Home Theater Forum and Systems banner

1 - 6 of 6 Posts

·
Registered
Joined
·
40 Posts
Discussion Starter #1
When I'm measuring SPL using speaker pink noise with my Galaxy CM-140 (on dBC, slow) I'm getting approx 5dB higher readings than REW does when using UMIK-1 with calibration data.

Could the calibration on my CM-140 be that much off? Can I be certain that the input levels from the UMIK-1 is correct? I believe I have seen that Windows might change the input levels, and that would affect the meas wouldn't it? I have checked that the volume is set at 100% in this case though.
 

·
Premium Member
Joined
·
1,562 Posts
Is the REW meter also set on "dBC, slow"?

There are several things that can be contributing to the difference. Normally the variations tend to offset/randomize more than stack so this much difference is probably a little unusual.

There may, or may not, be a setting wrong someplace, or the noise source selected may be inappropriate. I don't know the standard signal for SLM level calibration without looking it up, but, it's probably best to use the REW speaker level signal for the noise source. It is band limited I believe. A full range noise source would be a good 2nd choice. It could be very misleading the use a SW (low frequency) noise source.

It's possible there is that much difference even under similar and suitable conditions, but it is unlikely. Neither unit is individually calibrated for level per industry standards for SLM's and the 3 sigma is probably near ±3 dB for each.

The good thing is that the absolute level of calibration has no significant effect on most of the charts we normally use for tuning a speaker system. It can be a factor for some purposes, so if it is important for you, then neither unit is suitable.

This is intended as general directional advice to help your understanding of the possible issues. I am not qualified to offer more expert information.
 

·
Registered
Joined
·
40 Posts
Discussion Starter #3
I'm using the speaker cal pink noise in REW as source, so yes, that's a band limited source. So the differences I measure are both with dBC and slow. I checked again today and when I measure ~76dBC on the CM-140, I measure 71-72dBC with REW itself. Mics are only a two inches apart, both on indivdual stands.

Pink noise is random in nature, so unless the dBC analysis filter alogrithm is the same on both devices, I'd be surprised if the two measurements would give the same result. Hence I repeated the measurement, but this time, using something which is fixed: 1kHz sine wave. Essentially all mics and meas equipment are calibrated against 1kHz, so I'd expect more precision here. Here the results are steady 71.6dBC with CM-140, and 68.0dBC with UMIK-1. So there is a 3.4dB difference. Question is which one is correct.

Sorry, but I believe the UMIK-1 is indeed individually calibrated. I have downloaded the calibration file for that mic and are using this with REW. Unless I've mistaken, the tolerances then must be less than +-3dB.
 

·
Premium Member
Joined
·
1,562 Posts
You may well be correct about the UMIK-1. I know that it is calibrated for SPL response, but was unsure about whether the sensitivity is also individually calibrated. If the various cal files on their site have different values for the sensitivity then they are indeed providing individual level calibrations also. I have never checked.

The question is then how are the calibrations done at the 2 suppliers. As you demonstrated, we get different values with different methods. I would be surprised if either follow recommended practices. It would be even more surprising if either supplier has a an ISO certified lab with NIST traceable calibrations. It doesn't make sense to do all this at this price level.

I assumed your question as to which is correct was rhetorical. How would anyone know if either is correct?
It also depends on your requirements. At this price point I suggested both could be considered correct.
If you are expecting ±0.5 dB accuracy with high reliability. You need to find a well respected certified lab to test them for you. They can probably certify sensitivity to that level.

You could buy or rent an SLM calibrator, but it may be difficult to get an appropriated fixture (adaptor) for the nonstandard size of the UMIK-1. That way you could adjust them to be very nearly the same when measured in that way. That would be of questionable value however unless you also somehow also get the freq response accurate.

If you want them to agree for your own comfort level there is nothing wrong with just picking one that you decide to use as the reference and then adjust the other one to agree with it based you your own cal process. For one of my measurement mics I have data from 2 different cal locations. I just split the difference and use that value. I have no idea if it is correct however.

There is no easy answer. At least not one that is "correct"! :)
 

·
Super Moderator
Joined
·
9,260 Posts

Personally I would put more faith in the Galaxy, simply because it is a sound level meter. That is its sole intent and function. REW’s SPL readings with the UMIK are based on the mic’s sensitivity calibration, which (as jt noted) may or may not be accurate.

But as jt also noted in Post #2, it’s only relevant if you require absolute dB calibration for your REW graphs.

Regards,
Wayne

 

·
Registered
Joined
·
40 Posts
Discussion Starter #6
Thanks both of you

I'm sorry, but I did not intend to be rhetorical. From the little I've seen of mic sensitivity and SPL calibration, I have always seen them to use a 1 kHz reference signal to calibrate the sensitivity. I would assume that the CM-140 is individually level is calibrated to a reference in the final step in (Taiwan) production within some tolerance. Since the product has an adjustment pot available from the outside, and given that such pots are fixed in one end when coming out of production they need some adjustment at some point. But granted, I do assume a few things here.

I believe the sensitivity of the UMIK-1 is individually calibrated. The sensitivity figure in the calibration file is unique for each serial number. And from what I've seen the calibration adjustment is always 0dB for 1kHz.

Point is, I'm not trying to achieve absolute references here. I know that these products are not high-end products with precise calibrations. But I am surprised that they show such great differences. I noticed this difference when I calibrated the SPL level in HT to 75dBC (using speaker cal pink noise) with the CM-140. Then I did a UMIK-1 scan, and the resulting frequency response is centered around 70dB or so. But the absolute level of the curves are really not that interesting.

Secondly, I'm using the CM-140 to verify the SPL when working front-of-house during live concerts. It does not need to be exact, but it would be great to know if CM-140 is off as much as 6dB. I'd expect a few dBs, but not 6dB, that's all.
 
1 - 6 of 6 Posts
Top