Might seem like an obvious question but if I listen to someone speaking to me from say 6-10 feet away I'm not really aware of the 'room sound' - replace my head with a mic - even an omni - or a stereo pair and listen back to the recording and the room colours the sound sooo much. Why exactly is this. If the brain is filtering out the room in the real life situation why does it not do it in a recording? Even if video the scenario and listen while watching (to give my brain the same visual 'clues' as the real world situation) the results are the same.

Any thoughts?

JP

--------------------
Mac Pro, Logic Pro, lots of software and 17 hard drives!

Because when you're inthe room the brain has all the information from both ears to be able to work out the directions of reflectionsa nd their riming, and focus the attention on the wanted sound.

It doesn't get that wealth of information when you mic it, not even with binaural miking (although that comes close sometimes) because the exact spacing and the directional effects of the listener's own ears are missing from the equation.

Hugh

--------------------
Technical Editor, Sound On Sound

in addition to everything hugh said... there's the issue of the source no longer being in that physical position within that original physical location..

the source is now 2 speakers.. with a different physical relationship to the ears of the listener,, and a different interaction with a (possibly) different room...

so there is now potentially far more room sound involved, both that of the recording, and that of listening space.

Sometimes it helps to stick a finger in one ear when miking. Losing the information from the second ear can sometimes lessen the brain's ability to 'ignore' room reflections and focus on the direct sound, and you often perceive a much more accurate direct/reverberant ratio this way.

--------------------
"He that hears music feels his solitude peopled at once." - Robert Browning

Quote TurboD:

---

Sometimes it helps to stick a finger in one ear when miking. Losing the information from the second ear can sometimes lessen the brain's ability to 'ignore' room reflections and focus on the direct sound, and you often perceive a much more accurate direct/reverberant ratio this way.

---

Helpful tip of the day, I've not heard that one before. It seems easier than standing outside the doorway. Thanks!

Or, like me, you can become deaf in one ear. Just to add to the challenge!

--------------------
www.allyle.co.uk

Also bear in mind that when you listen to a sound in a room, you are making small involuntary head movements all the the time, and these help the brain "learn" and discount the room acoustic. This information is, of course, not available with any microphone.

Quote Hugh Robjohns:

---

Because when you're inthe room the brain has all the information from both ears to be able to work out the directions of reflectionsa nd their riming, and focus the attention on the wanted sound.

---

Some time ago I met a scientist who was working on "otoacoustic emissions". An otoacoustic emission is a "sound" (or more accurately, a signal) that is generated by the eardrum itself, and this signal can either be spontaneous or generated after an external stimulus. Generally speaking, what we hear is the sum (or the "mix") of the external audio sources and the otoacoustic emissions (which are usually very low level signals).
An obvious example of otoacoustic emission occurs when there is a very loud sound: the eardrum gets contracted so it vibrates less, as a protection system. This scientist gave me some other examples of otoacoustic emissions, but I only remember the following one: when we are listening to one sound source in a complex sonic environment, some otoacoustic emissions are generated in order to focus on the one sound source we are listening to (a major difference between us and a microphone is that our brain knows which sound we are trying to listen to). This should explain why we can have a conversation with someone else in a place where lots of other people are talking together, for example in a restaurant. I guess that in this case, the otoacoustic emissions work like a kind of brain-controlled "attenuator" based on "phase cancelling", but saying this is really not scientific and honestly I don't know how it works in practice.

From my intuition only (I don't have any medical nor neurologic background), I tend to think that these otoacoustic emissions work together with some other hearing mechanisms, like space perception (related to HRTFs), cultural factors (habits, civilisation, knowledge, ...), etc.

-j

Quote Bob Moose:

---

Generally speaking, what we hear is the sum (or the "mix") of the external audio sources and the otoacoustic emissions (which are usually very low level signals).

---

That makes it sound like otoacoustic emmissions are an active element of hearing... and I don't believe that is the case. That's not my understanding of it anyway.

The cochlea basically comprises a coiled cone, it's diminishing dimensions ensuring that differnt areas tend to respond to different wavelengths -- that's the basis of our frequency perception. However, the way the cochlea is built, it can change its internal shape slightly which alters the reonsant behaviour and essentially amplifies some frequency ranges to help frequency discrimination.

