Locating sound

[tulamide]

Is someone here able to explain without scientific special terms, but with simple words, why we are able to locate sounds threedimensional and 360° on all axes?

I understand the concept of a signal reaching one ear earlier than the other, thus carrying a location information. But how do we sense sound in front or in the back. I'm talking of natural sounds, not some 7.1 surround sound system.

My totally not funded in anything guess would be some reflections, but those would then still be coming from the back or the front, which still doesn't help me understanding, how we sense the front and back (or top and bottom) at all with just two ears on the left and right.

For example, I live by the sea. When one of the big cruise ships blow their horns, I can locate it exactly in direction (in front, slightly to the left) and even roughly in distance, without having heard the specific horns before. Why? How?

[aronb]

Hi,

The simple answer is "Phase AND Amplitude"...
Years back, I created a system (all analog too) that took an input sound (stereo or mono) and could move that "sound image" around a room with 2 speakers. In headphones it was amazing!

I believe the "Spacializer" unit does this as a professional audio product, but I do not know if it is still made.

So, Phase and Amplitude of:
Left - Right
Up - Down
Near - Far

Then the brain processes this phase information into a location we can track, Devolve and Convolve this phase information on the fly.

In a related topic - Take a look at the size of a dolphin brain, the area that processes echo location (From a Single Source no less) devolves the the echo into a 3D location!

Aron

[k brown]

Phase in the sense of time delay, yes. A concept to look up that explains much of this is HRTF - Head Related Transfer Function. It outlines the 'macro' of human directional hearing. It's mostly based on Interaural Time Differences (sounds arriving at the two ears at different times, due to the physical space between them), and frequency response differences due to the unique passive filtering of each pinna, and the high frequency blocking effect of the head itself. This effect disappears at around 400 Hz, as audio wavelengths below this are physically larger than the dimensions of the head, therefore it's very difficult to perceive the direction of sounds below 200-400 Hz. - and is the reason single subwoofers can cross over as high as 200 Hz.

Front-back distinction is mostly down to the forward-biased pinnae, combined with the fact that the ears are not centered between the the front of the head and the back (closer to the back of the head than the front) and is actually rather easily fooled. It's been shown that determining front from back requires very small, mostly unconscious head movement to percieve the fore/aft direction.

The ORTF mic array sounds so realistic because it crudely approximates the HRTF - directional mics spaced approximately head-distance apart, aimed at an angle that approximates the forward-bias of the pinna; the directionality (cardioid) approximating the blocking effects of the head. Other HRTF-related mics are the Crown SASS and the Schoeps sphere mic. For obvious reasons, any coincident mic system cannot approximate human directional hearing (with no space between the mics, no time difference).

[MichaelBenjamin]

.

[Spogg]

I think all the responses cover the topic well.

I looked into this in great detail before I made my Quilcom Whirlipan which aims to fake rotation around the head:
viewtopic.php?f=3&t=4464&start=0&hilit=whirlipan

There’s info in the download and the topic text that might be useful.

[adamszabo]

This video will answer all your questions

https://www.youtube.com/watch?v=Oai7HUqncAA

[tulamide]

This video will answer all your questions

https://www.youtube.com/watch?v=Oai7HUqncAA

That's exactly what I needed! Simple language, great visualizing. Thank you

[k brown]

One thing that may explain something they didn't go into in the video is why only the woman was able to perceive how high up the tree the drone was - my hunch is that it's related to women generally having more extended high frequency hearing than males of the same age; this may give them more resolution regarding height perception, especially from a distance, due to air absorption of high frequencies; her brain simply received more of this delicate spectral difference information.

A fantastic series of CDs recorded with the Neumann dummy head microphone, was the Gordon Hempton 'Earth Sounds'; now out of print, but other recordings can be purchased on his website.
https://www.soundtracker.com/

[Spogg]

This video will answer all your questions

https://www.youtube.com/watch?v=Oai7HUqncAA

Excellent video!

The same guy also makes videos with binaural recordings. I checked a couple and got my usual reaction: It’s nice but it’s just wide stereo to me. No front-back cues in my case.

Some time ago I assumed this was because everyone has different pinnae, as per the video. So I bought a pair of earphones with built-in microphones specifically to make personal binaural recordings. I used them with a portable sound recorder and went out and about.

I listened back on the recording earphones rather than any other headphones but unfortunately I got the exact same impression: wide but no front-back. My conclusion was that the ear-bud mics were effectively altering my personal ear shape enough to kill the perception.

Maybe it’s my brain that’s too experienced in assessing sound, I don’t know.

When I was developing my Whirlipan I found I could get something like a front-back effect but only when the sound was fake-rotating. When static the illusion disappeared. That’s why I called it the Whirlipan rather than a 360 degree panner.

[k brown]

I found that the effectiveness of binaural recordings (especially as regards front/back localization) varies a lot with different headphones - some scramble the localization cues pretty badly. Closed phones tend to be the worst. Open-air types tend to be much better. The very popular Sony MDR-V6, for example produces very little binaural illusion, whereas the humble little Koss Porta-Pros do very well.

This brings up an inherent flaw in the way a lot of binaural recordings are made and listened to. True binaural recordings are made with dummy heads with realistic pinnae, and mics embedded inside the artificial ear canal (like the Neumann head). For greatest accuracy of reproduction, these recordings should be listened to with in-ear headphones, not on-the-ear or over-the-ear types; simply because the listener's pinnae should not add their acoustic filtering to that captured on the recording by the artificial head's pinnae. The sound should be 'pinnae-filtered' only once. On/over-the-ear headphones work best with quasi-binarual recordings, made with mics that don't have artificial pinnae (the Crown SASS, the Schoeps sphere mic, Jecklin disc, etc.).

The very binaural illusion also tends to vary a bit from listener to listener. Same with 3D photographs; stereoscopy. There are many people that perceive little depth illusion from stereo photos.

Also the front/back perception can be very difficult for sounds exactly centered front or back - moved just a bit off-center and the front/back distinctions become clearer.

All that said, it's been my experience that the main impediment to the binaural illusion are the headphones used - some just mangle the binaural cues very badly.