ECM8000 Tests (lots of pics)

**Anthony** · **03-24-08, 06:52 PM**

Okay, so BoomieMCT and I did a battery of tests on our ECM8000 mics.

The setup was my M-Audio MobilePre preamp, test amp, Magnepan MG10.1. Mic stand was set 3 feet away, 2' off the ground aligned with the ribbon perpendicular to the panel.

Nothing on the speaker changed during all the tests, the mic stand was kept horizontal for all the horizontal tests and vertical for those tests (all done at once to minimize any changes).

Measurements were full sweeps in roomEQ wizard, 4 repeat average. 1/3 Octave smoothing.

We'll start with the comparing his and my mics in horizontal and vertical. The Top trace is comparing his and mine horizontal. The Bottom traces are the vertical ones.

In the second graphic, the top traces are my mic horizontal versus vertical and the bottom ones are his mic horizontal versus vertical. Horizontal is purple, green is vertical for mine.

I'll post more later, but you can see the mics agree very well across mics and horizontal versus vertical up to about 1kHz. Above that all bets are off, although some patterns emerge. More info to follow.

**Anthony** · **03-24-08, 07:48 PM**

So it looks like the vertical response is smoother than the horizontal (not surprising), and there is significant variation between the two mics in the upper octaves.

So the first conclusion I would draw is that for bass response only, these mics are probably okay. The two tested only started varying at 800 Hz, so not bad for subwoofer testing.

Another thing that struck me is the top octave roll-off that is in every measurement. I don't see this in any of the correction files for this mic, so I'm wondering if this is more of a soundcard issue.

It shows up even when using the left channel as calibration and when I use the amp output as calibration (through a voltage divider), so the only explanation was that the phantom power circuitry and preamp in the MobilePre might be adding that. Boomie and I have plans to text using his mixer to see if that still shows up.

But it's clear that this mic needs independent calibration for tweeter testing. There's just too much variance and uncertainty.

So first I'm going to confirm that my sound card is not introducing too much error in the FR (if it is I have bigger problems

), after that I'll have to see what my options are.

A $120 calibration fee seems reasonable, but lately I've been hearing about better mics that are cheaper and slightly more expensive mics that are ruler flat. So I'm not sure what I'll do, but it's always nice to have more information.

**Anthony** · **03-24-08, 07:54 PM**

Here's the same horizontal versus vertical with the amp calibration in the loop.

Not much change, although there is a slight correction for the amp that seems to roll off about 1dB at 20k (but that does not explain the rest of the rolloff completely).

Horizontal is Purple and vertical is Orange.

For those of you who are so inclined, I like the voltage divider setup. It allows the calibration to occur after the amplifier so your reference measurement includes amp response. This way, you've eliminated all but the mic, speaker, and room (sadly those are the biggest contributors, though). Do NOT use the amp output as calibration unless you know what you're doing, though. You can easily fry your soundcard. It took me a lot of research to get the input impedance of the card, and I measured the output of the amp at different setting and sized the jig/divider accordingly.

**brucek** · **03-25-08, 08:55 AM**

Quote:

So it looks like the vertical response is smoother than the horizontal (not surprising), and there is significant variation between the two mics in the upper octaves.

So the first conclusion I would draw is that for bass response only, these mics are probably okay. The two tested only started varying at 800 Hz, so not bad for subwoofer testing.

hehe, we're gonna have to agree to disagree on this issue.

Actually, I feel the mics show very good consistency between units (as we've always claimed).

It's very difficult to draw too many conclusions when comparing microphones this way. Small microphone element positional differences can account for very large changes in the measurements, particularily as the wavelength gets shorter and in areas where the response is changing rapidly. You could obtain the same differences using exactly the same mic under those test conditions. Better conditions would be to measure a near flat response in a very damped room ensuring the mic elements were perfectly at the same spot. Hard to do.

Even under the conditions used here, the different mics and different orientations showed remarkable closeness (although I'm not too fussed about the way you changed scales on every graph). Worst case under the worst condtions at high frequencies, the two mics still only show up to 4dB difference. You could get this by moving the same mic a fraction of an inch.....

