Cross-Spectrum Microphone Calibration Service - USA

[Sonnie]

We are happy to announce and recommend Cross-Spectrum Labs for microphone calibration in the U.S. ... particularly the Behringer ECM8000 and Galaxy CM-140. Of course, they will also calibrate other microphones as well.

Cross·Spectrum Labs
P.O. Box 90842
Springfield, MA 01139

I have personally spoken with Herb Singleton, founder of Cross-Spectrum, and he as agreed to calibrate these units 10Hz to 20KHz provided you specify you would like calibration to 10Hz. The charge is $55.00 and includes the cost of return shipping via USPS for customers in the continental United States..

Cross-Spectrum will use the pressure method to calibrate the low frequency response and the quasi-anechoic free-field on-axis method to calibrate the upper frequency response. The pressure method is one of the most accurate methods of measuring the frequency response of a microphone, however baffle diffraction must be accounted for when using this method. Generally mic manufacturers will provide the variance to offset the baffle diffraction. Unfortunately, Behringer nor Galaxy offers these variance corrections. Therefore, the free-field method of measuring the upper frequencies will be more accurate for these units. These combined calibration methods will give us the best calibration file for the ECM8000 and CM-140 mics. This is one of the first labs we have found that will calibrate the CM-140 SPL Meter microphone.

Cross-Spectrum also offers calibrated Behringer ECM-8000 microphones for sale. The cost is $100 + shipping. Please see their Behringer page for more information.

Herb will be joining us here at the Shack and has agreed to answer any questions anyone may have. Please check out their site before asking questions... you may very well find your answer. If you do not find the answer you are looking for, you may post a question here in this thread and Herb will do his best to get you an answer. Remember he is busy in the lab and does not monitor this forum 24/7, so please be patient... or you are more than happy to contact him by telephone (see their Contact Us page).

Remember, if you are wanting your mic calibrated to 10Hz, you must include a note requesting same, otherwise the calibration may only be to 20Hz.

We appreciate Herb's willingness to work with us in providing this service to our members here at the Shack.

[Anechoic]

Hi all,

Herb here. Sonnie covered most of the bases, but I wanted to stick my head in here and let you know that I am available to answer any questions. I will be here from time to time to check up on things, and I have subscribed to this thread, but it's likely that I may miss questions so feel free to email me or call my cell phone: 617-901-7472. (If I don't answer, *leave a message* and I promise I'll return your call. If you call multiple times and never leave a message, I guarantee I'll never answer it.) Also feel free to post any responses I email to you here for others to read.

I do want to mention that there are lots of things that go into mic cals, so I try to be as transparent as possible with respect to my reference equipment and techniques so you can evaluate the strengths and weaknesses for yourself. The one thing that I have learned from doing this is that when you look at lab to lab or manufacturer to manufacturer everyone seems to do it just a little differently - in fact I've learned how some of the Big Boys do it and it's funny how ghetto some of their methods are. I am on the AES working group SC-04-04 (Working Group on Microphone Measurement and Characterization) which is trying to update an old IEC standard on mic measurements. Hopefully this effort will result in more standardization in free-field microphone calibrations.

If you are interested in the cal service, please drop me an email and I'll give you the shipping address - I don't like to publish the shipping address because I've moved three times since I've started offering the service and my previous shipping addresses have wound up in various corners of the internet resulting in people shipping mics to my old addresses without letting me know. I generally try to turn them around in 48-hours but my normal duties require a lot of travel so I may not always get to it. I will let you know ahead of time if my schedule is tight.

Sonnie mentioned that many users here want 10 Hz to 24kHz cals. I'm happy to do that, but please indicate that you want that frequency range, otherwise I'm going to default to my normal 20 Hz to 20kHz cals.

I do sell pre-calibrated ECM8000, but lately it seems that everyone and their uncle wants one and I'm having a real hard time keeping up with the orders. Again, email or call ahead

Also, if anyone is interested in stepping up to professional measurement microphones (Bruel & Kjaer, GRAS, BSWA, ACO Pacific, Larson-Davis, etc), I have a lot of experience with them (in fact I own B&K, ACO, BSWA and Rionmics) so again feel free to ask.

