REW SPL cal craziness

[NEO Dan]

I just spent 2+ hours trying to figure out why I can't get a good SPL reference calibration in REW.

How is it that you can have a RS SPL Meter reading and go through the SPL Cal dialog and match the numbers(to 75dB) and then when you go to make a sweep the level is over 30dB low?

The RCA loop back calibration went well

The only speakers playing are the subs

We attempted another calibration and sweep run at over 95dB with the same ~30dB too low for the sweep.

When calibration noise is playing the SPL on the RS meter and the SPL display in REW closely matched.

Any pointers would most appreciated

Thanks
Dan

[brucek]

Did you run the Calibrate SPL routine after the Input level was set?

brucek

[NEO Dan]

brucek wrote:

Did you run the Calibrate SPL routine after the Input level was set?

brucek

Correct, that was the last thing before attempting a sweep.

[brucek]

~30dB too low for the sweep

One thing you should be aware of is that the Check Levels routine in the Measurement screen and the Check Levels routine in the Settings screen use different values to generate the pink noise level setup test signal.

The Check Levels routine in the the Settings screen uses the fact that you have the Check/Set levels with Subwoofer pull-down selected to generate its test noise (this a band limited signal with a low and high cut of 30Hz - 80Hz).

REW double checks and verifys that the end frequency of the sweep that you have selected in the Measurement panel roughly matches the test signal setting for sanity, and so its Check Level routine uses the end frequency as the test pink noise hi-limit cutoff. So, if you had 20KHz as the end frequency to measure for your sweep, you can see how the pink noise would be different (between the two Check Levels) and as such the low level signal you're experiencing.

Often, even if you have the levels set correctly, a peak in the response of the speaker being tested will have energy at a frequency that trips the sanity check and announces that levels are too high.

Simply return to the Settings screen and adjust the input level of the soundcard a bit lower and then run Calibrate SPL and carry on.

brucek

[Ricci]

I've noticed this exact thing that Dan is refering to. You set all of your levels and then calibrate SPL and match your levels from your mic to the SPL meter reading and everything is fine and the signal is strong and well juiced. Then you go to take a measurement without touching any settings and suddenly the input level is too low. I've had this happen when calibrating at everything from 75db to 100db levels between the mic and SPL meter.

It's not related to the difference between fullrange and bandwidth limited PN. It was with subs only for me too. When you calibrate SPL isn't the signal used always 0dbfs, or is it -3dbfs? I wish you could calibrate SPL using pure sine waves instead of pink noise , which you kind of can but it takes some work.

[brucek]

It's not related to the difference between fullrange and bandwidth limited PN

Actually it is.

The only difference between the Settings screen Check Levels and the Measure screen Check Levels is the bandwidth that the pink noise signal is limited to.

So, if the Check Levels screens pink noise of 30Hz-80Hz returns a certain SPL level, you can see how using the actual sweep setting in the Measure panel would pick up an incorrect setting - for example, trying to measure a sub using a full range sweep would immediately show a low level highlighting something is wrong. If the standard 30Hz-80Hz cal signal was used at this point, it wouldn't pick that up.

That's an extreme example, but the result still holds when you are measuring from 0Hz-200Hz. The pink noise will return a lower SPL level than if you were using the cal signal of 30Hz-80Hz.

My point that I'm trying to make is that if you have correctly set up your levels in the Settings screen, then ignore the low level warning in the Measure screen....

brucek

[NEO Dan]

in my attempt the SW cal noise was selected in the SPL cal routine and then verified to match by re-entering into the SPL cal and verifying that the level still matches. Then canceling out of that and attempting a measurement, it comes back and says the measurement is like 30dB low, while the RS meter registers a significant SPL during the sweep. The cal and sweep routine was attempted at close to 100dB, and still getting a low SPL error message.