OT: Can you tell the difference between sounds behind you and in front of you?

[Mr Macaw]

According to this:

"some people can tell if the sound is in front or behind them."

I've never heard of that before. Can you?

[Don Y]

Depends on the shape of the pinnae and the quality of your "auditory meatware".

Azimuth (left/right) is relatively easy for most folks. Elevation gets a bit harder for many. Front/rear is the hardest (esp if you want to express WHERE to the rear, etc.).

Your brain tends to rely on time differences, amplitude differences and frequency attenuations to discriminate direction.

A sound coming from the left is louder at the left ear than the right (because the right ear is in your "head's shadow"). It is also slightly earlier than its arrival at the right ear (a foot is about a millisecond delay). Additionally, higher frequencies are attenuated more than low due to the "meat" in the middle. And, finally, environmental characteristics (room acoustics, etc.) come into play (your brain adds context to better qualify the data that your ears are providing.

For a hoot, visit an elliptical room (my favorite exhibit at the Museum of Science & Industry, Chicago) and stand on a focus (ellipse can be considered a circle with two "centers") while a friend stands on the other. Your BRAIN knows your friend to be ~20 feet away. And, you are SURROUNDED by noisey tourists walking in every direction -- including directly behind you!

Yet, when your friend *whispers*, you will swear their lips are inches from your ears!

[Mr Macaw]

Now that I don't believe. I'm sure it's all volume. The time difference would be far too small to be of use. The volume is much more obvious.

There was a film once where a blind man could tell exactly where everything was in a room by echo location when he made a noise.

I've heard of those, but never tried one.

[Don Y]

You'd be wrong. ITD (Interaural Time Differences -- differences between left and right ears in the time domain) are used for low frequencies (e.g., below about 1KHz). Low frequencies are not attenuated much (which is why you can hear the bass notes from the guy's sound system in the car next to you through closed windows -- but can't hear any of the vocals!)

ILD (Interaural Level Differences) are used for higher frequencies -- where the signal experiences *more* natural attenuation.

*You* can perceive time differences on the order of 10 MICROseconds.

You can probably resolve azimuth (left-right) to within *1* degree if the sound is sort-of in front of you. Once it starts to move off to the sides, your "resolvability" gets much worse (10-15 degrees).

The most interesting aspect of this is that you can recreate these "situations" on a computer -- with headphones -- and the listener will hear the sound as originating *inside* his/her head!

[I've written a "3D spatializer" that allows me to present sounds to a (blind) user in much the same way that your eyes would perceive "notifications" in the visual field. Ideally, it has to be tailored to each individual user as each user's "audio meatware" is different.]

Notice how often you MOVE (tilt) your head when trying to identify the location/source of a sound! Or, "focus" on it.

Ever notice how you can hear someone mentioning your name in a crowded room -- despite being unable to hear anything else they are saying??

Hearing (like vision and the other senses) is delightfully complex!

[(PeteCresswell)]

Per Mr Macaw:

Depends on the sound.

Most sounds: yes.

But the warning sounds that the local emergency vehicles make are so loud or *something* that I cannot tell where they are coming from.

Whoever buys that stuff probably thinks louder is better - like the morons who set the sound level in theaters... but it seems counter-productive to me.... but maybe it's just me...

[Don Y]

It is a combination of loudness, narrow frequency range and "anxiety". Add to that the "canyon" effect in many roadways (buildings on each side funneling the sound in a corridor).

There has been some research that suggests you need a combination of an "attention getter" (like the current siren) coupled with a wide-band noise source (imagine the sound of a crowd of people) to help folks sort out the direction of the "alarm".

Each intersection, here, has a strobe light situated atop one of the lamp posts/traffic signals *in* the intersection. A strobe light on the emergency vehicle signals the intersection that it is approaching. The traffic controls then automatically adjust/abort the current cycle to favor the direction of the approaching emergency vehicle (the thinking being to get any cars sitting at the light MOVING, out into the intersection and "mobile" so they can avoid the emergency vehicle).