My conclusion would be that these are quite good mics for the price and full range measures can be taken with some confidence in the home environment.

brucek

**Anthony** · **03-25-08, 09:42 AM**

Hey Bruce, I think the only real disagreement comes in the top octave.

First off, the scale changed because I was constantly moving curves around in order to separate them and group them. In fact I couldn't even find the original scale on some of the measurements.

Overall, you are right, these results are close for a majority of the measurements.

In my case, I'm most interested in the tweeter section, though, and that's where things start to become different. when you are trying to set tweeter levels or design a notch filter, the response at that end is very important. 4dB is a lot in crossover design! Most of the effects, though can be explained. For instance, the waviness in the horizontal measurements coincide with a baffle effects for a very small baffle (the diameter of the mic). Keeping the mic vertical avoids this.

As to methodology, I can say with certainty that the center of the mic capsule was within a mic capsule diameter for the measurements. I was very careful to measure distance off the floor and use an index mark to set mic location (verified via crude plumb bob). Levels were never adjusted past the initial setup.

So I'm going to recommend people use this mic vertical whenever possible for full range stuff. It seems to be how it was designed (based on other research, not just the results posted here)

The good news is that for bass, the mic-to-mic variation is very small and seem to agree with the calibration files (not pictured here). Good news indeed.

So to summarize:
For REW -- great mic, cheap, available cal files good enough, use whatever orientation you want

For two-way crossover design -- good mic, available cal files good enough up to 3kHz or so. Orientation still probably does not matter (within +/- 1dB).

For three-way, or full range work -- okay mic. Good value, but the response gets peaky. Definitely get calibration done if you want accurate results. Pay attention to orientation of your cal file and beware the "baffle" effects in the 4k to 8kHz region. Use vertical where feasible (the peak is not as great and the response is a tad smoother).

Overall a good value, unless you want to bother building your own. Probably still the best choice for REW/BFD work.

**BoomieMCT** · **03-25-08, 12:40 PM**

Quote:

brucek wrote:

hehe, we're gonna have to agree to disagree on this issue.

. . .

You could obtain the same differences using exactly the same mic under those test conditions. Better conditions would be to measure a near flat response in a very damped room ensuring the mic elements were perfectly at the same spot. Hard to do.

I don't think Anthony is trying to say that what the plots show are the anechoic responses of the mikes. The things he's trying to point out is a) the difference in the ECM-8000's response in the vertical vs. horizontal mounting and b) differences between different units.

Given the graphs shown I'd say there is compelling evidence to show that there is a consistant difference between both mounting orientations and a consistant and repeatable difference between the two mikes in the top octave.

I don't think this was a slam on the ECM-8000 but rather some thoughts on better understanding how to use the mike as well as its limitations.

**brucek** · **03-25-08, 01:21 PM**

Quote:

I don't think Anthony is trying to say that what the plots show are the anechoic responses of the mikes. The things he's trying to point out is a) the difference in the ECM-8000's response in the vertical vs. horizontal mounting and b) differences between different units.

Yes, of course I realize that.

I just don't happen to agree with the conclusions (and gave my supporting arguments why I feel that way).

brucek

**<^..^>Smokey Joe** · **03-26-08, 07:10 PM**

I have to agree with Bruce, dispite your best efforts to align the mics to the same spot a very small difference could result in different reports. Try placing the same mic in and out several times with repeated runs(atleast 4 times) then take the average. Do the same for each mic.
This may help average out positional differences.

I used to work in the sound recording business a few years ago, so spent alot of time with mics. Moving a mic a fraction could make all the difference between good and bad no matter how good or bad a mic is.

Just imagion the room with lots of square boxes(interference patterns), each frequency representing a size, those that fit perfectly to the room shape will do different things to the ones that don't. If you are placing that mic right on the edge of those interference patterns you will get varying results that can be quite dramatic. ie a frequency can cancel completely with a pure tone creating areas with no sound and tone in zones that are shaped like a box.

Scuse the pun but you could be just looking at noise.

**Anthony** · **03-26-08, 09:23 PM**

I was one step ahead of you. I just did tests where I moved the mic a random bit. I also measured much closer to the speaker to eliminate as much of the room effects as possible.

The mic tip was around 16" from the speaker. Maximum move was about 3/4" in or out, side to side.

