Hi,

Over the last couple of years I've been doing a little research into binaural audio. I've also encountered holophonic audio. I understand that holophonic audio is a trademarked technology invented by Hugo Zuccarelli. To my ears
the holophonic audio seems even more realistic than the binaural audio I've heard.

My question is - is there anyway currently to produce holophonic audio. I know that there is the BS-3D from THE but I'm not sure that's what I'm looking for. A binaural dummy head is an option but it is not "holophonic"

After hearing a good deal of both technologies I've come to the conclusion that holophonic sounds far more "3D" to my ears than binaural (although there are some spectacular binaural presentations as well - see the ZBS production of "The Mist". How can one produce holophonic sound? Is there a dummy head or mic that can be purchased? I haven't seen anything available.

Firstly, not all binaural techniques are the same; some rely upon a mannequin head with ears (and some of these have shoulders, and still otehrs incorporate a torso), some use a sphere (BS-3D, Schoeps), some are worn against a person's head, some use microphones placed inside the user's ears. Each has benefits and drawbacks, however, here is what I know about binaural mannequins. There are several types of binaural mannequin heads that you will likely come across as you reasearch the threads. The ones that I know of are as follows:

* AKG (I forget the model number), circa 1974. Tough to find, but fairly inexpensive. Not very sensitive (2.0 mV/Pa if I remember) and equalized (I think) for free field response. Can probably find one for less than $1,000. Ears are affixed to the head, and I believe the transducers are dynamic. Some have wirtten favorably about it, and some not so favorably (on the GS forums), but I have never personally heard a recording from one.

* Neumann type KU-100: Probably the most common mannequin head out there. MSRP $9,999 or thereabouts. You will find them at 'discount' for around $7,999 new. You can also find them cheaper if you find them used, but expect to pay (most likely) around $6,000 used - there are not a lot of them in circulation, so supply and demand holds sway.

Sensitivity is 20 mV/Pa (and noise is, I believe somewhere around 15 dB(A)) and unlike the KU-80, is diffuse-field equalized (not free field as was the KU-80). This equalization is (IMHO) a pretty important detail. Note: The KU-80 and KU-81 preceded the KU-100. As I recall, the KU-81 was much like the KU-80 except that unlike the KU-80, the KU-81 was diffuse-field equalized. The ears in the KU-100are removable and the microphones allow for a pistonphone or other calibrator to be used to generate reference tones and thus, calibrated recordings. I am pretty sure that this mannequin head was used for "The Mist" though I am not 100% sure on that one.

Note: I checked on the web and indeed confirmed that what you heard in "The Mist" was in fact a Nuemann type KU 100. I also found this about ZBS' binaural work ["ZBS nicknamed their Neumann head "Fritz" and find that he picks up sounds "so perfectly that you'll be convinced that those heavy shuffling footsteps, demented little chuckles and wheezing gasps are actually six inches from the nape of your neck." They also report that "to get the full effect: the subtleties, nuances and all the three- dimensional As-Though- You-Were-There wowie-zowies -- headphones are best!"]

* Bruel & Kjaer type 4100. This is a head + torso variant. Mics require a special B&K power supply, but the supply allows the user to switch between free field and diffuse field equalization. I think this will set you back about $14k by the time the mic preamps / EQ is added. Removable ears (like the Neumann) and (if I am not mistaken) a 50 mV/Pa sensitivity.

* KEMAR (Knowles Electro-Acoustic Mannequin for Acoustic Research). I truly don't know where you get these, but people on the forum have used them. This is a torso + head like the B&K type 4100, and also like the type 4100, is used a LOT in binaural reaearch. I believe that the microphones in its ears are from G.R.A.S. and as such, are probably 50 mV/Pa. I truly don't know the other details.

* Head Acoustics: Probably the 'cadillac' (or BMW or Mercedes or Ferrari or what have you) of mannequin heads. Comprises a mannequin + shoulders, but no torso, so in this respect, it is more like the Neumann than it is the B&K or the KEMAR. Pricey - probably around $17k for a stand-alone head with SDRAM recording. One thing that Head Acoustics wares have is a third EQ (switchable) scheme called ID (Independent of Direction) as well as free-field and (I think) diffuse field.

* CORTEX. Like the Aachen and B&K variants, this one has been typically used in research work. It's been years since I have seen one, but I assume that they are still current offerings from them. I believe that its sesitivity is also around 50 mV/Pa. If I were to hazard a guess as to cost, I would guess around $20,000.

Apart from that, there are some DIY heads out there and some people have built them and like them. I have no direct experience, but fundamentally, I think you could probably get much of binaural's benefit by taking this approach. However, there are some tricky aspects to this (especially when it comes to placing the mics in the ear canals). So, that's the down-side (the 'players' have all the tooling and processes to make binaural heads, one after the next, that essentially offer the same performance as the previous ones.

