ECM8000 Orientation for Speaker Measurements

[arc2v]

I have noticed many pictures online of people using the ECM8000 to measure speakers and aiming the mic right at the speaker in question.

However, I have discovered that many of the calibration files floating around out there are for the mic oriented vertical.

so my question is: what's the proper way?

how do you all do it for measurements?

The reason I ask is that I'm seeing weird results in my top and bottom octaves that seem to correlate with the calibration file. So what I hear flat seems to measure way bright.

I'm considering paying for the calibration, but at $100+ I'd like to research my options more first.

Any thoughts on this?

Thanks,
A

[ShinOBIWAN]

Good question.

Personally speaking I point the mic at the driver/loudspeaker when aiming to measure loudspeaker performance. But point the mic vertically, at the listening position, when doing room correction or looking into performance at the listening position.

I have no solid reasons for working this way, it just feels sensible I guess.

[soongsc]

I'm wondering whether it has anything to do with the protection cap. Normally I don't like anything between the driving diaphragm and the receiving diaphragm. It creates some funny artifacts when measuring with MLS signals.

[arc2v]

that may be correct, but I see the same effect when using swept-sine (Room EQ Wizard), so it is definitely a frequency response thing -- regardless of signal type.

I will have an easy night tonight, so I'll measure the room and a few drivers vertical and horizontal -- with correction file and without, and see what the results look like.

[Ron E]

Originally posted by arc2v
I have discovered that many of the calibration files floating around out there are for the mic oriented vertical.

so my question is: what's the proper way?

The reason I ask is that I'm seeing weird results in my top and bottom octaves that seem to correlate with the calibration file. So what I hear flat seems to measure way bright.

Which calibration files are you speaking of? Can you show us some graphs?

You would want to point it at the speaker and calibrate the mic for on-axis measurements if that is what you are measuring. The vertical orientation is more used for a random incidence type measurement that you might use to measure reverberation in a room, etc.

I think you can still get your mic calibrated for ~$40 from Kim Girardin.

[Ouroboros]

You will always get faster HF roll-of if you orient the mic at 90 degrees to the source, compared with pointing straight at it. B&K have information on their site about this. It's much worse with large diaphragm mics.

When the wavelength of the signal gets small enough to be significant compared with the diaphragm diameter you'll get some signal cancellation on the mic output if the acoustic waveform is passing across the mic diaphragm.

I don't know if the Behringer calibration data is with the mic at 0 or 90 degrees to the signal source. I guess it's with it at 0 degrees.

[arc2v]

Actually, what I've seen coincides with that statement.

The behringer plot shows a fairly flat response that humps a bit (0.5 db) in the upper midrange and then just a very slight rolloff above 15 kHz.

The other cal file, from HomeTheaterShack was calibrated vertical and it shows a flat response across with a steep rolloff at 10 kHz. I confirmed that they tested theirs vertical.

so it may be that the Behringer is pretty flat for direct sound and needs that compensation for vertical use. I'll see if my measurements agree with that statement tonight.

[ShinOBIWAN]

Some test data on Behringer ECM8000 of various vintage:

http://www.hifi-selbstbau.de/text.php?id=70&s=read
http://www.hifi-selbstbau.de/text.php?id=121&s=read

The standard generic calibration file is far from ideal - not worth bothering with IMO.

The ECM8000 is accurate enough for fairly demanding tasks but if you need a flat mic then you should look to an individually calibrated one.

[Svante]

Here are measurements in three directions on an ECM8000 (relative to an Earthworks M30). The curves need smoothing, I know, but you can get the idea how it would look just by looking.

[ShinOBIWAN]

In that test setup can the earthworks M30 be considered flat?