We have accustomed ourselves to watch for the blinking strobe light *in* the intersection (it blinks as an apparent acknowledgement of the strobe on the emergency vehicle) to start looking (and listening) for the vehicle -- expected shortly.

[SeaNymph]

No I can't since I've lost a lot of my hearing.

[clare]

Standing out in front of the hangar (between 2 rows) when Gerry is punching holes in the air out in the acrobatics box you'd SWEAR the plane was out beyond the hangar across the way - when it is half a mile the other way. The noise is blocked by the hangar behind and bounces off the hangar in front. Combination of that, the low tone of the exhaust, and the volume really screws up the "aural postioning" sense.

[(PeteCresswell)]

Per Don Y:

We have the auto-green for approaching emergency vehicles but, AFIK, not the strobe lights.... OTOH, maybe they are there, but I have not seen them.

Last couple emergency vehicles I watched managed to charge into the intersection before the light had even changed for them.... And when it does change, it's just a *flick* and it's red.... no warning for the other traffic.

Sounds made-to-order for revenue producing tickets.

[Doug Miller]

"Mr Macaw" wrote in news: [e-mail]:

small to be of use. The volume is much more obvious.

OK, don't believe it. It's true just the same. The human brain is *very* sensitive at distinguishing the timing of sounds. We're much better able to tell which of two nearly simultaneous sounds occurred first than we are which of two nearly simultaneous sights.

[Don Y]

The strobe light tells us (me), "watch to see whether you get a quick red

*or* an early green". This then tells me the vehicle is either coming up from behind me/approaching head on (in the case of *me* getting a green) OR coming from the right/left sides (in the case where I get a red).

In essence, it cuts my choices in half so somewhat improves my odds of noticing the vehicle BEFORE it gets to the intersection.

Here, an emergency vehicle can often get tied up BEHIND traffic queued at a light. As we have median strips at most of the larger intersections, it's not easy for a vehicle to cross into the "wrong" lane to navigate around cars "parked" at the light.

Prior to the strobe gizmo, there were numerous cases where I found myself at the head of my lane and took the initiative to CAREFULLY cross the intersection on RED just to make a hole for the vehicles behind me to get out of the way.

[Of course, all the "seniors" sit their stupefied: "Should I go? But the light is red!?" Instead, they wait until the emergency vehicle is on their back bumper leaning on it's "foghorn" to make up their mind that they MUST go...]

[Paintedcow]

Yes I can for most sounds. Running a farm, I often have to determine which of my animals is making sounds. The direction often is important especially in the dark, and it sounds like one of them is having a problem. With the horses, I can usually know their voice and which one it is. They all make a different sound, and some make more sounds than others. I can usually determine the approximate location of wild animal sounds too.

Bit sirens and other very loud sounds are hard to locate. nd if I'm driving it's even harder because of the "shell" of the vehicle.

[makolber]

Yes small head movements allow you to tell the difference between front and back.

[Muggles]

I often can't tell where sounds are coming from, in addition to, I have a hard time identifying some sounds because I've never heard them before.

[gregz]

Don't need any research to figure backup beeper alarms have little direction capability. Sine waves don't work. Square waves do work.

Greg

[Robert Green]

difference would be far too

of two nearly

simultaneous sights.

Agree. The human bra It takes a rather simple neural comparator "circuit" in the brain to detect which sound arrived first. That type of ability is found in pretty primitive animals. But clearly intensity and timing are both integral parts of 3D auditory location techniques. IIRC, bats, owls and certain prawns easily can detect phase shifts and use the doppler effect to determine the speed of their prey.

It's remarkable that such complex structures arose within animals. The different neural net sensors in our bodies is much like photocells, microphones, thermocouples, piezo-electric pressure sensors and more evolving on their own. Your hair follicles are sensitive enough to know which way your hairs are aligned and to erect them on neural command.

[(PeteCresswell)]

Per Don Y:

At least some of them are thinking of the possibility of some Super-Trooper standing by to issue tickets for running a red light.

[Don Y]

"Photo-enforcement", here.