Note 2: There is a lot of debate out there as to what is the "correct" binaural approach (and indeed, what constitues "binaural" as well), so I would hazard a guess that some of what you have heard isn't necessarily true binaural. By that I mean that in the research community, only a binaural head mannequin is considered binaural. However, in the recording world, many different variants on binaural that do not use a mannequin head are called 'binaural' (by some). So, if you are comparing recordings, keep in mind this one fact, quite literally that, apart from the research community, not all binaural recording techniques are equal.

Then, you have the wearable variants. Sometimes these are simply omni's on eyeglasses, sometimes they are omni's near the ears, and sometimes they are omnis designed to be placed in a person's ears.

If you check the Binaural threads, you will find a lot of information. Also, if you check the thread "What is your favorite binaural recording" you can hear examples of just about all the types of mannequins and worn binaural sets that I have discussed.

I also started a thread specifically about how binaural can be advanced, and that thread is entitled "Best Bets For Binaural". Oh, now that I think about it, one of the members chimed in on that thread about his variation on the 'binaural theme' and if you cruise that thread, you will see his links to his own creations.

My advice? Before you give binaural short shrift, check out some of the examples that have been posted. I think that you may be surprised at the realism in some of them.

All the best... Mark

Thanks for the excellent post. I've been going through some of your other posts and I'd like you to know that the time you take in educating other people about binaural audio is really appreciated

I'm trying to decide what to implement at the moment. I will be recording audio drama and spoken word only. I grew up listening to audio drama and radio shows but even though there are some wonderful examples of binaural audio drama out there, it seems like it is a field that has not been exploited to the fullest yet. ZBS has done some wonderful things especially their drama called "Sticks" based on a Karl Edward Wagner short story. It is one of the creepiest things I've ever heard.

I'm pretty familiar with the sound of the ku81 and ku100 but since those are out of my budget I'm considering alternatives. One is the bs-3d. I've got some really thoughtful responses from the guys at T.H.E It is a really good sounding mic. They have taken over development of the holophonics system from what I've heard and read.

One thing that impressed me about the holophonic recordings that I've heard is the amount of frontal cues seems to be greater. I'm able to differentiate between back and front more easily. I've heard some say that this is because we tend to process things visually in front and that affects how we hear things. I've also heard some say that it is because our heads are all shaped differently and we will all hear different cues based on how close the mic matches our head. Either way, the holophonics recordings that I've heard seem more realistic to me.

I'm also contemplating doing the cheap mannequin head with church or core sounds omnis. Most of the things I know about making those came from this site. DigDagga's Binaural Dummy-Head. The guy that does the Bird Vs. Alligator stuff is very talented and has written a really good instructional guide. He also sells the silicone ears.

I'm sure it will not be perfect but it might be a good introduction. A good used ku-81 is also an option but the price is steadily rising if you can find them.

Tut: Thanks for the kind words. I do try to educate whenever I post (well, almost always), so it's nice to hear positive feedback about my posts.

I would suggest that, if possible, you read the AES paper 6704 entitled "The effect of audio compression techniques on binaural audio" written by Fabien Prezat and Brian FG Katz.

Anyway...there is some very interesting stuff in there not only in terms of what compression algorithms seem to affect binaural's spatial cues, but what seems to govern these spatial cues - also with regard to the front-to-back cues. Much of this is known and has been written about, but the paper references several other pivotal research works. The paper also spoke briefly about sonic realism, and in the words of Messrs Prezat and Katz, here is their take about 3D sound:

"Various methods exist for the reproduction of spatial audio. These methods include the popularized Surround Sound 5.1, Ambisonics, wave field synthesis, and binaural. Of these methods, binaural audio easily allows for full and precise 3D reproduction using either recorded or synthesized audio. For this reason, this study focuses solely on binaural audio rendering."

I know what I am about to say may anger some here (in G.S. proper) but I truly believe that such perceptual matters are best left to the educational institutions or at the least, those microphone companies working in concert with research organizations. To that end, all of the wares used to ascertain such cues have been binaural mannequins - some with and some without torsos, but always binaural mannequin heads. I have looked over Zuccarelli's site and find it very interesting, but I don't seem to find his recordings more convincing than those that I have heard using a binaural mannequin.

Of course (and I admit), that this may be my biases getting in the way, and I should probably have some more listens to the materials. Still, as an engineer who has worked in signal processing for the last 20 years, I can say that from the point of view of spatial accuracy in research / human perception of sound, the binaural mannequin has become a de-facto standard in the research community. Personally, it makes sense to me because it replicates, better than anything else that I have heard, the boundary conditions of the human head, and thus, best replicates (in a recording) that which we hear. With regrad to perceived realism, this is only my opinion though - again, my opinion, no more, no less, though many, many others in human perceptual research share this opinion.

Anyway, binaural has been entrenched in the research community for many years now, and there is a vast body of work to suggest what are the most important elements of localization (the 'obvious' ones are the inter-aural delay (time difference between ears), inter-ear level difference (magnitude and spectral differences borne of shadowing / pinna etc) and of course, the HRTF - you could argue theat the shadowing is part of the HRTF.

