Which note on a guitar corresponds to 440 Hz?
That's the note on the 5th fret of the 1st string. On a guitar, one fret is a semitone, so the next note, A, is on the 17th fret of the 1st string. So, a semitone is 12, and an octave is 1/12 of an octave. On a piano keyboard, adjacent notes, including the black keys, are a semitone, so they are also 1/12 of an octave apart.
So, a third octave is four semitones, which is two whole tones, so it's the interval between C and E, right?
That's right. An octave is a doubling of frequency, but for sound analysis, instead of using do-re-mi, a round number of center frequencies based on 1000 Hz is used. Like 63 Hz, 125 Hz, 250 Hz, 500 Hz, 1000 Hz, 2000 Hz, 4000 Hz, 8000 Hz, and so on. I told you before that humans can hear sounds from about 20 Hz to 20 kHz. But the sounds that exist in our living spaces are generally between 63 and 8000 Hz, so it's common to evaluate them using these 8 bands or so.
Oh right, I haven't talked about bands yet. An octave band is a frequency range with a width of one octave. The formula is complicated, so...
Without going into too much detail, an octave band with a center frequency of 1000 Hz has a frequency range of 770 to 1410 Hz, and an octave band with a center frequency of 500 Hz has a frequency range of 355 to 710 Hz. The level of sound intensity contained within one octave is called the octave band level. When analyzing sound, the easiest and most general approach is to evaluate it using a single number, such as dBA, for noise level. However, with noise level alone, you can't tell whether the noise is mainly low-frequency sounds or high-frequency sounds. Therefore, by analyzing using octave bands, or even more precisely, 1/3 octave bands, you can capture the frequency characteristics of the noise.
But why is such an analysis necessary? If there are no unusual noises, can't we evaluate it based on the noise level?
Consider taking measures to mitigate noise. Knowing the noise levels for each frequency will help you determine which frequencies need to be addressed. For example, when measuring road noise, if you have not only the noise level but also the octave band measurements, you can estimate how to design a sound barrier to achieve the optimal design, and also what the noise level will be like in living spaces such as houses near that road.
I see. Noise levels are sufficient for evaluation, but when considering countermeasures and predictions, octave band analysis is effective.