The audio signal is passed via the ear drum and the three ossicles through to the liquid-filled cochlea, where the actively adjustable resonance of the organ of corti and its nerve cells provides some of the signal discrimination.

Like any physical sound conduction mechanism, sound energy can be reflected and that's what otoacoustic emissions are. Sound energy is passed through to the cochlea and some of it bounces back and can be measured. The better the cochlea is working to enhance the resonance the more energy is reflfected because of the effective gain introduced by the active resonant tuning of the organ of corti.

Measuring otoacoustic emissions is an indication of the health of the ear because it indicates good active control within the cochlea. If the organ of corti is damaged it doesn't tune itself as well (or at all) whcih means less reflected energy and less ability to interpret sound (ie deafness).

Quote:

---

when we are listening to one sound source in a complex sonic environment, some otoacoustic emissions are generated in order to focus on the one sound source we are listening to (a major difference between us and a microphone is that our brain knows which sound we are trying to listen to).

---

Yes, kind of... but you have it arse about face. The organ or corti changes shape in order to provide greater frequency accuity to help differentiate the incoming signal, and because it chanegs shape and its effective gain increases, some of that enetrgy gets reflected back out and can be measured at very low levels at the ear drum. But It's not a reference signal being sent out by the ear, it's a reflection of the energy going into the ear.

Quote:

---

I guess that in this case, the otoacoustic emissions work like a kind of brain-controlled "attenuator" based on "phase cancelling", but saying this is really not scientific and honestly I don't know how it works in practice.

---

Not like that -- that's for sure

hugh

--------------------
Technical Editor, Sound On Sound

Quote TurboD:

---

Sometimes it helps to stick a finger in one ear when miking.

---

This routine also gives you a lot of unwanted credibility with beer-bellied folk singers in sandals.

--------------------
Dynamite with a laser beam...

Nope, we don't hear otoacoustic emissions.

--------------------
my music

Your head/ear system also contributes quite a lot of natural EQing which a microphone doesn't have. If you ever get the chance, try one of the 'head' microphone systems with two small capsules lodged in a generic head/shoulder unit (I think neumann make one) to synthesise the response of a head. This applies a generic HRTF (Head related transfer function, IIRC) to the recorded signal which, when played back through headphones to avoid further room interference, can sound quite uncannily like your own ears in the room. This I believe forms the basis for binaural recordings (these have been used for music too, Tchadd Blake used them for as a stereo pair for drums...). If you put the tiny capsules in your OWN ears, the reproduced effect can be pretty impressively like your own ears!

I believe the principle reason is that a microphone does not have a brain attached to the end of it.

"I believe the principle reason is that a microphone does not have a brain attached to the end of it."

And that^^^I think is the crux of the matter. Like the eye (which evolution stupidly never gave us a second lens for so our poor brain has to invert things!) the ear is largely a data gathering device (barring the feedback system Hugh mentions, a facility I seem to have largely lost as whilst I can still "hear" things, my discrimination is now bugger all!).

This I think is evidenced by the fact that we can learn and improve the use of all our senses with practice, e.g. few people are born with the ability to tune a guitar but most can learn, but few guiarists would ever be able to lay a scale on a piano!

Dave.

Exactly, as a philosopher I see it this way: all technology is an extension of the human body & mind: Software: mind, hard ware: body. Microphone: body, DAW: mind. Need I say more?

Quote ElecTrika-MixTek:

---

I believe the principle reason is that a microphone does not have a brain attached to the end of it.

---

That is the case of many a human being

Quote Bob Moose:

---

An obvious example of otoacoustic emission occurs when there is a very loud sound: the eardrum gets contracted so it vibrates less, as a protection system.

---

No. This is called the stapedian reflex, some kind of built-in acoustic limiter. Now, it is possible that the contraction of this tiny muscle produces some sound; that would be the equivalent of the thump produced by control-voltage feedthrough in a hardware limiter. But it would be a side effect.
It looks like some OAE's (there's at least three different types of OAE's) have a specific purpose, similar to dithering, but at the moment the understanding is very vague. But the observation of OAE's has a great medical value, since their measurement is used to evaluate the hearing process of babies.