My personal opinion on the 'dagga' DIY mannequin head is that he has the right idea. I advised someone else that he should try the project because provided he starts with some 'real' microphones (he was planning to use some DPA) that the rest is likely to go smoothly. Also, the ears he sells, while I don't have the details, are much cheaper than are the replacement ears from Neumann (which are around $500 / pair). I know that's a lot of money, but like everything in life, small-volume items cost more money (and so too does their maintenance).

I see no reason why with the proper data acquisition system and signal processing tools (and some test environments) that the response of a home-brewed mannequin could not be made into a diffuse-field equalized mannequin...or a free-field. Granted, given this it seems moot to worry about the frequency response of the microphone, but I believe in starting (any project) with the best critical-path element possible, and in a mannequin head, it's the transducers that are the critical-path items. He speaks to this (equalization)somewhat on his site, but there are some things to be mindful of that he doesn't fully cover, but on the otehr hand, I appluad him wholly for his illuminating promotion of DIY binaural.

As an aside, I wrote to the people at DPA about them making their own binaural mannequin. See, their progenitor is Bruel & Kjaer, who make the type 4100. I had suggested that they could possibly 'piggyback' that model and make a version for 'recording' rather than 'acoustic acquisition and analysis'. They were not too keen on the idea as I think they have a lot riding on their 5.1 microphone that they now market. I understand that - 5.1 is more 'marketable' than is binaural due to its (now) ubiquitous nature.

So this begs a question...what would you deem 'affordable' for a mannequin that delivered performance along the lines of a KU 100 with (perhaps) a few less features? You know, a sort of bare-bones model? I personally think that DPA is missing the boat here, but again, they probably see 5.1 as more 'marketable' than they do binaural.

Again, this may cause controversy, but IMHO, if you wanted to truly know what technique rendered the greatest realism, then then only way to do so (for options greater than one, i.e. three types of sound) and have any statistical confidence in the process would be do conduct double-blind juried studies presenting the same sound(s) as recorded using the various systems. That is (for example), binaural head(s) vs. Holographic 3-D. I wrote a diatribe in another post somewhere (and was derided by one reader for being a 'scientist' - I'm so ashamed of my engineering credentials...) about this juried testing, so I won't re-hash it. If you search on the keywords Bradley-Terry ot Bradlet-Terry-Luce you should be able to find it.

Once again, excellent post. Really thought provoking. I'm interested in the technical aspects of binaural recording as well (even if I may not fully understand some of it)

I read this about a new system that T.H.E. is doing. I haven't gotten a chance to ask about it but

The company says that it will launch the first commercially available Holophonics™ 3-D recording system in late 2010, in versions specially designed for the film and video industries, including a digital wireless system for field use.

This should be a unique product when it comes out. As far as affordability goes I would think that something along the lines of what the bs-3d costs ($2500) would be reasonable. Keep in mind that I am strictly amateur but I have spent more on less worthy pursuits in the past.

There definitely seems to be a lot of buzz going on about binaural at least. There have been a lot of good posts of late.

If the Holophonic system is 'priced right' then my guess is, at that price, they will sell a ton of them. Unique? Meh...

Granted, wireless, noise-free transmission would be very handy, no doubt.

I do have some doubts as to the validity of the holophonics 'argument' as put forth by Zuccarelli. That is, as I understand it, the argument is that we sense sound not by pressure perturbation that reach our ears and then excite the basilar membrane, but instead, that our ears actually emit sound, and we sense direction and such by an interference pattern that the incident sound makes with the sound our ears emit. That is, the argument (I hope that I am getting this right as it is all from memory) is that our ears emit a reference sort of tone / noise, and this interaction between that and the sound incident is how we localize sound.

Hmmm...I struggle with this, because this claim was laid back in the early 1980's, and I would think if it had any real validity, it would have thoroughly been researched by now as there are several universities that do tons of research on human auditory perception. To my knowledge, no other instutution has proven the hypothesis that is behind Holophonic sound. However, since the 1970s' and early 1980's, binaural (mennequin) based recording / analysis has become the cornerstone of pretty much all perceptual research. M.I.T., Michigan State University, The University of Aachen, The Institute of Sound and Vibration (ISVR) et al (and countless others) are all haevily involved (and have been) for years basing their work on sounds as perceived by binaural mannequins. Since the goal of such reasearch is to understand (among otehr things) how we localize sound, the fact that binaural mannequins are used is s a rather salient point. As far as I know, none of the research universities / institutions have based much (if any) work on Holophonics as envisioned by Zuccarelli.

As an engineer, it always troubles me when a hypothesis can never be proven by another scientist. I really need to research for more published works, especilly those that would have been published in peer-reviewed technical journals. That would give me a much better feeling about the technical merit and accuracy of what Zuccarelli et al have postulated.

See, I have read that ehe effect that the Holophonics system achieves is comparable to traditional binaural recordings made using mannequin heads. In other words, they sound the same as / similar to recordings that incorporate the HRTF and ILD (read: binaural) becuase they too incorporate them.

However, to my knowledge (based on technical papers that I have read in trustworthy journals and such) there is little, if any real, hard data or studies to support the claims of Holophonics that would suggest it's fundamentally different from binaural recording.

