My son is doing a middle school honors science project about low frequency noise. He has been going about town with a handheld recorder and recording 60 seconds in various areas to compare the presence of LFN.
The files look great using the Plot Spectrum tool (Analyze > Plot Spectrum > Size: 131072). It is clear that there are more LFNs in predictable areas. But then we have this problem where the dB are in full scale. The range is looking like -39 dBFS to -90 dBFS.
He has put in his literature review portion that LFN at low volume (0-30 dBA) can induce sleepiness or calm but at high volume (over 100 dBA) feelings of annoyance or dread. This is kind of the kicker of his little project. I don’t want to tell him to take out volume all together because it seems important.
What am I supposed to do to get these dBFS values into something explainable?
You can calibrate dB SPL if you have an SPL meter. But that assumes the recording volume isn’t changed/adjusted.
There are SPL phone apps but the mics built-into cell phones aren’t calibrated. You MIGHT get “close” if the app asks what phone model you are using.
Typically, you’d calibrate with a 1kHz test tone.
There is a direct correlation between dBFS and dB SPL but no automatic calibration. i.e. A 10dB decrease in SPL results in a 10dB drop in the dBFS level (a bigger negative number).
For the A-Weighting you might be able to approximate it with the Graphic EQ or EQ Curve.
Or REW (a “room measurement” application) has weighting options. But I think they are assuming that you’re using a calibrated microphone “live” and I’m not sure if it works with recordings.