Quote Giscol:

---

Your head/ear system also contributes quite a lot of natural EQing which a microphone doesn't have. If you ever get the chance, try one of the 'head' microphone systems with two small capsules lodged in a generic head/shoulder unit (I think neumann make one) to synthesise the response of a head. This applies a generic HRTF (Head related transfer function, IIRC) to the recorded signal which, when played back through headphones to avoid further room interference, can sound quite uncannily like your own ears in the room. This I believe forms the basis for binaural recordings (these have been used for music too, Tchadd Blake used them for as a stereo pair for drums...). If you put the tiny capsules in your OWN ears, the reproduced effect can be pretty impressively like your own ears!

---

On the other hand, the whole HRTF can be most easily supplied by a listener's own head and ears while listening to loudspeakers. Ideally (to some), a wide bandwidth coincident mic array captures and recreates something similar via loudspeakers to the point in real life where a listener's head would be. The incorporation of HRTF effects-including simple near coincidental mic arrays- in a recording intended for loudspeaker listening does the whole HRTF thing twice, it seems to me.

That's a gross over simplification, of course, but does raise some perceptual quesstions about using coincident vs non coincident microphone arrays for loudspeaker listening, assuming the engineer is trying to produce a realistic listening experience as opposed to producing a good sounding recording-not that the two are exclusive! Is the recording a documentary or an involving media creation or interpretation of an event?

I'm not proclaiming coincidental mic arrays as the bees knees, in fact I often prefer methods which hint at greater depth.

We hear what we see (and feel), so the memory of the live event that we compare recordings to may not be so accurate.

Also, in real life, a listener is usually positioned far enough away from an "ensemble," that there isn't much front to back depth. In recordings, the mic's are often closer than listeners would be, so the relative distances of sources (performers) to microphones are much more varied which can produce more depth in a recording than might occur in real life. This assumes that some sort of ambient mic technique is employed, rather than a bunch of spots panned and processed. If a recording duplicated what an audience member heard at a performance, the recording may be perceived as one devoid of depth and greatly HF attenuated.

This thread has impressed me with how much I don't know about ear physiology-time to do some reading!

The math of Mr Gerzon et al is well beyond me but would we not require mic arrays in the vertical plane to resolve depth( and two additional tracks)?

I know image "depth" is often talked of in monitor reviews but is it not in fact illusory?

Dave.

Quote ef37a:

---

The math of Mr Gerzon et al is well beyond me but would we not require mic arrays in the vertical plane to resolve depth( and two additional tracks)?

---

This is an issue I don't understand, and I haven't yet parsed Mr. Gerzon's writings either. Microphones have "3D" patterns, symmetrical or not. And mic's are often carefully pointed to include or exclude certain wall, ceiling and floor effects. Speakers are often located on the same horizontal plane. Perhaps Hugh can enlighten us. I haven't used a Soundfield like system, but some recordings I've heard from them have a certain "rightness" about them that other recordings seem not to have.

Quote ef37a:

---

I know image "depth" is often talked of in monitor reviews but is it not in fact illusory?

---

It has never seemed illusory to me. Amplifiers are part of the equation, too. Speaker loads are complex, so these effects may simply be the match of a specific speaker with a specific amp.

The SOS reviewers seem to have no trouble discerning differences in depth of sound stage. Maybe one of them might have something to say.....

Hi Jeraldo (you've worn well!)

I asked the question in all humility and ignorance!

A skilled artist can put the illusion of depth in a painting but there is no actual spatial information there, you cannot walk around the image.

People have made claims of hearing directional effects from mono recordings, if they say so, they do, fine with me but there is no such information in the signal.

Dave.

Quote ef37a:

---

A skilled artist can put the illusion of depth in a painting but there is no actual spatial information there, you cannot walk around the image.

---

You cannot walk around the image, but spatial information is there. How else would an illusion be created?

Quote ef37a:

---

People have made claims of hearing directional effects from mono recordings, if they say so, they do, fine with me but there is no such information in the signal.

---

I haven't heard these claims, and I've never experienced such directional effects.

Mono recordings can have depth information. For example, if there were two sources of sound that had some distance between them, and a mic was located very near one of the sources. Some structure reflections might help, or not.

If a mono mic was positioned in front of an orchestra and quite close to the orchestra, I don't think you'd have any problem determining which instruments nearest the mic, and which were positioned at various places toward the rear of the stage. Again, some reflections of walls or ceiling (or shell) would contribute to the effect.