UPDATE: There was a European patent filed by Zuccarelli, but I have not yet been able to get a full copy of it - only bits and pieces of it in articles and the like. However, what I have been able to find about the holophonics system as described in the patent is that it comprises a mannequin head, with artificial ears, and microphones in the ear canals. So far what we have described is... a binaural mannequin head. Additionally, the patent describes how frequency content is is eqialized to compensate for the sound passing through the ear canal. Again, this is part of the various equalization schemes used by the likes of Bruel & Kajer, Neumann, Cortex, and Head Acoustics; all offer some flavor of this equalization (as I have covered in other posts about binaural manneqin heads).

One thing that really troubles me (though I need to verify this) is that as far as I know, the details of the technique have never been published by Zuccarelli et al. I don't think he's allowed access to his research either.

I know...you're now saying "You're a real spoiled sport...the guy has something revolutionary and wants to protect it". OK, while in principle I agree, I also know that every 'novel' invention that passes scrutiny generally get patents awarded (I've been through the process myself three times now, though non of my patents are for recording hardware, two are related to acoustics). Seriously, if I were in his shoes, if if were the real deal, I would have had a tem of patent attornies documenting everything and would have announced it to great fanfare, and referenced my patents.

I could be dead-wrong here. It is entirely possible that Dr. Zuccarelli is incredibly gifted (I'm sure he is), and that he has managed to single-handedly revolutionize the field of perceptual studies. It is possible, I admit that. But...I also know that there are decades of work carried out over genereations of humanity by researchers who have done their work at various research institutions all over the world, and not one of them discovered what Zuccarelli has purportedly discovered. Or, it is also conceivable that what Holophincs amounts to is a re-labeled version of something that amounts to binaural...but with a more marketable name; everyone knows what a hologram is (well, almost everyone), so 'holographic' tends to transmit the message better than does 'binaural'. INdeed, it is remarkably common that people look at me like "what planet do you call home?" the first time they hear the word 'binaural'. To a science-minded type, it may convey three dimensional sound, but 'hologram' instantly conveys the concept of something that appears to be floating in space, having three dimensions, yet representing an illusion to us. The word 'binaural' does not immediately convey that.

All that I have read, save for that which comes from the Holophonic camp is that like binaural recording, Holophonics (re)produces the inter-ear (or interaural differences - this is the arrival time between the ears) and the shadowing effect and the HRTF. In essence, these are all of the things that binaural accomplishes.

One thing that makes me wonder...if the claims are indeed true, what is the reference noise / tone that the system emits into the transducers? My guess is that would be deemed 'proprietary' by them, but even so, what one would need to know is what that tone / noise is, and how to reproduce it. Would it be the same for all people?

I really should get someone I know with a very sensual voice to whisper close to the binaural mannequin and post that sound and see what people think.

Anyway, IMHO, if Holophonics is deemed to sound 'more accurate' than does binaural, I would like to see this proven using a double-blind juried listening software package. I would believe the results of such a test as it is a controlled, repeatable evaluation of sounds.

Sorry, but having done a lot of listening tests over the years, I happen to know how easy it is to get the wrong answer unless you are meticulous in construction of the sounds to be played, and a host of otehr factors. If you want a taste of what are the things that are de rigeur in such activities, look into Bradley-Terry (or Bradley-Terry Luce) as well as the method of paired comparisons as well as the method of circular triads.

I had the opportunity to speak with Christian Hohendahl who is the designer of the BS-3D and we discussed it and the new 3D Holophonic system they are coming out with. He seemed really exciting about both but especially the new product. There should be a lot of buzz when it is released. It's price will be considerably more than I can spend but it's meant for high end users anyway.

I think I will continue to evolve my homemade dummy head and when I hear some other clips of the bs-3d evaluate my situation then.

"Wouldn't it be Nice" if manufacturers of binaural / 3-D / Holophonic wares would actually publish things like the impulse responses? Of course, if they did that, having the impulse response of one would allow a person to morph the signal into the other (via convolution).

Honestly though, I would settle for a measurement 'event' of sorts. That is, place each device in a 'suitable' location in a 'suitable' space and excite each with a 'suitable' noise source and compute the complex frequency response functions (FRF) between the noise source reference microphone and each 'ear' (or channel) so that the results can be compared.

This would be a good starting point, and the nice thing about the FRF-based approach is that, unlike a power spectrum alone, the FRF approach would normalize all the measurements to that of the source - very handy. That is, the source would not need to have a flat frequency response (reasonably flat would do) because the complex FRF would take into account the source (if you prefer, think of the FRF as being derived by dividing the ear (or similar) transducer response by the response of the noise source). Note: by 'complex' I mean containing the real and imaginary portions, or in anoterh format, magnitude and phase.

The FRFs of each type of device could then be overlaid and thus, any deviations from type to type would be clearly illustrated. Measurement-wise, this is extremely simple and could be done with (it is my opinion) a high degree of both repeatability as well as reproducability. I know it's easy because I have done many measurements involving FRFs over the years for my work in noise control, sound quality, and sound synthesis, and there's really not much to it.