You can see some things that jump out as positional (green trace in both graphics), but some general things (humps, dips, rolloffs) are evident.

First one is horizontal, second is vertical, last is horizontal versus vertical (I picked one of each that was in the middle).

So, I stand by my assertion that there is a difference between horizontal versus vertical (it may just be the reflection/baffle effect of the mounting body of the mic). It might not mean anything to an RTA measurement or general wide-band EQ -- but it is significant enough for tweeter crossover/notch filter design (which is why I went down this road).

However, this mic is pretty flat up until the "unpleasantness" in the top two octaves. Good news if someone is just getting it to EQ a sub or see the effects of some room panels.

But for any serious tweeter work, calibration is definitely necessary.

**Anthony** · **03-26-08, 09:48 PM**

I also made a cal file using some of these measurements. It might not calibrate to flat -- but it will calibrate to my existing Magnepans (not exactly a bad speaker to design to

)

That'll work for now. I'll probably send the mic off to Caldwell next month. I'm interested to see if my "fake" calibration file is anywhere close

Thanks for the input. Disagreements or no, I just want to understand and get the bottom of this -- no matter how complicated. I guess that's part of the fun

**BoomieMCT** · **03-26-08, 09:49 PM**

Yeah, position errors aside, the delta between my mike and Anthony's stayed the same every time we did different runs. Also the difference between horizontal and vertical was very repeatable.

**Anthony** · **03-27-08, 07:41 AM**

I have some followup data (that was actually in front of me all along, I just did not have perspective for it)

Quote:

Omnidirectional microphones are capable of very flat response over the entire audio spectrum because only the front of the diaphragm is exposed to the incident wave, eliminating the phase cancellations experienced with directional microphones. However, there is one caveat. Omnidirectional microphone response to a plane wave becomes increasingly directional as the diameter of the diaphragm becomes comparable to the wavelength of the frequency being measured. Low frequencies "flow" past the diaphragm without incident. High frequencies, however, cannot bend around the microphone structure. They pile up on the diaphragm surface causing a rise in pressure that is higher than the pressure in the surrounding field (This is one manifestation of acoustic diffraction)

-- "Testing Loudspeakers", Joe D'Appolito, page 54

he goes on to say how "free field" microphones try to counter this by controlling the resonance and damping. I guess better mics do a better job at completely eliminating this phenomenon. Entry level mics like the ECM8000 do not. Which is a shame, because older cal files that I've seen for the 8k show only a small bump in the response. The newer ones (mine included) show a big peak and rapid falloff -- apparently some are worse than others.

It probably has something to do with the capsule change (as Panasonic no longer makes what was used in the older ECM8000's).

Hmm, maybe a DIY microphone is in my future

**brucek** · **03-27-08, 09:54 AM**

Quote:

High frequencies, however, cannot bend around the microphone structure

Exactly, and this was my original disagreement with comparing these mics. At short wavelengths, they are extremely sensitive. I like your idea of the DIY. I've come across lots of exposed capsule DIY mic tests with the old Panasonic unit - and others. That would make a great tweeter test mic. I've seen some really inexpensive (couple bucks) capsules that are sold with calibration files that would work great.

brucek

**BoomieMCT** · **03-27-08, 10:03 AM**

Quote:

brucek wrote:

Exactly, and this was my original disagreement with comparing these mics. At short wavelengths, they are extremely sensitive.

Two reasons I don't agree. First off we are mainly noticing differences in the last two octaves (5kHz - 20kHz). This represents wavelengths of 2.7 - 0.7 inches. I'd say that assuming you have to get the mikes lined up to within a half wavelength that is still well within our test stand / plumb line capabilites. Anthony shows in post #9 that the position errors were minimal.

The other point is that our results were repeatable. We could switch back and forth between Anthony's and my mike and the difference in the readings stayed the same. Similarily the differences with the same mike between vertical and horizontal were very repeatable.

**Anthony** · **03-27-08, 10:07 AM**

Bruce, I came across this schematic for the ECM8000 online. The words needlessly complicated come to mind. It amazes me how those capsule mics have such simple schematics and the ECM uses this one.

I wish I knew more about op-amp circuits (I'm good with filters and other LCR stuff, but transistors and diodes are out of my league) -- so I could tell what's going on there.

