[zmix]

I find that I use REW to test electronics more frequently than I do to measure the acoustics of my studio.

It's the most incredible tool I've found for evaluating distortion vs frequency, phase anomalies, harmonic behavior, etc.

One problem I have with REW in this capacity is that the measurement data is scaled for SPL rather than relative differences.

For instance, if I want to measure the frequency response of a piece of hardware and have a meaningful and simple display of the deviations form "zero" , I'll need to shift each reading down by 85dB or so, and because the scale of REW only goes to -60dB my THD measurements typically scroll off the bottom, never to be seen again.

I wish there was a calibration or a preference that would allow "85dB SPL = 0VU" so that I could see the levels relative to a calibrated unity setting.

Thank you.

 

[JohnM]

The lower limit of the graphs has been reduced to -180 dB for V5.10, so it is now possible to calibrate to a 0 dB reference.

 

[zmix]

Thanks John..!!

Is it possible to start at a reference level of 0? I find is a bit tedious to have to subtract ~85db from each sweep after the fact when running multiple passes, as I frequently do when adjusting a circuit parameter.

 

[JohnM]

You can calibrate the SPL meter to read whatever you like at your chosen reference input level, so just go through the SPL meter calibration.

 

[zmix]

You can calibrate the SPL meter to read whatever you like at your chosen reference input level, so just go through the SPL meter calibration.

That sounds promising...

However, I do this every time, but never seen anything other than "Adjust the SPL figure below until it matches the reading on your SPL meter then hit "Finished" " When feeding a normal signal level the reading is ~85dB SPL.

How can I calibrate in Vu, dBu, etc?

If I set the reading to 0dB I get a warning that the signal level is very low (-87dBfs)...

 

[JohnM]

You lie

All REW knows is the rms dB FS level of the signal on its input. When you enter the figure in the SPL meter you are telling REW how to translate that dB FS reading into the display figure you want to see, it is used to calculate a dB offset that is applied to the input dB FS values. For example, if you fed 775 mV rms into the input and enter 0 as the SPL figure the SPL readings would then be in dBu.

 

[zmix]

You lie

All REW knows is the rms dB FS level of the signal on its input. When you enter the figure in the SPL meter you are telling REW how to translate that dB FS reading into the display figure you want to see, it is used to calculate a dB offset that is applied to the input dB FS values. For example, if you fed 775 mV rms into the input and enter 0 as the SPL figure the SPL readings would then be in dBu.

Perhaps I was doing it wrong... !

I realized that you can arbitrarily set the reference based on how much headroom you need in your converter.

I've managed to calibrate it several times now to several references..

Fantastic software, REW is...!

 

[JohnM]

Apply a signal to the input you use for measurement - that could be a feed from an external signal generator or a loopback connection from the soundcard output. Open the SPL meter and click Calibrate. If you are not using the REW generator for the input signal select "Use an external signal" as the signal source. In the SPL Reading Calibration box enter the level that corresponds to the input you are applying - if you have set things up to apply a 775 mV rms sine wave to the input, for example, and you wanted REW to read in dBu, then enter 0 in the box since 775 mV is 0 dBu. Then click finished.

 

[zmix]

Got it. Thank you..!!!

 

[zmix]

Ok, I've got the reference calibration all sorted out for my interfaces, but I am a bit confused about how the reference level applies to the scaling of the RTA display. The numerical display in the pop-up field matches the level indicated on the generator, i.e.-23dBfs (my 0vu reference) but the RTA and RTA cursor indicate a signal level about 23dB lower than that:

 

[JohnM]

The peak looks unusually broad, have you applied smoothing? If smoothing is applied the peak will be lower and broader.

 

[jtalden]

Maybe it should be Rectangular window instead of Hann?

 

[zmix]

The peak looks unusually broad, have you applied smoothing? If smoothing is applied the peak will be lower and broader.

Not sure what you're referring to?

My question was about how the reference level doesn't match the scale in the RTA display, here's a JPEG of the exact same measurement (1kHz at -23dBfs), though generated internally from REW, rather than a screen capture. Note that the text looks terribly mangled compared to the onscreen display:

 

[JohnM]

In the RTA graph controls it is possible to select smoothing to apply to the spectrum plot. That plot looks like smoothing has been applied, which is why the peak is quite broad and the level of the peak has been reduced, both a result of smoothing - or so it appears, at least.

 

[zmix]

In the RTA graph controls it is possible to select smoothing to apply to the spectrum plot. That plot looks like smoothing has been applied, which is why the peak is quite broad and the level of the peak has been reduced, both a result of smoothing - or so it appears, at least.

Thanks again, John.

Upon further investigation I discovered that I had the RTA scale set to the absolute "dBfs" rather than the relative "dB" which is why the SPL calibration was not being applied..! (The dB/dBfs selector disappears once the "Distortion" mode is engaged).