Syntrillium Backup Forums :: View topic - Help needed in measuring a polar pattern of mic.

brock · Posted - Wed Apr 03, 2002 4:24 am

I've been asked to do a crude test to get some idea of the polar pattern of a mic. One test frequency is to be used (1KHz) at 2ft away and 3 angles, 0, 45 degrees left and 45 degress right, some other angles might be thrown in. I've not done anything like this before so any help is appreciated. My idea is to use a signal generator to play the test tone and record the results from the mic (cannot use a PC to record, remote location) onto my Sony DCD100 portable DAT. Can CE2000 be used to give me some information on the polar pattern? I have limited equipment, but is there any other way of going about this?

SteveG · Location: United Kingdom

For a start, do this out of doors away from buildings - not in a room, where the wall reflections will much up the results. And you need to do more than 3 angles- try doing every 20 degrees from front to back. It's hardly worth bothering with the symmetrical 45 degree meaurements - they are likely to be pretty much the same! And polar patterns vary quite a bit according to the frequency you measure them at - they tend to get more directional at higher frequencies, so you might want to do the tests more than once. And I'd go for a constant distance between 4-6 feet and turn the microphone - this is much easier to do (causes major problems if you do it indoors though!).

You really ought to do a quick test at the 'grazing angle' - 90 degrees, as this will give you a big clue as to whether you are looking at a pressure or velocity response - this might actually tell you all you need to know!

Playing the results back into CE will let you measure the amplitudes of the responses at different frequencies, but you will have to plot these onto a piece of circular graph paper by hand. If you plot the dB numbers, you should end up with a meaningful graph. Just make sure that 0 degrees is perpendicular to the diaphragm.

Usually, it's much easier to just look up the manufacturer's website and see if they've published a polar diagram, though...

Edited by - SteveG on 04/03/2002 04:54:06 AM

brock · Posted - Wed Apr 03, 2002 5:25 am

Thanks for tips. The recordings will be done in a soundbooth, 15ft by 10ft, and I was hoping this will be Okay. The mic in questions is mounted in a plastic unit which also has lights and switches incorporated into it, it attaches to the headlining in a car so I have to test it as a complete finished unit. As far as I know this is an OEM mic which has no identification on it but I might be able to found out a bit more, but having said that I'm fairly certain it would have been tested without the plastic housing unit so a manufacturers plot might not be that accurate. I wish it was that easy!!

SteveG · Location: United Kingdom

If you've got to do it like that, then all bets are off! You can usually tell quite a bit about a mic from just looking at it though, but in this particular situation, other effects start to predominate...

The thing about any mic that's flush-mounted is that it will start to operate in the the pressure boundary at the surface of the flat mounting, and effectively have an even hemispherical response anyway, regardless of the original polar pattern. The frequency at which the polar pattern starts to deviate from hemispherical rather depends upon the dimensions of the box. This means that usually, you get a response that's far less space-dominated, and usually a considerable improvement over a self-cased one in the same situation as far as clarity, etc is concerned. I'd be quite intrigued to know if this is what actually happens, so can you post some results?

Steve

brock · Posted - Wed Apr 03, 2002 6:19 am

Yep, I'll let you know how I get on. Its for testing an in-vehicle voice recognition system, thankfully this doesn't directly control the car!!