you), record the result through both of your microphones, and make a spectrum analysis of each track to compare.

You don't need (and won't get) absolute accuracy--your results won't match the manufacturer's spec sheet curves even if those curves are perfectly honest (which most are not)--and you won't get meaningful results near the extremes of the frequency range, but you should be able to get a reasonable idea of how well matched your mikes are in overall sensitivity and frequency response in the midrange.

If you're going to record both mikes at the same time, they can't be in exactly the same point in the room pointing in the same direction. In that case I'd suggest making several sets of measurements so that you can see how much of the difference in the results is a matter of one or two inches difference in the microphone positions, or other more or less random factors.

--best regards

[John Willett]

You can try to do a "null test" if you have the equipment for it--a preamp or recorder that lets you invert the signal polarity of one microphone (or maybe an inline adapter that does the same thing--flipping pins 2 and 3 on an XLR), plus a way to mix the two outputs and fine-tune the gain so that they mostly cancel. Then you can listen to the difference signal and/or run a spectrum analysis on it.

This is a good way to check how close two mics are.

Face to face, very close, one with the reverse-polarity of the other.

With the same signal, in a perfect world, adjusted equal, you should get zero output.

What you do get tells you how different they are.

Simple and inexpensive. 

[timP]

I am interested in seeing really how close my mics/ caps are to each other and am wondering if I could take/send them somewhere to get plots produced, ala: the ones that come with matched pairs of mics.........




just curious......
thanks

i have a copy of Smaart 6.0 we could play around with.
somehow we can produce plots with it...but i'm not 100% versed on the software yet, so we'd have to do some playing around.
maybe chris church can instruct us how this is done. i'd imagine it's not much more involved than pumping tones through a speaker and taking readings with each mic, then plotting them out. i know the software can do it...it's just a matter of figuring it out.

hmmm...... maybe next week one day before i go to work we could play around?

I have these 4 card caps and I'd lke to pair up the closest ones
I'll bring lunch

++++Ts for all the info!

[DSatz]

Moke, you don't aim the microphones apart at all--you aim them in parallel at the sound source, and you put them as close together as you can. The idea is to feed them exactly the same acoustic signal.

But then you take their two outputs, and you invert the signal polarity of only one. Since the two microphones ideally put out identical signals, these two signals are now "equal and opposite." As a result, if you mix the signals together (possibly in your editing software) you will get a result that is relatively close to zero.

But it's never exactly zero. To the extent that the two microphones received exactly the same acoustical input at the same moment, the residue will tell you how much difference there is between the two microphones, and in what region of the frequency spectrum that difference mainly occurs.

It's a little similar to what happens in M/S matrixing, with the right channel output being "M-S"--except that here if you have two microphones numbered 1 and 2, the signal that you're trying to obtain will be either 1-2 or 2-1 (it doesn't matter which).

--best regards