Good luck all.

[brucek]

We've recently been discussing the proper orientation for a measurement mic such as the ECM8000 in this thread and have found a difference between vertical and horizontal orientation.

I understand that your calibrations are for a vertical orientation (EDIT incorrect - the default calibration is horizontal on-axis), but I wonder if the calibration is still accurate when measuring near-field (<= 12") or even at close diffuse soundfield distances such as 50". I did some tests of my tweeter in my loudspeakers here, but still used the horizontal orientation.

Can you comment on taking measurements at a listening position compared to near-field, with respect to mic orientation and a single calibration file.

brucek

[Ricci]

Welcome Herb. It's great to have someone with your experience around. You'll most likely be hearing from me soon about a mic calibration.

Can you calibrate mics even lower than 10hz? Could you do 5 or even 3hz?

[Anechoic]

I have to head out for the day, but I will address the questions as soon as I can.

[Anechoic]

Okay, it took me a while, but I'm back.

I understand that your calibrations are for a vertical orientation, but I wonder if the calibration is still accurate when measuring near-field (<= 12") or even at close diffuse soundfield distances such as 50". I did some tests of my tweeter in my loudspeakers here, but still used the horizontal orientation.

Just to be clear - my measurements results give the on-axis mic frequency response; that is, the mic diaphragm is facing the sound source.

Can you comment on taking measurements at a listening position compared to near-field, with respect to mic orientation and a single calibration file.

Well there's really no simple answer. I skimmed the thread you linked to and for the most part, the exposition there is on the right track. The important thing is that at the higher frequencies, the orientation of the microphone does matter. At the highest frequencies, the wavelength becomes comparable to the size of the microphone and therefore the mic does have an influence on the soundfield. For a 1/2-inch mic, the effects really start around 1 kHz but become noticeable around 2-4 kHz. A 1/4-inch or 3/8-inch mic (like the dbx RTA-M) gives you another octave or so before the effects become noticeable.

You shouldn't see any proximity affects in the high frequencies with condenser mics. You also shouldn't see any proximity effects in the low freqs with condenser mics like the ECM8000 - however the most recent runs of ECM8000 come with a chart that claims there is one, but I haven't observed it (although to be honest, I haven't investigated it all that much.) But since Behringer has never released a freq response curve for the ECM8000 that has agreed with what I and others have measured, I'm not going to put too much stock into it.

Since the response of most measurement mics will depend on the orientation, what to do? Well, if you're trying to measure the freq response of a particular source, measure with the mic pointed at the source (which will minimize the effect of the mic body) and apply your calibration curve to correct the response. If you're doing a room measurement for DRC or other purposes (STC or RT60 measurements) what you really should be using a diffuse field microphone (which pretty much no one does) or you should measure at various locations with a free-field mic at various orientations (as recommended in ASTM E336), and then average the responses together. Of course most of us don't have the time and/or patience to do that, so when I do room measurements, I put the mic in random incidence position (mic facing the source and angled 70-80 degrees with respect to the floor).

Of course the catch is that typically the cal files you have are for one orientation. If you really care that much about it, you can get polar measurements of the mic (which I can measure for additional cost) and apply the specific cal file, or just do some trial and error and see what works for you. As I wrote above, in practice I use random-incidence positioning and apply the cal file I get from my 3rd party cal lab and I've always gotten reasonable results.

If your measuring loudspeaker frequency response, you shouldn't have a diffuse field, you should be gating the results so that you only have the direct sound. If you do that, the cal file should apply no matter the distance (where the distance is limited to how long your gate is of course).

Can you calibrate mics even lower than 10hz? Could you do 5 or even 3hz?

Sonnie mentioned that a bunch of people here where interested in getting down below 10 Hz. I have put together a pressure field jig and I can get down to 5 Hz, but that's as low as I can go. Below that, you start getting into regions where ambient air pressures have an effect. If I want to get much below 5 Hz, I either need to invest in an electrostatic actuator or build an air-tight pressure chamber, either of which would be a substantial investment on my part and since mic measurements are a tiny part of what I do, it's not worth it at this point. However I may get lucky on eBay, so stay tuned.