However, I doubt that manufacturers would knowingly loan their gear to just anyone (I would think a university would have to be involved) for such an undertaking. As I see it, they are equally concerned with something in the 'competition' being present and perceived as 'better' than what they have.

Anyway, the BS-3D is probably pretty close to a mannequin in some respects and under some boundary conditions, but my expereince in acoustics and with binaural tells me that this would be so in environments where the primary factor is the shadowing (basically, the inter-ear level difference or ILD) piece of the binaural puzzle. Keep in mind that the sphere (BS-3D and the Schoeps variant) shape approximtaes some aspects of a true binaural mannequin head, but it lacks the ears and the ear canal(s). Please note that I am not saying that the BS-3D is a bad-sounding microphone, I am merely saying that it lacks all of the elements of the binaural head to be mathematically correct in order to truly be binaural - but this is true of other quasi-binaural microphones as well. The fact that they are asking somewhere around $2,700 US for the BS-3D unit, makes it way more affordable than a new KU 100 (roughly $8k US).

The funny thing is this...If a company wanted to sell a binaural mannequin for $2,500, don't you think they could do it? Seriously, if the 'dagga' DIY head posts prove anything it's that a decent binaural mannequin can be made for far less than the cost of a KU 100 etc; it need not be research-grade. With some minor upgrades to the mic elements and the head forms I'll bet you could get something that, in practical terms, is very close to a KU 100 or a B&K type 4100.

If Neumann, B&K, Schoeps, DPA et al really wanted to make a 'recording only / non research-grade' mannequin head for $2,500, you just know that they could. However, none appear to be interested .

Ironically, if these companies sold them as 'loss leaders' they would do much to proliferate the technology of binaural, but then again, there is not much room in budgets for things that cost more to make than what they can seel them for. In a way, it's like the rotary engine for Mazda - it doesn't make 'economic sense' to keep making those engines, but the rotary is a big part of Mazda's heritage / pedigree, so even if their piston engine-powered cars outsell the rotary-powered ones ten times over, they'll still keep making rotary engine-powered cars...because not to do so would make them a company other than "Mazda".

Anyway, were I you, I would definitely keep working on the DIY mannequin head. I do believe that you can arrive at a decent sounding binaural mannequin head, and (especially) if you keep the costs low, you'll have that sense of value plus pride of having done it yourself. Even if it's not the 'best' binaural head microphone known to mankind, it'll always be a special mike in your gear wheelhouse.

So, I did a lot of digging on the web and found the foreign patent application for the Holophonics system as envisioned by Zuccarelli, and later, I fould the U.S. patent proper.

There are many, many things in the patent that scream 'binaural' and there appear to be (in my opinion) few, if any discriminators that clearly make the claimed holophonic process significantly different than a typical binaural as manufactured by Bruel & Kjaer, Head Acoustics, Neumann, Cortex et al.

In a nutshell, the U.S. patent describes both the idea of 'holography' as well as the process itself for making the mannequin head. The differences between what they refer to as 'other binaural' techniques comes down to this:

The process assumes that an x-ray (or series of x-rays) are taken of a human head in order to define its geometry. The mannequin head incorporates a sort of 'mass' like that of the human brain as well as nasal cavaties and such. What I can ascertain from the patent is that these things are all incorporated, but the mechanism by which sound is sensed is pressure transducers (i.e. an omni-directional microphone) mated to an ear canal (what they call the meatus) and that ear canal being part of an artificial ear. The patent also specifies that the ears should be made of something like silicone rubber (which is what is used in the B&K 4100, Neumann KU 100, the KEMAR, and the Cortex - only the Head Acoustics variants utilize an ear that more resembles a wedge than it does a pinna).

Basically, this like all true mannequin heads is unlike a BS-3D in that the BS-3Dis a sphere and not more of an ovoid like a typical human head (and thus mannequin) shape and lacks ear-located pressure transducers at the end of an artificial ear canal. What Zuccarelli refers to as "holography" is IMHO simply a binaural mannequin.

At first I thought that somehow they were using the nasal cavity as part of the transduction scheme, but as far as I can tell in the patent these are really just 'place-holders' of anatomy. While he does speak to the idea that the skull should have human-like properties, this isn't really given specific parameters (i.e. optimal mass, stiffness, damping etc). I copied the image and attached it. If you look, you will see a tube, and at the end of each tube, a pressure transducer.

The patent continues:

"hair is also of considerable importance in spatial discrimination, allowing a precise perception of sound coming from the front or the rear by giving rise to an asymmetric hologram. In fact, bald persons have been observed to have somewhat reduced front-rear discrimination along the axis of symmetry of the head"

Quick poll: For those readers out there who are bald, do you find it difficult to know when a sound is in front of you or in back of you?

Frankly, there are numerous reasons why sounds directly in front / back are harder to localize that those slightly off axis, but mush depends upon a) boundary conditions (acoustics), b) source-to receiver distance, and c) spectral as well as temporal content. However, this straight ahead / straight behind localization is difficult for all humans, whether having hair or being bald.

