at school we’re doing a project about noise pollution, and it requires us to record background noise and get some data on it, such as average dB level and where the noise is in terms of frequency.
The frequency part seems fairly easy with a spectrogram, but I have a few questions about the dB levels.
I already know that in a recording it’s not possible to get actual dB levels as they’re gonna be relative to the device that outputs the sound, but if I were to find what’s the smallest level of sound my microphone can pick up (let’s say -58 dBFS, which if I understand correctly is the unit of measurement Audacity uses, ranging from -60 to 0, which is the loudest sound the device can output without clipping) and consider that as “0 dB” (even though there would still be some background noise when I measured the lowest sound the microphone could pick up) would that mean when the sound in the waveform reaches 0 it would equal 58 dB at time of recording? If so, how do I find values above 58 dB? Also is there a way to get the average decibel level for every T seconds? Thank you.
Correct. You would need to calibrate your system some how.
This might not provide a reliable calibration. The “smallest level” will be determined by the self-noise of the recording device, including microphone self-noise and sound card self noise, and mechanical noise (computer fan) picked up by the mic.
Better would be to record a sound of known loudness. The way I would do this would be to use a calibrated SPL meter and note the measured SPL with the SPL meter next to the computer’s mic. So, for example, if the SPL meter shows 70 dB (unweighted) when recording the school bell, and that produces a waveform with an RMS level of -6 dB in Audacity, then you know that “Full scale” (0 dB) in Audacity is equivalent to 70 + 6 = 76 dB FS SPL.