[brucek]

Thanks for your detailed answer.

I guess our recommendations for measuring loudspeakers (near field) using a horizontal on-axis orientation, combined with proper gating, are correct.

Where we differ from your recommendations, is when measuring a room at the listening position. We have been advising members to use a vertical orientation for their ECM8000 mics, so as to take in all reflections equally, since the ECM8000 is an omni-directional mic. If I understand correctly, you're advising to face the source using an almost horizontal mic position (70-80deg). Since your calibrations are for on-axis, facing the source orientation, I guess that would make sense.

Do you offer calibrations for both vertical and horizontal positions of the mic?

brucek

[Anechoic]

If I understand correctly, you're advising to face the source using an almost horizontal mic position (70-80deg).

No, I'm saying use an almost-vertical orientation where "vertical" is 90 degrees and I rotate the mic toward the floor by 10-20 degrees (if "vertical" is 12-o'clock, I measure with the mic pointing at ~1 or 2-o'clock). There will be some error since the cal file is based on an horizontal alignment, but even with the "correct" cal file there will be some error at the highest frequencies regardless because you're using a freefield mic for what is essentially a diffuse-field measurement. But it's not something that causes me to lose a lot of sleep although I'm sure others feel differently.

Do you offer calibrations for both vertical and horizontal positions of the mic?

I can perform additional measurements at 45 and/or 90 degree angles for free, but any more than that (or full-blown polar measurements) is extra.

[brucek]

I'm saying use an almost-vertical orientation where "vertical" is 90 degrees and I rotate the mic toward the floor by 10-20 degrees

Ahh OK. I usually consider vertical to mean 0 degrees (straight up and down), and so your 70-80 degrees meant a near horizontal position to me. But, I get it now.

So we're not far apart on this then. I usually recommend vertical, and you're recommending vertical with a slight tilt (10-20deg) toward the mains speakers at the listening position. Seems fair. And again, since this is for home use for most of us, a few dB isn't going to cause any problems either way, since most of us are using a generic calibration file...

brucek

[Ricci]

Thanks for the detailed answers. I have an additional one. How much would a an ECM calibrated down to 5hz be with an additional 45 and 90 degree measurement? I'm most likely going to get on the waiting list for one.

[Anechoic]

And again, since this is for home use for most of us, a few dB isn't going to cause any problems either way, since most of us are using a generic calibration file...

Yep. When it comes to acoustical measurements the reality is that there are always limitations that we have to live with. For example, if you really want to know what's going on a 20 kHz (say, for a speaker measurement) you really need to be using a 1/4-inch mic, since a 1/2-inch mic is so big relative to the wavelength so that you get diffraction effects and the diaphragm doesn't necessarily act like a piston (there's a reason why IEC and ANSI standards for Type 1 mics go to -infinity above 16 kHz). If you're careful, you can get a "good enough" number up that high but it's not going to be anywhere near as accurate as what you're getting in the 100 Hz - 3 kHz range.

How much would a an ECM calibrated down to 5hz be with an additional 45 and 90 degree measurement? I'm most likely going to get on the waiting list for one.

I won't charge extra for that, but you'll have to ask for it when you order it. And yes, there is a waiting list - it's been really crazy lately, everytime I get some in stock, they immediately go out again.

[redliner]

What's the turnaround time once you're on the waiting list?

[Anechoic]

redliner wrote:

What's the turnaround time once you're on the waiting list?

It really depends on where you are on the queue and my schedule. Right now I have no back orders and zero stock (which should get refreshed early next week) so the first few orders I get now should go out no later than March 2.

[Dennis H]

Hi Herb,

Good to see you posting here. Do you have any guess about the high and low end roll-offs of the Behringer? Say X dB/oct at Y Hz? With a 96K sound card and a 128K FFT, you could be generating measurement files covering .7-48K Hz. It would be nice to have some non-zero numbers in the cal file at those extremes, even if they were just a wild guess, so you didn't have your CAD software assuming there was a sudden big jump in the response. Seems like that could lead to phase errors in your crossover design.

[Anechoic]

The ECM8000s have so much unit to unit variation (see here) that I'm not comfortable giving generalized numbers - well other than +/- 15 from 20Hz to 20kHz .

