How do you Match Mics? key test ect...

[Booda]

I bought 6 mics (sdc) w/ the hopes to get 2 closely matched pairs... one for me and one for a friend. How does this method sound too you?

Record w/ one of my better mics (cemc6) some keys jangling.
Set up a mic stand about a foot and a half away from one of my Monitors.
Play and Record the jangling keys w/ all 6 Mics (1 at a time)
Play and record sine wave clips (50hz, 100hz, 125hz, 150hz). w/ all 6 Mics. (1 at a time)
Run the wav. files through a spectrum analyzer (I use SX3)
and compare the results.

This way all the variables would stay constant.
I'm thinking I could do the recording w/ each mic as one WAV. file that would include the 4 Sine waves and the key jangle. The SA does a analysis of the whole wav. I could record them all individually but I'm thinking it's not necessary. This way I'd just have 6 files to analyze/compare rather than 30.

I'd basically look for 2 files that have similar output and visual EQ graph.

Any thoughts? Thanks,
B.

 

[mshilarious]

Don't bother with anything that's not full-range and repeatable. The key jangle is a good suggestion of Harvey's if you're in a store and have to use what you have. It's important to focus on high frequencies, such as those that keys generate, because that's where stereo location information comes from. But in your studio, you can just use any repeatable source that is full-spectrum. I just use white noise.

A full-range frequency sweep works too; that is used in a pro lab when they generate plotted charts, but that's because the output voltage is used to drive the plotter. I can't measure any difference between white noise and a frequency sweep with FFT. In theory, if you mic resonates at a particular frequency, but then a frequency sweep is probably still too fast, and let's hope your mics don't resonate!)

Anyway, a good 10 sec. of white noise--the same 10 sec. of white noise. Analyze RMS level and FFT--at a high level of resolution on the FFT. Pay more attention to frequencies above 1kHz than below. The RMS analysis will help you quickly focus on which mics are more likely to match, but ultimately the FFT is the judge.

Use a speaker to mic distance of about 2 ft. Too close, and if you don't get each mic within 1 or 2 mm of the same place, you'll get variation. If you have room treatments, bring 'em in close. That doesn't matter a whole lot, because whatever flaws your room has, it has for all the mics. But it looks nicer on the chart

Don't go too crazy, because you will have at least 0.1dBRMS variation between samples of the same mic. If you get within 0.3dBRMS and nice close frequency response--it won't be exactly the same level, but it shouldn't have any peaks headed opposite directions--that's a success.

As an addendum, my suggestion for store matching: ask to borrow one of the store's tuners or metronomes, something with a A=440 tone. Those tones aren't sine waves, they are usually rather rich triangle or sawtooth waves, lots of upper harmonics, and no bass to throw things off. Match level using meters on the pre, then take the mics close in level and compare with one channel reverse polarity (null test).

 

[Booda]

Thanks MS! That worked much better than my idea. I have a CD w/ Pink Noise on it and now I know more about what White and Pink noise is all about as well. I was just using the Pink noise to find the "Sweet Spot" at my listening positions, not really knowing what all the Hiss was about.
& you are so right about the 1k mark. All the mics were very similar under 1k... over 1k is where they started having some variation. Actually I was pretty impressed w/ the consistency of these mics. There was only 1 that was quite different from the rest.

I started off by listening to the recorded Pink noise and made notes of which mics sounded similar... then analyzed the graphs and came up w/ 2 pretty close matched pair. Not sure if this makes a difference but I'm thinking the pair that has the smoothest graph (smallest peaks and valleys) would be the better pair.

Thanks for the Help!
B.