Well, until recent citizen's initiative rewrote the city laws to make it illegal to use a camera for traffic violations.

Yesterday, I was at one of the intersections that had previously been monitored by a camera. THE "intersection", here, is defined by extending the curb lines of the crossing streets. I.e., the "stop line" (in this case) is a full 30 feet BEFORE the "intersection".

Must stop before your vehicle ENTERS the intersection, as light turns red.

So, I found myself stopped 30 feet PAST the stop line, PAST the crosswalk feeling pretty much "exposed" -- just to ensure I was in compliance with the law... "just in case". Lots of folks behind me annoyed that I'd not continued through the intersection ("Hey, why are you stopping? Cameras have been turned off!")

[Don Y]

Play two sine waves at equal amplitudes and fudge the phase of one relative to the other.

Note that you "consciously" perceive sounds arriving within several milliseconds of each other as "one" sound. (I think around 50ms is where you start to become conscious of "echoes", etc.). Yet, your brain is processing times 1000 times shorter than that as it attempts to "localize" those sounds!

"Three times greater" cuz we have fatter heads! :> Three times longer distances involved!

Some creatures rely on more "mechanical" means (e.g., directly coupling the eardrums so the "difference" is available as a directly observable signal).

So much for the "intelligent design" theory! I.e., if you can create EVERYTHING (including "radiation"), then why create a system that requires so many different solutions to the same problem? Why don't

*our* ears move like those of dogs? Why aren't our eardrums directly coupled like flies? Why don't all of our pinnae look the same? etc.

OTOH, if you are ADAPTING to an environment IMPOSED on you, you EVOLVE solutions that address those particular needs in the context of your own orgainsm, etc.

[Don Y]

In addition to time and intensity, there are also differences in how particular frequencies are perceived. I.e., higher frequencies are attenuated more (by the "acoustical shadow" cast by your head) than low frequencies.

The same sort of experiment (outlined above) is performed on people to quantify their HRTF (Head Related Transfer Function -- how different types of sounds are perceived IN EACH EAR CANAL based on where they originate in the three-dimensional space surrounding them).

In: you can see microphones inserted in the subject's ear canals (as close to the tympanic membrane as feasible).

The baffles in the anechoic chamber (floor/walls/ceiling) damp out reflections.

The speakers arranged in the arch over the subjects head present the sound source in one of N "relative directions" (the subject's head is located in the "center" of that arc).

The arc pivots so that the speaker array can be behind, above, infront, etc of the subject. So, in effect, you can issue the sound source from any position in a SPHERE surrounding the subject's head.

You can replace the human subject with a KEMAR head and obtain data for a generic head (vary the size, change the shape/orientation of the pinnae, etc.).

Unfortunately, the data from one subject (or, a KEMAR) isn't directly applicable to other subjects. Normal variation in our anatomies means that the HRTF for *your* head is different from that of any other person.

The takeaway from all this is that, in theory, you can manipulate any sound source, mathematically, and present it to a user via headphones and convince him that the source is at a specific point in space relative to his head. *If* you have HIS particular HRTF encoded in that mathematical algorithm!

Where this falls down is in the dynamics of perception: you will tend to react to a sound stimulus by altering your position, head orientation, etc. relative to your guesstimate as to where the sound is located. I.e., you'll twist your neck, tilt your head, etc. -- waiting for a repeat of that signal to give you a refined estimation of where the source is located.

With the synthesized approach, moving your head MOVES THE SOURCE (because the headphones don't know that your head is now pointed in a different direction -- so, the signal that it presents is STILL "off to the left" just as much as it was before you moved your head!) You can work-around this by using a head-tracking device and dynamically tweeking the math (signal processing) to reflect the head's motions (to adjust the signals to each ear so the sound source stays fixed in space as the head reorients to try to refine its position).

[In my case, I rely on different *sounds* in different *places* so the user doesn't have to know exactly where a sound originates: "Ah, that signal off to my left means someone is at the front door; had it appeared to my right, I would know it was the BACK door!"]