Binaural effect

father

Today, I'd like to talk about the "binaural effect" that I briefly mentioned the other day.

Otokun

You previously explained that there's a time difference between sounds entering each ear, and that the shape of your head is also a factor. You also mentioned that when you're equidistant from both stereo speakers, the sound is localized in the middle, and when you move closer to one of the speakers, the sound source's localization shifts to the closer speaker.

father

Even with several sound sources, the first sound that reaches the ear allows us to perceive the location of the sound source—this is called the "preceding sound effect." It might be better to say that the ability to pinpoint the direction of that sound is due to the "binaural effect." (Figure 1)
Also, the sense of spatial expansion is a perception made possible because we have ears on both sides of our heads, and this is also called the "binaural effect."

Otokun

When you say "spatial expanse," do you mean the acoustics in a concert hall?

father

You can still perceive the reverberation of a space even if you cover one ear. To get a little technical, it's known that there are two types of perceptions of "spaciousness" within a space. One is the feeling of being "enveloped in sound," which is closer to the image of spatial expanse. The other is the sensation that the sound source itself has expanded, which we call the "apparent width of the sound source." In both cases, the perceived intensity changes depending on the structure (direction, time, and magnitude) of the reflected sound that arrives after the direct sound. (Figure 2)

Otokun

That's kind of complicated. It's easy to imagine that we can tell the direction of a sound because we have ears on both sides, though.

father

That's right. So let's consider the case where there is only one sound source in a room with no reverberation (an anechoic chamber). When the sound comes from directly in front of you, there is no time gap between the sound waves entering both ears, so you perceive the sound source as being directly in front of you. However, the same is true if the sound comes from directly behind, so because there is no time difference, we often mistake sounds from directly behind for sounds from directly in front of us.

Otokun

Oh, really? I don't think I've ever mistaken a sound coming from directly behind for one coming from directly in front.

father

One reason is that in a normal room, reflected sounds from the walls and ceiling, other than sounds coming directly from behind, are what we perceive as sound coming from behind. Another reason is that sounds almost always carry meaning, so the moment we hear a sound, we combine it with the surrounding environment information we normally perceive to determine what kind of sound it is.

Otokun

Experiments that involve presenting sound in a laboratory setting where there is no information other than sound, and no reflected sound from walls or ceilings, are different from what we normally experience.

father

That's true. But with these kinds of experiments, you have to eliminate as many uncertain factors as possible and keep them simple, otherwise you won't get the real results. You need to be able to synthesize what you've learned from each experiment and explain what's happening in the real world.

Otokun

I see. What about angles other than directly in front?

father

"That's right. Oto, at what angle do you think people can distinguish the direction of a sound?"

Otokun

I'm not entirely sure, but maybe around 10 degrees?

father

Actually, in directions close to directly in front, there's a resolution of about 1 to 3 degrees, although it varies depending on the frequency. This is quite amazing, as the time difference between sounds entering both ears is only about 1/100,000th of a second (10 μs).

Otokun

You can distinguish even such subtle angles, huh? What happens when you move away from the front?

father

When viewed from about 60 degrees to the side of the front, the accuracy of directional localization drops considerably, and when viewed from directly to the side, it can only detect objects within a width of about 40 degrees. (Figure 3)

Otokun

Wow, I didn't realize there was such a big difference between the front view and the side view.

father

That's because, as I explained earlier, the left-right difference in the head-related transfer function becomes larger when the sound is directly to the side. So, even if the angle of incidence changes slightly, the difference between the left and right sides becomes smaller compared to when the sound is closer to being directly in front. Another reason is that when a person tries to determine the direction of a sound, they turn in that direction, so sensitivity near the front might naturally be higher. And importantly, although I've only talked about time differences so far, the clues to distinguishing direction are both time differences and sound pressure differences.

Otokun

Is the difference in sound pressure due to the different ways diffraction occurs on the left and right sides, which in turn causes a difference in sound pressure?

father

That's right. The two clues are related to frequency, and it's said that the time difference is effective below 1500 Hz, while the sound pressure difference is effective above 1500 Hz.

Otokun

It's complicated, isn't it? It seems like we need to gradually unravel it in a place like an anechoic chamber.

father

That's right. We talked about the cocktail party effect before, didn't we? This is also due to the fact that both ears can determine direction in order to selectively hear. Research into this kind of directional localization has a wide range of applications, including in robotics.