It may be as simple as Behringer designed a circuit for the old capsule, didn't change it for the new one, now the new ones are "compensated" for a different response -- therefore the peak that was tamed in earlier versions is now a problem.

I'll look into it, it seems like a fun side project. Also, a guy over at DIYAudio e-mailed me about an article testing a bunch of different mics. Apparently there are some offerings from Naiant and Beyerdynamic that blow away the ECM8000 (one cheaper, one more expensive -- although both are much cheaper than the EarthWorks models).

This weekend I'm going to rerun some frequency tests of my center channel baffle using these corrections. If all goes well, I may just use my "quasi correction" and be done with it.

**Anthony** · **03-27-08, 10:18 AM**

Boom,
We seem to all be nit-picking very small differences and disagree on what we consider important. Like any religious argument, the smaller the difference, the bigger the controversy

Anyways, I think we've shown that for the vast majority of uses, these mics behave rather wall unit to unit and are reasonably flat up to 5kHz, regardless of orientation.

It is comforting that using the same unit in different measurement batches shows almost no variation. so you can believe what you measure on Tuesday is as accurate as what you measured on Monday.

As for orientation differences, I still claim that there is a difference in both bandwidth and amplitude of the hump and the measurements show it. Now, how important this is debatable. For room equalization, bass response, heck, even 2-way crossover design, it would make absolutely no difference. 4dB in the top two octaves is only marginally outside the overall system accuracy (+/- 1dB).

But for more sensitive tweeter work, it's definitely important (as my trial and error notch filter attempts have shown me) to have as accurate a measurement as you can and 4dB or a couple kHz of bandwidth are very important and make the difference between a bright speaker and a dead one.

So everyone who bought an ECM8000 can relax, unless you have to put a notch filter at 13khz, your mic is pretty good.

**Rich Jura** · **04-02-08, 12:01 PM**

I have 9 of the ECM 8000 and access to An Earthworks TC 30K. Next week I'll post a comparison
graph .If I recall correctly they are quite similar.I'm not that concerned about 10K to 20K as
motorcycles and gunfire in my youth took care of that.

Later
Rich

**glaufman** · **04-03-08, 08:39 AM**

Quote:

Anthony wrote:

I wish I knew more about op-amp circuits (I'm good with filters and other LCR stuff, but transistors and diodes are out of my league) -- so I could tell what's going on there.

I may be able to help here, depending on exactly how in depth you want to go, and how far you can take it yourself... but this schematic is not completely correct...I have to make a few assumptions or answer a few questions....

this circuit is in the mic itself? not a separate mic amp? It gets "phantom power" from an external device? that voltage comes in on BOTH XLR 2 and 3, and is referenced (or returned) to XLR1?

**Anthony** · **04-03-08, 08:52 AM**

I think so. I got this off of a website that shows how to mod the ECM8000, so this is the schematic of what's in the mic (not a preamp).

Phantom voltage is applied at 48V to pins 2 and 3 relative to pin 1 (chassis/earth ground).

Someone over on DIYaudio looked at this and had some comments on the design. It serves two functions: to break the signal from the capsule into two components to send balanced over the cable, and to make sure the bias voltage stays consistent on the capsule.

I was confused because similar design mics use a bias voltage across the condenser element, but just put out an unbalanced signal back to the preamp. That is obviously a much simpler circuit.

**brucek** · **04-03-08, 09:00 AM**

Quote:

this circuit is in the mic itself? not a separate mic amp?

Yep....

Quote:

It gets "phantom power" from an external device?

Yep...

Quote:

that voltage comes in on BOTH XLR 2 and 3, and is referenced (or returned) to XLR1?

Yep...

brucek

**glaufman** · **04-03-08, 10:40 AM**

OK... so how much more info are you looking for? Would I be telling you things you already know if I said the SA transistors are forming a differential amplifier that provides the balanced output, and some gain, the BC transistor is an input stage for that amp, to provide impedance matching between the mic element and the diff stage (matched impedances make for the most efficient energy transfer) and also to provide some gain...
The resistors all set the bias for the transistors, but at the same time define the input/output impedances and gain of their particular stages? The 1uF caps allow the signal to pass from stage to stage while keeping the biases separate...

Advertisement