As for the extremes, yes it is possible with a long FFT and wideband sound card to give a wide frequency response, but acoustically it's very difficult to generate reliable test data at those extremes. I could give numbers at those frequencies, but they would be meaningless.

[lcaillo]

Amen to that!

[Dennis H]

Anechoic wrote:

The ECM8000s have so much unit to unit variation that I'm not comfortable giving generalized numbers - well other than +/- 15 from 20Hz to 20kHz .

As for the extremes, yes it is possible with a long FFT and wideband sound card to give a wide frequency response, but acoustically it's very difficult to generate reliable test data at those extremes. I could give numbers at those frequencies, but they would be meaningless.

Okay, forget the Behringer. Generally speaking, what is the slope of the roll-off of generic condenser mics on the high and low end? 2nd order? Steeper? I'm just trying to learn how these mics work.

[Anechoic]

Dennis H wrote:

I'm just trying to learn how these mics work.

So am I.

The short answer is "it depends on how the microphone and microphone pre-amp are designed." I have to head out for the day, but I'll give a more detailed explanation later this evening.

[Dennis H]

Thanks, Herb. To keep things simple, could we start with the response of a plain old electret capsule with no extra circuits? Most software takes 2-channel measurements these days so the response of the preamp gets subtracted out.

[Anechoic]

Dennis H wrote:

Thanks, Herb. To keep things simple, could we start with the response of a plain old electret capsule with no extra circuits? Most software takes 2-channel measurements these days so the response of the preamp gets subtracted out.

Just to make sure we're all on the same page: when I talk about the capsule, I'm refering to the "tip" of the microphone that has the microphone diaphragm, blackplate, isolator and electrical terminals(and grill if applicable). When I talk about the pre-amp, I'm not referring to the phantom power supply or the sound-card interface, I'm talking about the electronics that convert the varying charge into an output voltage and steps up the voltage. In the parlance of my reference equipment, this is the mic capsule, and this/this is the pre-amp. In mics like the dbx RTA-M or ECM8000, the unit contains both the mic and the preamp (which gets plugged into another pre-amp to provide phantom power and bump up the output more).

A condenser mic can go down to DC, in which case you could use it measure atmospheric pressure. In practice, the rear of the diaphragm is usually vented to prevent the diaphragm from breaking during sudden pressure changes or exposure to very high SPLs. This vent creates a low-frequency roll-off which limits the low-frequency performance of the microphone. The vent hole and the volume of air behind the diaphragm is a small resonant chamber; the size of these parameters determine the roll-off, so you need to know those parameters to predict the roll-off of any particular microphone.

The theoretical roll-off on the high frequencies for condenser mics (normal incidence) is 12 dB per octave if you only consider an undamped system. However, at mid-audio frequencies (8-13kHz depending on a variety of factors including diaphragm diameter) there is a resonance that manufacturers try to damp out though adjusting the vent sizes, grill design, diaphragm tension and diaphragm mass. This damping can change the slope of the high-frequency roll-off depending on the proximity of the resonance to the high-frequency limit of the mic. So if you want to use a number for the high-frequency rolloff, use 12dB/octave but it's going to be a wild guess at best.

We also have to account for the performance of the pre-amp response. Manufacturers may try to use the electronics to tweak the frequency response of the capsule, which may also have an effect on the rolloff.

HTH.

[bxxk]

Hi Herb,

I've built some mics before, including an ambisonic tetrahedral mic, and used a B&K reference mic to manage the calibration. But I would not mind my own for my home DRC work... I'm sure I'm not alone in wondering what you would recommend in various price ranges for a cheap or inexpensive reference mic.

The reason I'm asking is because the ECM8000 is not all that good (eg the bass roll off, which is not in the original Panasonic WM51a capsules, so is possibly in their preamp), even for the price - I believe the similar NADY is better and I've seen home made units which are better. Whatever mic is used will still need calibration, but using a better (flatter / quieter) mic to start with is usually a good idea.

