Positional audio question

[lostatlantis]

A friend was arguing that humans can only determine positional audio in the horizontal plane; more specifically, mainly left/right. I think he's wrong... who's right?

 

[Matthias99]

According to my neuroscience classes: We can't do it very reliably, but we CAN do it to some extent.

It has to do with the shape of our ears/ear canals -- there's a subtle difference in the way something sounds when it is 'above' you as opposed to 'below' you, since the sound waves hit different parts of your ear with different timing.

 

[KoolAidKid]

We certainly can, although not as accurately as in the horizontal plane. As I recall localization error when estimating elevation is around 30 degrees on average, but considerably less for azimuth judgments. The fact that our ears are mounted on the sides of our heads allows for many left/right cues to location, but fewer up/down cues. This is likely to be evolutionarily adaptive, as most prey/predators/mates are at eye level.

 

[CycloWizard]

If you want to demonstrate this, ask him what the first thing he does is when he hears a sound at a different elevation. He cocks his head, of course, to move the sound into the plane that his ears are most effective. I never thought about this before, but it seems obvious now after what the others here mentioned.

 

[lostatlantis]

Thanks for the feedbacks. That's basically what I had in mind, essentially similar to our vision which favors horizontal plane (hence the widescreen craze).

CycloWizard that's a very effective and intuitve exercise you pointed out.

So how come the music/movie industry hasn't been all over binaural recording? I've heard some samples and they sound pretty amazing in terms of positional audio. I can only imagine what a live concert recorded that way would like...

 

[blahblah99]

I don't think our ears have evolved to listen to sounds in the vertical plane... We INFER where the sound is coming from based on other sensory input and/or memory such as vision (seeing a bell ring above you), or memory - such as the sound of a helicopter.

 

[KoolAidKid]

Originally posted by: blahblah99
I don't think our ears have evolved to listen to sounds in the vertical plane... We INFER where the sound is coming from based on other sensory input and/or memory such as vision (seeing a bell ring above you), or memory - such as the sound of a helicopter.

This is incorrect. Listeners are quite able to make somewhat accurate elevation judgments using auditory information exclusively.

 

[harrkev]

See those funny thing sticking out of the side of your head? Most people call them ears. Doctors call that the "penna." The actual "ear" consists of the penna, plus all of the bits inside your head (ear drum, cochlea, bones, etc).

The penna is shaped to adjust the frequency of sounds coming from different directions, and it is this that helps people to determine the direction of sound. In engineering, a "transfer function" can be thought of as a filter. You put something in, and you get something else out. People who study 3D sound have developed a "HRTF (head-related transfer function). So, if you put in a sound and a position, you get a slightly different sound out. It is exactly this technology that lets 3D sound be emulated using 2 speakers or headphones.

Note that HRTFs do not work exceedingly well. But they do work a little. Just google "HRTF" for hours of reading fun.