How low can we go?

[PeteD]

OK. I am not looking for tweaking opinions on my freq response, yet. That will be in another post soon - after some more testing/tweaking. Please excuse the non-standard graphs - for illustrative purposes only...

Anyway, I am am curious about way the VLF response (i.e. <10Hz). I have a RS meter and downloaded the date-corrected cal file (shown in the graph below). I have a BFD and have no plans of upgrading to something to allow a fuller range EQ (assuming they exist) anytime soon. My main curiosity is how much of this low output is real?


The cal goes to 7Hz and extrapolates to the low end from there. This is a 4x15" IB, so it can put it out down there - you can feel it, especially at higher SPL:

(yes, I wore hearing protection - I am not a loud volume guy anyway - my hearing is still pretty good)

If the RS meter tails off more than we think (or more than represented by the curve) at low HZ, won't this increase the measured SPL, because the correction values will be larger?

In other words, if you can measure it is probably there, but if you can't measure it, that doesn't mean it isn't there?

Is the meter so erratic that low so it is all garbage?

Thanks,
Pete

[brucek]

Is the meter so erratic that low so it is all garbage?

Any response information that is below the meter calibration file is not meaningful.

Below that, the falling meter calibration waveform (which is incorrect, since you have C-Weight turned on- it's not extrapolation) is exaggerating the signal that's in the noise and making it appear as though the response down low is good. I'm afraid that it's not though. Uncheck the C-Weight box to help alleviate the problem, or at least don't look at your graphs below the calibration file of the meter.

Here's an impulse response showing the noise about -45dB down and then the subsequent frequency response graph with C-Weight on and C-Weight off.

Note that at about -45dB down from peak, the response appears to rise.... it's bogus.

In both cases though you can see where the signal is actually into the noise because the signal looks like it starts rising. It doesn't really........

RISING LEVEL impulse.jpg
C-Weight checked in REW
RISING LEVEL with c.jpg
C-Weight unchecked in REW
RISING LEVEL no c.jpg

brucek

[PeteD]

OK. That makes sense.

So, the response graph should be cutoff at the lowest frequency where there is a cal value.

How were the cal files generated? Is the fact that they stop a function of signal generation (difficulty generating low Hz reference tone) or that the RS meters are not capable of measuring below that level?

How low do we feel comfortable measuring with the Galaxy meters...5Hz?

How about the ECM8000?

Frankly, I don't see a need for output below 5Hz, I would be more worried about transients at those frequencies. I seem to remember that most electronics start to roll off at those frequencies anyway...

Thanks,
Pete

[brucek]

So, the response graph should be cutoff at the lowest frequency where there is a cal value

Yes.

How were the cal files generated?

Comparison testing against a professionally calibrated ECM8000 using several samples of each meter type. The lowest frequency chosen for each meter type was determined experimentally.

How low do we feel comfortable measuring with the Galaxy meters...5Hz?

Personally, I think it's a bit silly to measure response below 10Hz.....

brucek

[PeteD]

brucek wrote:

Comparison testing against a professionally calibrated ECM8000 using several samples of each meter type. The lowest frequency chosen for each meter type was determined experimentally.

Personally, I think it's a bit silly to measure response below 10Hz.....
brucek

I realize they were compared to a reference mic - I meant how were the tones measured generated? Are they at a certain SPL?

Since there are already movie soundtracks with content in this area (War of the Worlds, Black Hawk Down, Hitchhikers Guide) of less than 10Hz, I don't think it is that silly...unless the sub rolls off below 20Hz, then I agree.

Thanks,
Pete

[brucek]

meant how were the tones measured generated? Are they at a certain SPL?

REW (of course) generated the signals and measured at 75dBSPL.

less than 10Hz, I don't think it is that silly

You missed my point. I don't dispute the frequency content, only the use of cheap RS meters, Galaxy meters, ECM8000 mics for any frequency less than 10Hz is nuts......

brucek

[PeteD]

brucek wrote:

REW (of course) generated the signals and measured at 75dBSPL.brucek

I am obviously not asking the question well. For example, is the mic placed in close proximity to a 15" woofer to make the measurements in air? i.e. Is generating the tone in air part of the error involved with measuring this low?

brucek wrote:

You missed my point. I don't dispute the frequency content, only the use of cheap RS meters, Galaxy meters, ECM8000 mics for any frequency less than 10Hz is nuts......
brucek

I see! I was thinking that the Galaxy meters were reproducible even at some of those single digit frequencies. Is the fact that ECM8000 was used and not really accurate that low the problem?

i.e. Can you calibrate a specific lower end microphone with a high end mic (not ECM8000) to measure these low frequencies or do you really need to use a high end mic to measure the frequencies directly to plot response down this low?

Thanks,
Pete

[brucek]

is the mic placed in close proximity to a 15" woofer to make the measurements in air? i.e. Is generating the tone in air part of the error involved with measuring this low?

The mic(s) were placed near field using Sonnies behemoth subwoofer and his ECM8000 microphone that was specially calibrated professionally down to 5Hz. Believe me, his mic is accurate to 5Hz. Certainly, using measurement data to derive corrections is fraught with problems though. Where the response changes rapidly, if you have positional differences in the mic elements, you can get errors. We placed a lot of effort in getting a response that was as smooth as possible to obtain as little error as possible. We make no guarantees. If you need a guarantee, you have to have your mic calibrated.

Once the calibrate files were all created and normalized, we loaded them in REW to test several like models to see how they compared to the ECM standard (which is quite accurate since it was calibrated). They all tracked fairly close and so as far as the ones we tested they were good to use, but there's no guarantee that your meter is the same. But it will be close enough for home use. It's not worth even bothering about a few dB, since a foot of listening position change will be at least that much change....

The Galaxies tracked the ECM fairly good down to about 7Hz and then were erratic below that. I think your fairly good down to 10Hz though..

brucek

[PeteD]

brucek wrote:

If you need a guarantee, you have to have your mic calibrated.
brucek

Does this mic need to be of some sufficient quality to measure say to 5Hz, or can any of the above mics be individually calibrated to read that low?

Sounds like ECM8000 is best choice with side benefit of measuring higher frequencies more accurately, too.

Thanks for the patient replies,
Pete

[brucek]

Sounds like ECM8000 is best choice with side benefit of measuring higher frequencies more accurately, too

Oh for sure. The ECM8000 is a great omni-directional mic at a reasonable price. It is also quite consistent between units (as we found by experimentation of quite a few). This says that you can buy an ECM and be quite confident for home use by using our calibration file on the site. No guarantees though..
It's also quite low in noise too, since it's a balanced output you can run an XLR mic cable a good distance to your preamp/mixer.

brucek