My eyebrows arose when I read that, according to the patent, the mannequin is equipped with "a wig which serves to generate asymmetry for front-rear discrimination". No data are referenced / cited to support the 'hair' claim. Sure, hair can affect diffraction as well as absorption, but the type of hair, its absorption (which, lest we forget, is a function of frequency) and such are all variables that are not defined in the patent literature. So, as far as I can tell, the purported difference on between a Zuccarelli mannequin and those of all others (on the outside) out there is that the inside elements of the head are reproduced (brain, nasal passages etc), and there's a wig.

Now, I can see how those endo-skeletal elements might contribute to how the transducers react to excitation (airborne and structureborne), and certainly when we speak (all these things affect the timbre of our voice for instance, but this mannequin is not used for telephony and as such, does not radiate sound through its mouth), but keep in mind that the diaphragms of the cited pressure transducers react almost entirely to forces normal to or at some angle incident upon the diaphragm, but the diaphragm is mated to the closed-open tube; the diaphragm, like all microphone diaphragms, would thus be relatively immune to forces acting in shear (along the surface of the diaphragm).

So, in my engineering opinion, it cannot be acoustically affected by the nasal passages nor the brain (in the mannequin) in the mannequin because there are no acoustic paths between the nasal cavaties and the artificial ear canals. This leaves only structure-borne as a potential contributor. However, one would have to know the relative mass and compliances of the human brain, nasal cavity, and ear (auditory) system if one wanted the vibration characteristics to be correct. This is, from a structural point of view, or paramount importance because the mass, stiffness, and damping values would define the equations of motion and thus, the frequencies at which the physical system is efficient at passing them along, and the frequencies that are likely to be attenuated.

That...is a very tall order, and...to my knowledge, I don't know of any published data that cites the normative distributions of mass, stiffness, and damping of the human brain, nasal passages, and auditory system, but assuming one could in fact get these data, and assuming that one could faithfully execute a 'representative' compliance, mass, and so forth, then a representative skeletal structure and skin would also be required to accurately complete the system. Since the purported argument seems to be that the pseudo brain and nasal passages play a role, then one would have to make the surrogates in the mannequin head possess a representative mass, compilance, and damping, because if there is transmissibility to the ear mechanism, that transmissibility, which is a function of frequency, is goverened by those equations which have those elements in them (like any spring-mass-damper system). I know that work in audiology has been done on cadavers, but even so, I am not sure how the compliance and damping of the things inside the human head (brain, ear, nasal passages) change when we are not living (i.e. such as a cadaver). So, one would really have to know the properties of these living tissues. I'm reaching the limit of my knowledge here, and perhaps there is a database / a bunch of published work in peer-reviewed journals about the compliance and damping of these oragns and their measured transmissibility to the ear...but I really don't know how one would accurately measure such things.

Now, all of the other binaural mannquins out there (and I have used all but one variant) have a lot in common, but a big one is that they are all heavily damped. That is, rap one of them with a knuckle and you hear a decidedly dead 'thump' radiate from them. Clearly the idea among all manufacturers of commercial recording as well as research-based binaural mannequins is to minimize the structure borne element of the head, and thus prevent the sound from being 'colored' (having its frequency response affected) by any structure-borne vinration, leaving the mannequin's ears to react to the pressure perturbations incident upon each diaphragm.

So, we have a lot of claims being made, and it would seem the way to answer them would be to measure the system in a controlled test environment so that the claims could be substantiated. This would tell us a great deal, and in fact, along with the complex FRFs observed, the impulse response of the systems would tell us pretty much everything we need to know. That is, if they are mathematically highly similar (they will never, ever be the exact same - even two mannequins by the same manufacturer will be slightly different due to manufacturing tolerances) then they will sense sound in highly similar ways. This priniciple is why every sm57 sounds alike or why every BS-3D sounds alike - because according to their complex FRF and impulse functions, they basically are the same.

If you want another trial, a subjective trial, then simply record the test sounds with each system, and perform a structured juried listening test - even if paired only, ABX would so. Of course, for a third variant, place a wig on each binaural mannequin...and go from there...but...what's the correct wig???

The one thing that I can see as being a possible unique approach is how the ear canals are implemented. There is some discussion in the patent about the geometry of the ear canals, and it would stand to reason that these would in fact affect the frequency response of the pinna / canal / transducer system. However, even taking these possible factors into account, it is still ostensibly a binaural process based on pressure transducers inside a replica of an ear, inside of an artificial head. I really can't find any real details in the patent as to the oto-radiation properties that are claimed by Zuccarelli.

Bottom line (IMHO), this is binaural, but may be different than other binaural systems due to changes in equalization. That's my best guess as to what this really is. Then again, I could be dead wrong...but I don't think I am.

