What Is A Noise Floor?

Meanwhile, noise floors are the minimum level of a system. Any acoustical or audio system has a noise floor, right down to and including the thermal noise inherent in wires (approximately -125 dBM). The noise floor of a recording is the sum of all the noise floors encountered during the recording process. In acoustical systems, the noise floors have their Lmax at low frequencies. If you measure the noise floor of a typical room, you’ll notice that the primary energy is in the bottom two octaves, and that the noise level rolls off steeply with frequency, getting quite low in level in the top several octaves.

In audio systems, 60 Hz. hum is often a limiting factor for the noise floor. Such hum is, usually, a sine wave at 60 Hz. Sometimes, it will include harmonic components at 120 Hz., 180 Hz. and so on. If it’s 60 cycle buzz, then you’ll hear harmonics spread all across the audible spectrum. Bad, bad.

Aside from hum, the noise floor of electronic devices in general (like mic preamps, f’rinstance, or analog magnetic tape) is white noise, which is equal power per frequency unit. Therefore its output rises on a 3 dB per octave slope. White noise sure is audible! While the acoustical noise floor of rooms occurs generally at low frequencies that we don’t hear very well (consult the Equal Loudness Contours, a/k/a Fletcher-Munson Curves), the noise floor of electronics is right where we hear best. Not good.

Now consider what happens when we sum the noise floor of the electronics with the noise floor of the room: The resulting noise floor has high levels at low frequencies overlaid with white noise hiss. Rising and relatively high levels at both ends of the spectrum! Definitely not good.

However, it’s not the end of the world. We can record in very quiet rooms with little low-frequency content, and we can use sensitive microphones fairly close to our artists. This way, we keeps both noise floors low. We figure that, hopefully, we’ll never hear them because of the presence of the signal.

It is the phenomenon called masking that lets us assume that (a) we won’t near the noise floor when a signal is present, and (b) that the noise floor is the lower level of audibility of a system or signal. So let’s look at masking briefly.

What Is Masking?

Masking refers to the physical and psychological effects of one sound concealing the presence of another. This is a complex topic from the field of psychoacoustics, and there are a lot of weird variables. In it’s general form, however, masking is the phenomenon where one component of a sound is rendered inaudible by the presence of another component. For instance, you can have a trombone player play middle C at fortissimo along with a flute player playing the same middle C at pianissimo. When you do this, you aren’t going to be able to hear the flute. The trombone masks the flute. That’s masking, in a nutshell. A couple of verities (from Cyril Harris’ Handbook of Acoustical Measurements) are worth noting:

1. Louder sounds tend to mask softer sounds.
2. Sounds with the same frequency are the most likely to be invoke masking.
3. The softer sound, if it is the same frequency as the louder masking sound, can be detected even when it is more than 10 dB softer than the louder one.
4. Louder, lower frequencies tend to mask softer higher frequencies more than the reverse. (Louder, higher frequencies have less tendency to mask lower, softer ones.)
5. The greater the ratio between the masking and masked frequencies, the less masking occurs, so that the softer tone can be increasingly soft and remain audible as the ratio between it and the louder frequency increases.
6. The louder the level of the masking tone, the broader the range of frequencies it will mask.

With a steady state signal present at both ears, an 80 dB SPL 400 Hz. tone or noise band will render a 1 kHz. tone at 40 dB SPL inaudible. That same 80 dB SPL 400 Hz. tone or noise will not mask a 4 kHz. tone at all!

There’s lots, lots more (I mean LOTS more) to this masking stuff, but this is enough for right now.

So, How Big A Dynamic Range Can A Single Sound Have?

I decided to see for myself if the above statements were really true, so I hooked up a couple of oscillators and pink noise in the studio, sent ‘em all to one speaker and to the RTA and started messing around.

Sure enough, I could hear a low level 4 kHz. tone almost regardless of what or how much low frequency was present. For instance, even with a 400 Hz. tone at 95 dB SPL, I could hear a 4 kHz. tone at 26 dB SPL. This was particularly interesting because I had the windows open and here, in late August, the crickets really get loud. They were 6-10 dB louder than the 4 kHz. tone in a band from 4 to 6 kHz., and I could still clearly hear the tone below them!

Studying noise vs. tone was also interesting. I could hear a 4 kHz. tone @ 50 dB SPL below 67 dB SPL broadband pink noise. I could hear music at 35 dB SPL below noise of 42 dB SPL. Interestingly, at louder levels of pink noise (80 dB) the music was pretty well masked until it was only about 3 dB below the noise.