Are comparisons odious?

[Chrisbee]

Hi

The purchase of an ECM8000/Xenon 802 today instantly provided a chance to confirm the accuracy of the various testing devices in my possession.

Using the correct REW cal files my Galaxy 140 so closely matches my ECM8000/preamp combo as to make the latter merely desirable luxuries.

You chaps obviously did a superb job on the cal files and these devices do seem very similar from example to example.

The problems start when I try to compare my old/new RS analogue SPL meters with the above two.

There is obviously some disparity here which is quite probably the result of a simple testing error on my part: I had similar problems in matching the RS curves when I first obtained the Galaxy 140. I have tried both C-weighting and disabled. The ECM8000 and Galaxy 140 both had C-weighting disabled for their tests.

Rather than further damaging my few remaining brain cells it is very likely that you can point straight to the problem.

My old and new RS meters have near identical curves despite there being 10 years in age between them. They do not seem to share the same curves with the other two more expensive devices despite careful loading of the correct cal files for each REW test. Close matching of the position of the various mic capsules [on the tripod at the listening position] was achieved with rubber bands.



Thanks.

[Chrisbee]

Retested with all three SPL meters. One old RS , one new RS and Galaxy 140.

I didn't bother with the ECM8000 today because the Galaxy is so close.

C-ticked and not ticked for each meter in turn.

Correct cal.files loaded for each meter after clearing cal.file each time.

I downloaded the latest cal files from the download area.



The two RS meters match well despite the 10-12 year age gap. (Upper green curves)

The Galaxy 140 and ECM8000 match within a hairsbreadth.

Neither pair agrees with the other.

I started out hoping this test would be a useful comparison to save people wasting money upgrading their test equipment unnecessarily.

[brucek]

I tried to come up with a reason why the Radio Shack meters don't compare well to the better Galaxy and ECM, but I don't have any idea.

The C-Weight must be enabled on the meters themselves. The ECM is a flat mic so it has no C-Weight switch.

The C-Weight in REW is only a factor when measuring outside the range of the REW calibration file.

Most meters and microphones are quite accurate at 80Hz and require little or no compensation, In fact you'll find all the cal files are basically zero at that point. Your comparisons should be normalized to 80Hz.
My RS meter is fairly close to my ECM and so are Sonnies, so I have no answer for you. Wish I did. We have had discussions about RS changing capsule suppliers at various times, so I suppose it makes the RS meters a **** shoot.

brucek

[DougMac]

Your dilema reminds me of an old Chinese saying:

"A man with one watch always knows what time it is. A man with two watches never knows what time it is."

Doug

[Chrisbee]

brucek

Thanks.

I feel this has to be related to the cal files.

Nothing else changes when the next meter is connected.



DougMac

Abandon hope all ye who enter here! (I have a lifetime's interest in clocks)

[thewire]

My theory is that the differnet mic except the RS meters are running on a different voltage and this is why they read different. Here is an example of with a digital RS meter mic plugged in to my sound card and (green) without. Turning off the mic has a similar result.

mic and no mic.jpg

[Chrisbee]

thewire wrote:

My theory is that the differnet mic except the RS meters are running on a different voltage and this is why they read different. Here is an example of with a digital RS meter mic plugged in to my sound card and (green) without. Turning off the mic has a similar result.

I am swapping the same 9 Volt battery between the meters to exclude random variations.

Here's an new REW graph for your entertainment. RS meter V Galaxy 140.

I hope the annotation is easily understandable.

CW and AW refers to meter switch weight settings.

Cal/No Cal refers to REW meter cal file loaded under Mic/Meter. [33-2050.cal]

[3ll3d00d]

cross-post from avforums for completeness

it looks to me like that meter needs the RADIOSHACK-33-2055_4050.cal file. This would take off 6dB @ 20Hz, ~13.5dB @ 10Hz and 10dB @ 7Hz which I reckon would make it map reasonably closely to the galaxy curve.

if so then this goes back to the "how do you know which cal file to use?" issue

[Chrisbee]

3ll3d00d wrote:

cross-post from avforums for completeness

it looks to me like that meter needs the RADIOSHACK-33-2055_4050.cal file. This would take off 6dB @ 20Hz, ~13.5dB @ 10Hz and 10dB @ 7Hz which I reckon would make it map reasonably closely to the galaxy curve.

if so then this goes back to the "how do you know which cal file to use?" issue

Well done!

I was using the 33-2050 cal file for my new RS meter.

It really needed the old meter cal file. 33-2055-4050

Another senior moment!

I'll post a properly annotated graph later.

ECM8000/Galaxy 140 and New-Old RS SPL meters.

For the moment here's just the Galaxy 140 and RS meters with the correct cal. files and 1M sweep.

[Dent]

Chrisbee,

Would you be able to do a sweep from 5 to 20000 Hz with the ECM8000 and the Galaxy CM-140? I'm just curious to see how they compare when you do it. I believe when you hook up the Galaxy to do the sweeps that the C-weighting option in REW should definitely be checked. Someone else can correct me if I am wrong.

[Chrisbee]

Okay.

First graph ECM8000 at listening position head height C-weighting ticked and unticked.



Second graph Galaxy 140 C-weighted ticked and C-weighting unticked.



Third graph ECM8000 and Galaxy 140 together. With and without C-weighting.



All graphs 1/2 octave smoothing.

I need to work on my frequency response!