Mark A. Jay
I like the Hendrix quote, never heard it before, it's worthy of Socrates, Christ, Ghandi, King etc...
I think to be truly real, the listener would have to have a sensor attached to its head to shift the mix depending on the such said angle and have speakers (stereo pairs) all around (double quad, if you will)(quad hi and quad low)
food for thought

Grazie mille. Mi dispiace per il mio Italiano (perche Italiano non e la mia lingua prima, e spero che parli Italiano), ma devo dire che si, e importante, ma la cosa piu importante e per fare le condizione correto al'lorecchie. Giusto?

Allora, e piu importante per le cuffie suonare come era la suona originale, e per quello, binaural e la moda piu correto.

Di nuove, dispiace per il mio Italiano; forse, nel'futuro, e meglio che rispondero in Englese.

Anche, grazie mille per le parole (circa Hendrix, Gandhi, Cristo etc).

non male!
I'm an American trapped here
I'd probably want to forget Italian if I could
Long story
Headphones have always confused me but when I do a good mix, I perceive that there's a horizontal and vertical element as well as LR and front to back.

Ah, OK. Well, if you have to be 'trapped' you could do much worse than Italy - you could be in Bulgaria (not that it's a bad place...). Try the grignolino wine when you get a chance. It goes so well when you are having some mascarpone and gorgonzola on some grissini...but I digress...

Anyway, I've heard a lot / read a lot about this sort of head tracking proposition.

Are you referring to the idea that a listener should hear the performance based on how he / she moves his / her head? If so, what you are proposing is essentially on-the-fly convolution. The tricky part is that you would need to have not just the impulse response of the venue, but a whole family of binaural room impulse responses (BRIR) at that location (physical) at that venue.

Assuming that you had this family of BRIR (binaural room impulse responses) you could likely use some sort of positional sensors on the headphones to tell a system the angle of rotation, elevation and so on (of the listener's head) and convolve the original signal with the BRIR that is most-correct for the head position being read.

Of course, the hard part would be adapting to these changes seamlessly (or close to seamlessly). In other words, let's assume that the lsitener is looking up and to the right (pick an angle...this is just a thought experiment) and then very s-l-o-w-l-y turns their head, then the change from one convolved response to the next could be managable (it's really just CPU horsepower).

However, what happens when the listener quickly changes their head direction? Could the system keep up? Could such a system 'learn' a user's responses to music and know that he / she will nod their head up and down with the beat (or maybe side to side)? If so, such a system could 'stack the deck' with a short list of the BRIRs that will apply to the anticipated head motions of the listener.

What would likely have to happen, though I don't know how this would be implemented, would be for a given BRIR to sort of 'cross-fade' with the BRIR that is on-deck. You would not want abrupt changes in the response because this would defeat the purpose of trying to create such realism.

It comes down to the acoustics of the space. For example, for the sake or argument, if you were in a wholly diffuse space (i.e. a reverb chamber), the BRIR's would basically be all the same, so the rotation and elevation of the listener's head would be moot.

Now, let's step toward reality and into a concert space that has fairly long decay times. Clearly, this is not wholly diffuse, but nor is it free-field. The BRIRs start to become more defined as a function of the listener's head. And, as we march toward free-field conditions, the BRIRs now reside at the opposite end of the impulse response continuum.

So truly, the adaptable system would have to be able to assess, based on the similarity / disparity in the BRIRs, whether or not to 'morph' the response heard through the headphones or not. For example, if you compared all the BRIRs from a reverb chamber, you get a sort of "NOP" condition - the system could morph one BRIR to another, but in such a set of acoustic conditions, this would be academic. The listener would perceive the same thing whether the system was convolving in real time or simply doing nothing (because there is no real difference in the BRIRs).

A hybrid approach could work. Maybe something like a model of a space based on ray-tracing is accessed to synthesize the BRIRs based on the ray tracing database.

Very thought-provoking...

Try comparing some electrostatics to some dynamics to some closed dynamics. Talk about a world of difference...

Anyway, if you are interested in stuff that sounds great in the headphones, check out the "What is Your Favorite Binaural Recording" thread, as well as a post of mine entitled "make a jazz noise here" : https://gearspace.com/board/remote-p...ml#post5314962

...or search on KU 100 to hear still more excerpts of mine from Gershwin (Rhapsody in Blue orcesatra rehearsal excerpt) to Bach (pipe organ in a small cathedral): https://gearspace.com/board/remote-p...ml#post5348691

Hope you like them...

Here are a few videos of Zuccarelli discussing his "holophonics" invention.

Unearthed Archives: Hugo Zuccarelli Part 1 - AOL Video

Unearthed Archives: Hugo Zuccarelli Part 2 - AOL Video

I've seen these videos...like I said...the contention is (was) that we localize by oto-acoustic emissions and not by ITDs, ILDs, or HRTFs.

I am not saying such things (oto-acoustic emissions) don't exist - I recall having read about them and their discovery circa 1980 - roughly the same time Zuccarelli took the stage and used that as a sort of purported backbone for the system.

