[gutz]

Hi Guys ,
I should be getting my calibrated ecm8000 from cs in a couple of days

i also own a non-calibrated ecm8000
and i wanted to know if its possible to semi-calibrate my old ecm8000 with the calibrated one
of course it won't be the same as professional calibration but it should bring it to a semi-calibration level right?

has anyone tried it?
im planing on doing it at home ( regular not-treated room ) using white-noise
and a tripod or something to keep the placement as close as possible when i switch between the mics
should i gate the measurment? to avoid as many reflections as possible?

thanks in advance ,
gutz

 

[brucek]

I don't like to discourage you, but it's near impossible to do what you suggest at anything but the lower frequencies. You need to be very cautious in using measurement data to derive corrections.

In areas where the response is changing rapidly (especially where it is dropping) very slight differences in the positions of the mic elements will produce significant differences in the measurements, but these are due to the positional differences and not the inherent responses of the microphones.

Note that it is the microphone capsules that need to be at the same location, which are usually located just inside the tip of the mic. Measuring a response at low frequencies or one that is already fairly smooth (e.g. after correcting resonances with the BFD, or in a heavily treated room) would give more reliable data.

At higher frequencies where comb filtering will make it impossible, I wouldn't attempt it. You'd be closer by simply using the generic cal file we offer.

brucek

 

[Wayne A. Pflughaupt]

At higher frequencies where comb filtering will make it impossible, I wouldn't attempt it. You'd be closer by simply using the generic cal file we offer.

You don't think it might work using a constant noise source like white or pink noise)? I doesn't seem like comb filtering would be an issue of the RTA graph was smoothed for 1/6 or 1/3 octave. The only issue I can see is that something like a 1/3-octave EQ would be needed to get the graph as flat as possible before the comparison. That would make it easy to see the difference when switching to the uncalibrated mic. Sure, it's not perfect but it seems to me like it would better than "hit or miss" with the generic calibration file. Without the prior EQing, though, it seems to me like it would be difficult to do this.

Regards,
Wayne

 

[brucek]

You don't think it might work

No, I really don't think it would produce very good results. This comes from experience when Sonnie and I use to create the cal files ourselves for the different Radio Shack meters, compared against a calibrated ECM. Once you got above about 100Hz it was impossible.

brucek

 

[Speedskater]

Earthworks has a paper on how to do it.
ARTA also has some information.
But at the end of the day it's cheaper, more accurate and less time consuming to send the mic out for calibration.

 

[Anechoic]

I talk about how I do my measurements here. There's no resaon why you couldn't do the same with your cal'd mic.

A couple of things to look out for:

- placement of the reference mic and the mic under test (MUT) are critical. The placement is more critical the closer you are to the sound source, which is one of the reason why I measure 20-30 inches away from the source. The tradeoff with measuring away from the source is the greater influence of reflections if you're measuring in a room.

- accurate free-field low-frequency measurements below ~100 Hz are extremely difficult to perform in a room unless you know how to correct for the room effects. Placement is also especially critical at low frequencies. It may only be a difference of milimeters between the peak at one frequency and a null at another.

 

[gutz]

wow great thanks alot guys
more than i expected to hear
if i can get pretty good results above 100hz at 1/3 octave resolution it would be great

Anechoic -
what about near-field mesurments? will it be more accurate this way or worse?
im talking about taking a full-range driver ( or 3 deticated frequency-range drivers and prefrorm 3 tests ) and place the mic as close as possible to the drivers

lets say i want to get as accurate as i can above 100hz
should i gate the measuments? if so what gate to use? 5ms?
go with quick sweep?
or just plain non gated white-noise?

Speedskater -
i live in israel so actually getting the mic calibrated costs more than buying a calibrated one ( due to shipping )

 

[bjs]

Hi Guys ,
has anyone tried it?

Yes. I have a couple calibrated microphones and a couple non-calibrated ones. After my "home calibration" the non-calibrated microphones are as good as the calibrated ones. It is relatively easy and works well but one does have to take some care in putting together the test setup.

I made sure the capsule was in exactly the same place. Also I used a speaker with smooth response and no crossover (to avoid sharp cancellation nulls). I used REW's standard Frequency Response measurement and gated the window at the first reflection. I also measured at about 0.5 meter from the driver as getting too close can cause strong variations in frequency with position (above the bass frequencies). If you can't mount and remount the same calibrated microphone with exactly the same results then your test setup needs more work. I can't remember other details but I believe those were the main ones. It worked perfectly!

 

[Speedskater]

I can post links now

First read the Earthworks paper:

http://www.earthworksaudio.com/tech/ew_measures_mics.pdf

 

[Anechoic]

Anechoic -
what about near-field mesurments? will it be more accurate this way or worse?
im talking about taking a full-range driver ( or 3 deticated frequency-range drivers and prefrorm 3 tests ) and place the mic as close as possible to the drivers

The closer to the mic you get, the smaller your margin for error is. You don't want to get too close to the speaker or you may get weird effects from the modal behavior of the speaker cone.

lets say i want to get as accurate as i can above 100hz
should i gate the measuments? if so what gate to use? 5ms? go with quick sweep?
or just plain non gated white-noise?

If you're doing this in a room, should use some gated impulse-response method (sweep, mls, impulses, etc). As for how long the gate should be, that will depend on your exact setup. 5ms is too short to get you to 100 Hz, if you want 100 Hz the gate should be at least 10ms (or maybe a few ms longer).

 

[gutz]

thx guys
i got the info , now i need to work