Many thanks for your input, it is much appreciated.
BK

[Anechoic]

bxxk wrote:

Hi Herb,

I've built some mics before, including an ambisonic tetrahedral mic, and used a B&K reference mic to manage the calibration. But I would not mind my own for my home DRC work... I'm sure I'm not alone in wondering what you would recommend in various price ranges for a cheap or inexpensive reference mic.

It's kind of a broad question that's difficult to answer, but I'll try.

In the same (rough) price range as the ECM8000 I've measured the Nady CM 100 and the dbx RTA-M. The Nady is flatter to 20 kHz although I didn't look much lower. The RTA-M rolls off a bit above 20 Hz.

If you're willing to spend a couple of hundred dollars, the Josephson and Earthworks lines are great mics for the price, but pricewise are geared for the serious enthusiast. The Beyerdynamic MM 1 has a great reputation and can sometimes be found for under $200, but I've never personally handled one. There's the APEX220 which falls into the ECM8000 price range and may or may not be a rebadged RTA-M.

One microphone that I really like is the Radio Shack lapel microphone. They're cheap, but they are amazingly flat up to 20 kHz and much more consistent from unit to unit. The downside is that the noise floor is pretty high (around 60 dBA IIRC) which limits its use for low SPL testing. You also have to find a way to mount it such that the mounting doesn't muddy up the signal. Caveat: it's been several years since I last measured one so it's entirely possible that there has been a revision and the mic is now.

[Dennis H]

I think the Behringer and the Nady are both OEM'd from Superlux in China -- probably the same mic with a different label. They seem to have lost their flat response when Panasonic discontinued the WM-60 and everyone switched to the WM-61. I thought it was a Superlux problem but I recently read a paper where someone built 8 mics using the WM-61, 4 with the stock wiring and 4 with the 'Linkwitz mod'. They all had the 10K hump although the ones with the mod didn't hump quite as high. The ones with the mod also had lower distortion at high SPL.

Herb, the Panasonic electret capsules don't really require a preamp like you describe for a traditional condenser mic. They work fine with just a battery and a resistor as long as they are feeding a fairly high impedance preamp or sound card. Superlux is including more of a preamp built in to make them compatible with a pro mic preamp with low impedance and 48V.

I was looking at some measured 5-48K curves for the Behringer and it looks like the tails are more like 24dB/oct. The tails don't really matter for someone measuring a room but they could make a difference for someone designing a crossover because of errors in the calculated phase using a Hilbert transform. As an experiment, I built a fake cal file extending down to 1Hz and ran a Hilbert transform to get the phase. Then I truncated the file to 20Hz and ran the HBT again. The phase error at 20Hz was about 70 degrees with the truncated file.

[Anechoic]

Dennis H wrote:

I think the Behringer and the Nady are both OEM'd from Superlux in China -- probably the same mic with a different label.

There was a discussion a few years back on Usenet (Google groups search is completely fubar'd right now so I can't find it) where there was some dispute over that - I think the mic element was probably the same but the electronics are different. The ECM8000 electronics have changed a couple of times over the years so if may be that they're the same now.

I was looking at some measured 5-48K curves for the Behringer and it looks like the tails are more like 24dB/oct. The tails don't really matter for someone measuring a room but they could make a difference for someone designing a crossover because of errors in the calculated phase using a Hilbert transform. As an experiment, I built a fake cal file extending down to 1Hz and ran a Hilbert transform to get the phase. Then I truncated the file to 20Hz and ran the HBT again. The phase error at 20Hz was about 70 degrees with the truncated file.

You have to be careful with using Hilbert Transforms on condenser mic freq response data since condenser mics aren't minimum phase.

[Dennis H]

Anechoic wrote:

You have to be careful with using Hilbert Transforms on condenser mic freq response data since condenser mics aren't minimum phase.

Hmmm, interesting. I'm no expert but I can't imagine a mechanism where a mechanical transducer and analog electronics wouldn't be minimum phase. There have been a lot of discussions about that with speaker drivers, cone breakup modes, baffle diffraction, etc. and it always turns out (as far as I know) that it is behaving as a minimum phase system even though it doesn't appear that way at first. BUT.... the HBT is making assumptions and it's subject to computation errors if you don't do things like extending the tails beyond the measured values. You have to do that with speaker drivers and I'd imagine you have to do it with mics as well.