Several questions on calibration and level check.

gallantus

New Member
Thread Starter
Joined
May 19, 2024
Posts
9
REW newbie here.

Several questions on calibration and level check.

  1. I read the manual, and it didn't make sense to me. What exactly does the check level do? I get it that measurements are usually made at 75 dB and the manual tells you how to use the check level function, but what does it do? What is the difference between check level and Calibrating the SPL Reading?
  2. When you calibrate the SPL or check level, do you use one speaker or both?
Thanks in advance for your help.
 
Calibrating the SPL reading allows REW to show the input as SPL, it lets REW know the dB offset between the input digital data level and SPL figures.

Check levels makes sure the input signal will be in a suitable range when the measurement is made by sending out a signal at the measurement level and checking the level of the input.

Usually one speaker would be used.
 
Calibrating the SPL reading allows REW to show the input as SPL, it lets REW know the dB offset between the input digital data level and SPL figures.

Check levels makes sure the input signal will be in a suitable range when the measurement is made by sending out a signal at the measurement level and checking the level of the input.

Usually one speaker would be used.
Thanks John! I'm a complete newbie to REW and not so verse in sound measurements in general, so thank you for your patience in advance.

I'm still not completely getting it. It sounds like one of them ensures the output level is accurate and one of them ensures the input level is accurate?


I assume I have to do both before I do any measurements and that I should target 75dBC with my handheld SPL Meter?
 
SPL calibration only needs to be done once as long as you do not change any gains along the input path. Its purpose is to let REW understand how to show levels as correct SPL figures. It is only concerned with the input level. You won't need our handheld meter after you have calibrated REW's SPL reading.

Check levels makes sure that the levels are appropriate to get a good measurement. If your measurement output level were very low or your mic were a long way from your source the input level could be too low. If the output level were too high or the mic too close to the source the input level could be too high. You usually only need to check levels if something has changed significantly since you last measured, such as a big change in mic position or measuring a speaker a very different distance from the mic.
 
Thanks John.

Few more follow up questions.
  1. At what dBFS level should I aim for the input to be? Some sources recommend anywhere from -24dBFS to -12dBFS, some just straight out say use -18dBFS. What does REW recommend?
  2. If I am calibrating the input level to be at, say, 75 dB, using my hand-held dBA meter and the sweep level is set to the default -12dBFS. Then why would I still need to do the SPL calibration if the sweep level remains at -12dBFS?


Edit:
3. I know you said earlier that one would generally use one speaker to calibrate the level and the SPL reading. But if I am measuring two speakers at the same time (meaning playing the sweep through both speakers), wouldn't I calibrate it using both speakers? Because my hand-held dB meter will give a reading of 75dBC for one speaker, but with two speakers playing equal sweep at equal level, then the actual SPL would be higher than the original 75dBC when measured when the sweep was playing through one speaker.
 
Last edited:
Hi Gallantus. In a 24 Bit Recording Resolution even at -48dBFS would still be 16 Bit, i.e. CD quality. Sweeps or Noise on Two Speakers would be 6dB Louder I guess if the speakers were stacked, but let's say 3-5 dB for spaced speakers. Sweeps of say 75-85dB should deliver a very decent S/N Ratio, but 85dB at HF can get old pretty quick.
 
In a 24 Bit Recording Resolution even at -48dBFS would still be 16 Bit, i.e. CD quality.
Can you elaborate this? -48dBFS in the output or input?


Sorry, as I've mentioned earlier, I'm a newbie.
 
Sorry, I have to spell it out.

I stated that: "Some sources recommend anywhere from -24dBFS to -12dBFS, some just straight out say use -18dBFS."

And you are saying I can calibrate the input level of REW to as low as -48dBFS and it will be fine as that is still 16-bit resolution? Is that what you mean? If so, do you mind sharing the calculation of how -48dBFS equates to 16-bit resolution?