[Dent]

Thanks a lot Chrisbee! This is exactly what I wanted to see. I guess there are variations when measurements are compared with those done on different days. In your graphs at the very beginning of this thread with the Galaxy and ECM8000 they were pretty much spot on from 5 to 200 Hz (which were the limits of your graph at that time). With these latest graphs, the Galaxy and ECM8000 only seem to match from about 19 Hz to 225 Hz except for that "dippy" area at about 175 Hz where it doesn't match. Above 225 Hz the two meters compared to each other are all over the place.

I assume with the middle Galaxy graph that the response that is higher from the 2500 Hz area and up (greenish-blue line) is the C-weighting checked line and the lower (more bluish color line) one is the C-weighting unchecked line.

From "maybe" about 5000 Hz or so to 20000 Hz the Galaxy follows a general trend of the ECM8000 but it is by no means matched. I wonder how Ilkka gets his calibrated mic and Voltcraft/Galaxy meter/mic to match so closely:

Ilkka wrote:

Sorry that it took so long, but here are the measurements. It looks pretty good.

Again, Chrisbee, thanks a lot for doing this.

[brucek]

C-weighting unchecked line

The ECM8000 is a flat mic. It has a calibration file that covers the entire audio band. The C-Weight checkbox has no effect.

The Galaxy is a C-Weight meter. The calibration file only extends to 200Hz, therefore the C-Weight box must always be checked.

brucek

[Dent]

brucek wrote:

The ECM8000 is a flat mic. It has a calibration file that covers the entire audio band. The C-Weight checkbox has no effect.

The Galaxy is a C-Weight meter. The calibration file only extends to 200Hz, therefore the C-Weight box must always be checked.

brucek

That is what I figured was correct also.

Was my assumption correct regarding the middle Galaxy graph, i.e. that the response line that is higher (greenish-blue line) from 2500 Hz and up is the measurement with the C-Weighted check box checked in REW?

[brucek]

I assume so, but really, with respect to the measurer, doing graphs with a C-Weight meter with the box unchecked just adds confusion. In fact, comparisons are quite difficult to do between two mics and fraught with all sorts of chances of errors with regard to element positional differences and quckly changing responses.......

The ECM and Galaxy are fairly reliable for home use with the cal files we provide for full range. The Radio Shack meters are not. That's about it.....

brucek

[Chrisbee]

It was all a bit of rush so I rubber banded the two together to ensure coincidence of the microphone capsules. They were then mounted on the tripod together. There were almost insignificant variations of level which I adjusted out in REW. I'm still rather surprised at the variations at different frequencies. It comes down to: Which do you trust?

Thanks brucek. I should have posted the final graph of the ECM8000 sans C-weighting against the Galaxy 140 with C-weighting for a simpler comparison.

[SRR]

Try a better preamp then the behringer stuff, even a m-audio DMP3 will give better results. I don't think the 802 mixer has an EQ bypass and I would never trust it for measurements.

[Chrisbee]

SRR wrote:

Try a better preamp then the behringer stuff, even a m-audio DMP3 will give better results. I don't think the 802 mixer has an EQ bypass and I would never trust it for measurements.

One of our members has checked the Xenyx802 FR in loop with REW: It doesn't look too bad at all provided we all avoid the area below 4Hz. The M-Audio model is 4 times the price and will only ever be used for REW testing. Moving the microphone 4" in my room would make a far bigger difference in FR than changing the preamp.

http://www.hometheatershack.com/foru...x802-loop.html

[brucek]

Try a better preamp then the behringer stuff, even a m-audio DMP3 will give better results. I don't think the 802 mixer has an EQ bypass and I would never trust it for measurements.

I'll have to disagree. The 802 is actually quite a good preamp, and bypass isn't necessary since you can calibrate out any response anomolies to zero by including the preamp in the REW soundcard calibration file. It's quite simple to do. This then removes its response from the measurements and renders it flat.

But, if didn't want to include the preamp in the cal file it has a completely acceptable response.

Here's a graph of the 802 with the EQ in detent, using standard REW vertical graph axis, from 5Hz to 20KHz. The response is completely adequate.

EURO ub802loop_response.jpg

brucek

[SRR]

Just wondering but are you taking into account the impedance of the microphone plus phantom power into the mic VS the input impedance of the board, to come to the conclusions that the lowest priced mixer in existence is good enough? Please ask your local pro shop what's the return rate on them happens to be. V.S. a very respected for the price DPM3 happens to be?; QC is not in Behringers natural language, I would never trust a behringer mic pre in my studio. Volume control for my monitors was a stretch, other then that a tracking headphone amp, patchbay, and now a 1124P for my subs....no further I say, but watch I turn around and buy a XXXXXX Behringer cause I just can't afford proper stuff. I am not trying to be confrontational just if you are going to do it, I guess do it right or go home?

[brucek]

if you are going to do it, I guess do it right or go home?

I guess that's correct since most people on this forum do use this mixer only for the home. After a lifetime in engineering, I've come to know that it's all about compromise. The 802 is a perfect example.

The 802 mixer is being used to measure subwoofer response for home use, where the slightest microphone element positional difference can result in a response change of several dB. Do you really think I care about the impedance spread between the mic and line-in inputs? This mixer is completely adequate for home use and anything more would be somewhat wasteful. The money would be better spent elsewhere.

Please ask your local pro shop what's the return rate on them happens to be

Why would I care? Mine has been working for years. We've not had anyone here complaining about the 802's reliability. If it breaks, take it back and they'll give you another one.

brucek