No, Sorry J,
There is no 3D information in a 2D picture. It IS an illusion.

Way back in the daft days of Quadraphonics someone proposed the J.O.K.E. 4 channel matrix system.

The whole thing was a spoof to prove that you cannot encode 4 channels onto a 2 track system. At least that is how I recall the article in HF News.

Dave.

You can encode four channels onto a two-channel system. You just can't extract the four channels with total independence again -- inherently there will be a large degree of crosstalk. That can be disguised by various cunning means (see Mr Dobley's clever 'active steering' technology as used in Pro-Logic), but it will never provide truely 'discrete' channels.

Any sound presented over loudspeakers -- whether one, two, or many -- is portraying an illusion of real life that most people find acceptable. Just as TV, film, photography, paintings, sculptures and books all do.

We willingly suspend disbelief in order to 'enjoy' the experience and our minds allow us that priviledge.

An impression of depth can be portrayed over a single speaker by using a combination of tonal and level changes combined with early reflection patterns which suggest relative distances within an acoustic space. Two speaker stereo does exactly the same, but with the additional benefit of an impression of left-right width or spatialness as well as front-back.

There is no height element and no way of encoding height reliably in two-channel stereo. The use of specific comb filtering tricks to replicate some of the spatial localisation capability of the pinna (outer ears) can create an impression of height, but it tends not to be reliable -- by which I mean different people interpret the processing as giving very different positions up/down, left/right and even front/back. And room reflections can bugger it up completely.

However, the effectiveness and 'realism' of stereo generally is determined by many things including the design and quality of the speakers and the listening environment.

A lot of speakers don't generate stereo images very well, and anything that adds early reflections in the listening environment can destroy the illusion very quickly too. This makes it a challenge to generate decent stereo in a project studio with computer screens and other junk cluttering up the space between the monitors, for example.

Adding height to the party helps to make the illusion even more convincing (if done well), and there are various systems around offering that facility -- although I doubt any will become domestically acceptable in the foreseeable future.

The SoundField mic solution (especially the new models using digital matrixing algorithms) work better than anything else principally because of the very high degree of coincidence in all three planes.

Spaced mic stereo is inherently less accurate (spatially) than coincident stereo, but often the extra time differences captured make for a more subjectively pleasing effect.

It's a bit like comparing a waterclour painting (spaced mic approach) to a photograph (coincident approach). They both represent the same scene in different ways, and depending on what you want from the experience, one will be more enjoyable than the other, but neither is inhererently better or worse.

hugh


--------------------
Technical Editor, Sound On Sound

'S right Hugh,
All smoke amd mirrors but we enjoy it!

Dave.

There is no 3D discrete info in a picture. There is no discrete depth info in a mono recording. But, just as the auditory system has the capability of figuring out rather accurately the depth position of an instrument in a mono mix (or to be fooled by artificial reverb and addition of time-related FX). Stereo offers a discrete additional width information.

Quote lead ears:

---

There is no 3D discrete info in a picture. There is no discrete depth info in a mono recording.

---

There is relative depth information in a mono recording, depending on the circumstances of the recording. A "picture" most often contains perspective information.

Quote lead ears:

---

But, just as the auditory system has the capability of figuring out rather accurately the depth position of an instrument in a mono mix

---

You can't have it both ways. If the "auditory auditory system has [that] capability," then there's got to be information for the system to process.

A "mono recording" is not the same as a "mono mix;" you can't use the terms interchangeably, as you may have done.

Lets call "depth" distance information.

On a stage, in a room, whatever- Using a single microphone, a source closer to the microphone will have relatively higher SPL's and have relatively fewer reflections than a more distant source. The more distant source may have attenuated HF.

This impression will be increased by a relatively close proximity of the mic to the nearest source, greater distance between sources, and the timing, number, and amplitude of reflections from each source relative to the other.

As to a "picture," all competent photographers know how to manipulate depth perception by the thoughtful use of long v short lenses-or put differently-changing perspective by moving the camera. (Not to be confused with the photographic term: "depth of field".)

Quote Jeraldo:

---

Quote lead ears:

---

There is no 3D discrete info in a picture. There is no discrete depth info in a mono recording.