EDIT:
I think this is the calculation.
(2^24)/(2^16) = 256
1/256 = .0039
20xlog(.0039) = -48dBFS

I get the math, but I don't see the relationship as it relates to the mic gain level. When a mic gain is increase, it increases the signal, but does it not increase the noise level in the electronics?
 
Last edited:
You seem to be trying hard to overcomplicate things. SPL calibration will work for any input level more than about 10 dB above the noise floor. Typical input levels with a mic are from about -60 dBFS to about -20 dBFS. The measurement process is very tolerant.
 
You seem to be trying hard to overcomplicate things. SPL calibration will work for any input level more than about 10 dB above the noise floor. Typical input levels with a mic are from about -60 dBFS to about -20 dBFS. The measurement process is very tolerant.
Yeah, I have self-diagnose OCD, so I tend to unconsciously go into the rabbit hole. I really do appreciate you being patient and taking the time to help.

More importantly, I also don't want to screw up the measurements, because I plan to send it to GIK for them to look at and make recommendations on room treatment options. If I screw up on the measurements, then I am not providing good data for them to recommend the right solution.

Having said that, let me recap for your confirmation:

1) When calibrating mic input level, -60dBFS to -20dBFS will do fine, so long it's 10 dB above noise floor?

2) For a two-channel stereo system, when calibrating output level or SPL reading, use one speaker when measuring one speaker and use both speakers when measuring both speakers?

3) If the output level is calibrated at 75 dBC using a handheld dB meter already, then when calibrating the SPL reading it should already be 75 dBC, assuming I do not change the sweep level, mic positioning and any gain level along the signal path? But I still have to go through the moments to calibrate the SPL reading so that REW knows how to show levels as correct SPL figures?

Thanks again John!
 
1) When calibrating mic input level, -60dBFS to -20dBFS will do fine, so long it's 10 dB above noise floor?
Yes.

2) For a two-channel stereo system, when calibrating output level or SPL reading, use one speaker when measuring one speaker and use both speakers when measuring both speakers?
Use one speaker to calibrate the SPL reading, it won't need to be done again unless you change the gain along the input path.

3) If the output level is calibrated at 75 dBC using a handheld dB meter already, then when calibrating the SPL reading it should already be 75 dBC, assuming I do not change the sweep level, mic positioning and any gain level along the signal path? But I still have to go through the moments to calibrate the SPL reading so that REW knows how to show levels as correct SPL figures?
There is no output level calibration.
 
One follow up question. You mentioned that the mic level input can be -60dBFS to -20dBFS, so long it's 10 dB above noise floor. Do you mean 10 dB above the noise floor of the room's ambience environment? If so, I presume I would measure the noise floor of the room using my hand held dB meter, which measured in dB SPL. How do you correlate dB SPL to dBFS?
 
@gallantus Sorry to add to the complexity of your stew. That -48 example was to illustrate that in a world of 24 Bit pretty much any old level mid meter is perfectly useable. What input chain are you using, what mic, what preamp?
 
I am strggling to find why the input level doesn't match the output level when doing a loopback test.

MY DMM reading matches the sweep or generator output and hence the input level, but nothing I do will make the input match. It's always about 10% less - ie 1v out, .9v shows on the level check displays and sweeps.

Tried two Behringetr UAC222 with identical results. Windows 10, with excl usb speaker and mic for out and in respectively. Using the latest version.

SOLVED. (seemingly)

I went to Preferences/Cal files and under mic cal files chose the first one, clicked on the setting icon and then altered the FS sine Vrms until the loopback input matched the output.
 
Last edited:
The level matching refers to dBFS values, i.e. digital levels. You can calibrate the input levels to show correct voltages using the button on the RTA or by manually editing the calibration value, as you did.
 
@gallantus Sorry to add to the complexity of your stew. That -48 example was to illustrate that in a world of 24 Bit pretty much any old level mid meter is perfectly useable. What input chain are you using, what mic, what preamp?
I am using a Neumann MA1 (yes, with factory calibration file), with a Topping 2x2e sound interface. I also have a miniDSP mic, but not using it right now.
 
Back
Top