Cocktail party phenomenon

The other day, when Dad was talking about harmonics, he mentioned that one of the ways humans hear sounds is through something called the "party phenomenon."

Ah, you must be talking about the cocktail party phenomenon.

What is a cocktail party?

"It's a standing-room-only party with alcohol and light snacks. Party venues are usually noisy with a mix of conversations and music. Even in such a place, you can still hear the voices of the people you're talking to, and even if they're a little far away, you can still hear things you're interested in or things you're trying to listen to carefully, which is how it got its name. It refers to the way human hearing works, allowing us to selectively extract and hear specific sound sources from among multiple overlapping sound sources."

So, even if there's a lot of noise around you, if you focus your attention, you can pick out only the sounds you want to hear? Come to think of it, during recess, even when everyone in class is making a racket, if I hear my name and think they're talking about me, I can sometimes understand what they're saying. It's like you can hear through the power of attention, isn't it?

"Of course, if the level of the sound you want to hear is too low compared to the level of surrounding noise, you won't be able to hear it. But, as Oto said, it becomes possible to hear it through the workings of attention."

Also, how is that related to the overtones you told us about the other day?

This is going to be a bit complicated, but do you mind? In nature, it's very rare for only one harmonic component to be emphasized within a complex sound emitted from a single sound source. This means that if there is an emphasized harmonic, it implies that there is another sound source. Therefore, humans separate and interpret it as a sound coming from a different source.

So, if 10 people are chatting loudly in a class, can humans distinguish between 10 different sound sources?

I wonder how it works with that many. Research has shown that the process of distinguishing between different sound sources is handled by the part of the brain that controls hearing, which is connected to the inner ear by the auditory nerve.

The inner ear, which you explained earlier, has the function of analyzing the pitch of sounds, right? And then, further down the line, in the brain, it distinguishes between different sound sources, right?

That's right. A researcher named Pregman apparently named the process of reintegrating each sound source as auditory scene analysis. It's been found that, in addition to the common harmonics, sounds emitted from the same source start and end simultaneously, and that the characteristics of frequency and the direction from which the sound arrives change similarly, and these changes are gradual.

That's kind of complicated. The direction of the sound is also a factor, huh?

Since sound recorded with a single microphone cannot reproduce directional information, it becomes difficult to separate sound sources.

Human hearing is amazing, isn't it? Even if you record sounds and analyze them with a computer, I wonder if it's still impossible to distinguish between different sound sources?

While speech recognition technology seems to be improving, if we can apply the power of human attention to engineering, it might not be a dream for computers to distinguish between sounds playing simultaneously. It's believed that humans possess this ability because it's a function that vertebrates have evolved to survive by escaping larger animals and eating smaller organisms.