---

There is relative depth information in a mono recording, depending on the circumstances of the recording. A "picture" most often contains perspective information.

Quote lead ears:

---

But, just as the auditory system has the capability of figuring out rather accurately the depth position of an instrument in a mono mix

---

You can't have it both ways. If the "auditory auditory system has [that] capability," then there's got to be information for the system to process.

A "mono recording" is not the same as a "mono mix;" you can't use the terms interchangeably, as you may have done.

Lets call "depth" distance information.

On a stage, in a room, whatever- Using a single microphone, a source closer to the microphone will have relatively higher SPL's and have relatively fewer reflections than a more distant source. The more distant source may have attenuated HF.

This impression will be increased by a relatively close proximity of the mic to the nearest source, greater distance between sources, and the timing, number, and amplitude of reflections from each source relative to the other.

As to a "picture," all competent photographers know how to manipulate depth perception by the thoughtful use of long v short lenses-or put differently-changing perspective by moving the camera. (Not to be confused with the photographic term: "depth of field".)

---

Language problem? You're saying the same as me with different words.
I put discrete in bold characters for a reason. In a mono signal (let's not debate about recording vs. mix), depth information is embedded (you say relative).

Quote lead ears:

---

In a mono signal (let's not debate about recording vs. mix), depth information is embedded (you say relative).

---

The depth information, such as it is, is embedded in any recorded (or mixed) signal, surely, whether mono or stereo?

I don't think I understand the point you're trying to make. Sorry.

Hugh

--------------------
Technical Editor, Sound On Sound

Quote ef37a:

---

No, Sorry J,
There is no 3D information in a 2D picture. It IS an illusion.

---

I think you have it confused. Your real life view of an oil barrel is not the barrel. It's your view of it. It's the same with a photo. Much of the same data is processed the same way. The Camera Obscura informed painters in the ways humans see. The Camera Obscura identified and quantified data. Not the reverse. Not magic, either.

A little experiment: Next time you're buying a house or renting a flat, take your own camera along. Take pictures of everything. I've found I see all kinds of things in the photos that I never saw "in real life." Even things that I definitely saw differently than the photos show. Which is "more accurate," the flawed attention and memory of the renter/buyer or the "illusion" of your own photograph?

You're in court. Is the best evidence a photograph, movie, audio recording, or someone's eyewitness account or memory?

Quote ef37a:

---

Way back in the daft days of Quadraphonics someone proposed the J.O.K.E. 4 channel matrix system.

The whole thing was a spoof to prove that you cannot encode 4 channels onto a 2 track system. At least that is how I recall the article in HF News.

---

They were wrong. It would have been equally logical to posit that you could not present a solid phantom mono image from a recording and playback system that was stereo from beginning to end.

Long after Quadraph(e)onia, there was a very simple passive scheme to derive surround information from broadcast stereo for video, which did not simply move L and R to rear. It perhaps worked better than it should have because a lot of attention at the time was being devoted to more fully exploiting stereo playback.

So the issue becomes semantic: is the question of the ability to encode more than two channels into a stereo format a valid question, when the entire method of audio for video changes, allowing for effects that couldn't have been accomplished even with a simple discreet four channel system to be done with stereo-not to mention stereo/surround?

On to more concrete subjects:

There are (were) the famous(ly expensive) K boxes for spatial imaging *from* stereo sources. These expensive boxes did not perform illusions and magic, rather they manipulated and supplied data for the perceptual systems of humans.

And then that Neve *Stereo Field Editor* recently reviewed in SOS. In addition to the predicable Side Signal amplitude and EQ variations, it offers a center depth control that I am told operates by phase rotation (of either S or M-not sure). (One for Hugh to explain. The reviewer didn't.)


But there's more: recall that Shuffler Circuits were/are intended (but generally aren't) to be used with various arrays, such as the Jecklin Disk.

And that "Spatial Equalization" (also done by the Neve) is a common software and hardware production tools.



My point, finally, is that this entire discussion seems to be resting on semantics.

Is the question this: Are all of the phenomena discussed in this thread "illusory" or are they "real"?

What exactly would be the difference between a real oil drum and a completely convincing representation, other than the representation may (or may not) be easier to walk around?

Is this not how human perception works? And that does seem to be the question.

What is the difference between being "fooled" by an illusion, or experiencing a "real" object or phenomena with the same data?

