Even with my eyes closed, I can detect which direction a sound is coming from. Now, I can see how I’d be able to tell whether a sound was coming from my left or my right, since it would reach my two eardrums at slightly different volumes and times. But how can my ears and brain tell the difference between, say, someone speaking ten feet in front of me and someone speaking ten feet behind me?
A given sound goes in both ears, but due to the spacing and shape of your head and ears, will hit the eardrum at different times, and with different attenuations. Your brain figures out which part of the resulting signal correspond to the same origin, and from the difference in attenuation and time that the signals got to it, estimate the location of that origin.
Can it? I’ve occasionally been confused by a sound coming from directly behind me, thinking it’s coming from directly ahead. This doesn’t come up very often, admittedly, (since the sound would have to be something that could reasonably have no visible source, or be in the dark, or with my eyes closed…).
It would be an interesting experiment: put a couple speakers directly ahead of and behind someone, have someone else randomly send sounds to the two speakers, and see if the person in the center could tell which was which. If you got the speakers positioned just right, I don’t think I could.
It depends upon the frequency spectrum of the original sound and the presence of scattering surfaces. As for doing experiments, that’s been done a lot. It’s a very big thing in 3d games, since the designers want sound-producing objects in a virtual environment, and want the player to feel immersed in the game. Look up head-related transfer functions (HRTFs) for far more information on the subject than you probably want.
Like Punoqllads said, there’s been a lot of research done on this topic. Here’s one book on the topic if you want to delve into it.
Again, it’s been done; and yes, there’s feedback between your sensory modalities. I did some time in a VR lab; one of the canned demos for spatialized audio was a recording of a helicopter flying in a circle around your head. If I kept my eyes closed, I could hear it flying in a circle around me; if I kept my eyes open, I could hear it flying in a semicircle back and forth behind me just out of vision range.
It’s not just timings, but the way that your outer ears, and in fact your whole head attenuate various frequencies. If you have a decent pair of headphones, track down a binaural demo recording. Those are done with a dummy “head” with microphones where the ears should be. The sound is as close to being there in person as I’ve ever experienced.
I can tell you for a fact that you can discern with a great deal of accuracy where a bolt has dropped just by listening carefully. I have learned to stop and listen whenever I drop a bolt or nut. I can usually localize it resting place to within about 1 square foot. I don’t twist my head to try and follow it visually, I just stay still and listen.
I believe that the shape of the ear, it being very deliberately (if you can call such things “deliberate”) non-symmetrical on any axis, are what allows you to triangulate where a sound is originating from. How it arrives at your ear drum will be different because of how it bounces around in the shell of your ear.
That’s my best assumption, anyway.
One thing which I think hasn’t been mentioned is that we constantly cock our heads to try and figure out how moving our ears affects the sound which helps localise.