But why did the 200 Hz sound lower than the 4 Hz sound when we compared them? They both started at the same 1000 Hz...
Human hearing is fundamentally designed to perceive periods. Pitch is nothing more than a period. However, as we approach the lower limit of the audible frequency range and the period lengthens, it no longer sounds like pitch, but rather like a pattern with a certain rhythm.
"Beat" follows the same pattern, doesn't it? Humans can't hear a 4 Hz sound as a pitch, so it must be that the original waveform has a frequency of 1000 Hz, and the 4 Hz period is what we hear as a pattern. But that's unrelated to why a 200 Hz sound sounds low, right?
I vaguely remember that when two sounds with similar frequencies are present, one sound masks the other.
That's right. The lower note masks the higher note, doesn't it?
I see. I understand. In this graph, the 1200 Hz sound is masked by the 1000 Hz sound, but the 800 Hz sound is not masked, so does that mean that sounds lower than 1000 Hz were heard?
Correct! Masking occurs when the difference between each frequency is within the critical bandwidth (the bandwidth in which human hearing can distinguish between different pitches), and in this case, it can be considered to be just within that range, so the masking effect is at work.
When Dad said that the fluctuating frequency starts at around 30 Hz and gradually gets lower, it means that we are hearing a peak sound that is lower than 1000 Hz by the fluctuating frequency, and the difference in that sound gradually increases downwards from 1000 Hz.
That's right. In fact, when the difference between the two sounds that Oto initially asked about widens, the resulting superposition of the two sounds takes on almost the same shape as a waveform with amplitude modulation. This diagram (Figure 4) shows the superposition of the waveforms of 1000 Hz and 1200 Hz sounds.