There is no difference.

Blumlein was not a magician. Neither were the Camera Obscura people. Gerzon doesn't seem like a illusionist, nor does Mr. Neve.

It's the science of perception, not illusion.

Quote Hugh Robjohns:

---

Quote lead ears:

---

In a mono signal (let's not debate about recording vs. mix), depth information is embedded (you say relative).

---

The depth information, such as it is, is embedded in any recorded (or mixed) signal, surely, whether mono or stereo?

I don't think I understand the point you're trying to make. Sorry.

Hugh

---

That was in reference to "There is no 3D information in a 2D picture. It IS an illusion." by EF37. He calls it illusion, we call it "embedded".

"That was in reference to "There is no 3D information in a 2D picture. It IS an illusion." by EF37. He calls it illusion, we call it "embedded"."

Call "it" what you like, still not there!

Dave.

Quote ef37a:

---

"That was in reference to "There is no 3D information in a 2D picture. It IS an illusion." by EF37. He calls it illusion, we call it "embedded"."

Call "it" what you like, still not there!

Dave.

---

Listen to "rock around the clock" by Bill Haley, and tell me if the snare is at the same distance as the rhythm guitar...a mono recording. Although nobody can tell exactly at which distance, everybody hears it coming from behind. Could it be collective hallucination?

" Could it be collective hallucination?"

Or could it be that "everybody" has seen the film?

That sound reproduction is a monstrous illusion is perhaps best demonstrated by the "phantom mono image" effect. Try as we might it is almost immpossible to discern two spaced identical phase coherent signals, we hear just one centrally placed signal.

THAT is not physical reality! We can measure the signals to the speakers and the near field SPLs but at the sweet spot our brain refuses that information and puts the signal plumb centre (this even works for me and my right ear is at least 25dB down at 2kHz and beyond compared with my left. Don't ask me how!).

At this stage of technology we cannot duplicate how real sound sources radiate (I remember a paper on a dummy head for speech reproduction in an AES journal but that is but one, very limited result). The best we might hope for is some kind of acoustic hologram?

The various headphone systems mentioned I cannot comment upon, not being in a position to check them out. I would assume they use some very clever techniquies akin to dummy head recording? I have dabbled with this and the results are scarily lifelike, 360dgr reproduction and yes, a very accurate impression of distance (no vertical component as I recall but since we evolved on the plains of Africa we would not need height resolution).

Our brains ARE easily fooled, the concave face, the "candlestick" picture, colour telly (only 3 primaries you know!).

Dave.

Quote ef37a:

---

" Could it be collective hallucination?"

Or could it be that "everybody" has seen the film?

---

I should think you know the images in the film have nothing in common with the actual stage where the track was recorded (the Pythian temple in New-York).

Quote:

---

That sound reproduction is a monstrous illusion is perhaps best demonstrated by the "phantom mono image" effect. Try as we might it is almost immpossible to discern two spaced identical phase coherent signals, we hear just one centrally placed signal.

---

I really think it's a language problem.
For all purposes, these illusions are so repeatable and so consistent from person to person, I consider them as useful tools.

Quote:

---

THAT is not physical reality! We can measure the signals to the speakers and the near field SPLs but at the sweet spot our brain refuses that information and puts the signal plumb centre (this even works for me and my right ear is at least 25dB down at 2kHz and beyond compared with my left. Don't ask me how!).

---

You seem to neglect the fact that these two speakers manage to produce a sound pressure at the listening point that is not too dissimilar to that that would be produced by a single center speaker. We don't hear sources , we hear the sound pressure these sources generate at our listening position.

Quote:

---

At this stage of technology we cannot duplicate how real sound sources radiate (I remember a paper on a dummy head for speech reproduction in an AES journal but that is but one, very limited result). The best we might hope for is some kind of acoustic hologram?

---

It's been proved since long that it is not necessary to recreate the whole map of the soundfield to render a believable auditory sensation. I'm not meaning accurate, I just say it's a coherent image that allows positive (i.e. all listeners have convergent experiences) evaluation of the positions of the different sources.

Quote:

---

Our brains ARE easily fooled, colour telly (only 3 primaries you know!).

---

Are you suggesting that creating a colour by mixing those primary colours is also an illusion?