If memory serves, they are sometimes emitted by some people under some circumstances, but Zuccarelli's contention (unless I completely misunderstood him) was that these are essentially occurring 24/7. That is, there are the spontaneous variety, and there are those that can be excited by a stimulus. Again, what I have read of them is that they are sporadic (unless stimulated) and are (I think) indicators of a certain type of hearing deficiency, though I can't recall which.

Still, what he is claiming, and as I mentioned in my posts above, what I take issue with, is not that it's basically a binaural process - not in the least, I am OK with that.

What I take I take exception to is the idea that the Holophonics system is purportedly active (i.e. generates tones / sound that interact / interfere with sound incident upon the artificial ears) , but none of the methodology is described - it is merely alluded to.

Patents are as much for generating revenue as much as they are for protecting the invention (and the intellectual property of the inventor). In other words, many a patent is written solely with the idea of licensing the technology to another company for some set licensing fee / piece-price royalty. Believe it or not, often times the company that patents an invention has no intention whatsoever of manufacturing the invention - they just want to prevent the 'other guys' from making it unless they are going to pay a fee to do so.

So, if anything, full-disclosure of how the invention works and what makes it unique are critical to the inventor's reputation as well as economic well-being. Failure to fully describe but merely to allude greatly erodes the strength of the patent - and...makes it easier for someone to 'work around'. So, if you have something really novel and something unique...you spell out, in painstaking detail how the invention works. It's a long process...I know...I've been involved in three patents and believe me, it's a long row to hoe.

Anyway, if one does not do this (explicitly call out details as to how the process works), then you have to wonder why this is the case.

The videos are from 1992. It's now 18 years later. The patent has expired (if I am not mistaken). Where are all the Holophonic systems that could have been made after the patent lapsed?

Again, he seems like a very bright guy, and indeed, he could have managed to discover what thousands upon thousands of other researchers have failed to do over the last 70 years. Things like that do indeed happen, but it's been a full 30 years since the initial patents; computers and analysis methodologies have gotten monumentally better and faster and yet, no one (to my knowledge) has been able to step forward and say "indeed, we are viewing an axion; here are the data that prove he was right all along".

I'm sorry, but it was never embraced by research facilities or universities around the globe...and that pretty much tears it for me. If it's revolutionary and can be replicated in various labs, then universities and research facilities would have been all over it like a cheap suit.

It's 2010 and pretty much all percpetual research is still being conducted with conventional mannequin head microphones.

I guess the last nail for me is always when an inventor doesn't 'tell' and doesn't discuss details (not even in a patent), nor makes his invention available to the scientific community to be measured and compared to other systems and thus, be proven.

I just re-watched part 2 and I wanted to pont this one thing out because it goes by extremely quickly...

First, the room is shown from the speakers' perspective; you see brick etc which would seem to indicate a typical room. However, for just a brief second, the area where the speakers are is shown and toward the listeners on the side walls there appears to be a heavily-damped region, thus absorbing a lot of the high-frequency content. Effectively, depending upon how much and the type of absorption, this minimizes the reflections that (for a pure binaural recording) cause crosstalk and inhibit good localization over speakers (which is why binaural always better localizes over headphones). This is not to say the crosstalk is removed, but it is minimized.

Also, as far as a juried test goes...it's very bad science the ways it's being run- no one that I know of, working in percpetual studies, would ever invoke this methodology. Normally, when you run any kind of feedback-based test, you make provisions to ensure that one juror is not influenced by another's actions.

That is, some of the jurors could be pointing where others are pointing because others are pointing and not necessarily because they are localizing the sound. Are they influencing one another? Are they not influencing one another? We can't say, because the experiment wasn't properly set up and thus, isn't valid.

I have seen this first-hand when jurors take listening tests. Even you you stress to them that there is no wrong answer, you would be amazed to see just how many jurors try to watch other jurors' vote and then themselves vote only after they have tried to see how the other jurors have voted.

So, if you want to do this sort of thing right, you have to make provisions so that one juror / listener cannot observe what another is doing, and thus, you remove environmental bias.

It could be six of one and a half-dozen of another in this set-up, but arguably, this is NOT how one should conduct a listening test / demonstration.

Mark your dismantling of the whole "holophonics" idea is a joy to read. I know there is no malice in it - only science.

It's very interesting to read comparisons of the various technologies. I came across this admittedly semi-scientific thesis written about a comparison of the OKM Soundman binaurals and the KU-81i. It's an interesting read even if one doesn't agree with it.

http://www.soundman.de/pdf/RA303-Thesis.pdf

I have the OKMs and have had good luck with them. I should get them out and do some more work recording "stuff" around here. I find it hard to sit through a concert as still as the Neumann head, however. The binaural effect is good or at least I think it is good.

I am reading the PDF from Soundman's site that you linked. Thanks

Tut: I'll definitely read through this soon. I am much more acquainted with the KU 100 than I am the 80/81 (though I know a little bit about them). I'm not sure why they chose a KU81 as part of the work - as far as I know, Neumann have not manufactured them for some time (I think the 81 is circa 1986...). I know very little about the OKMs. Time to hit the woodshed...

Anyway, about the differences between the '81 and the '100...I believe that the equilization is supposed to be the same (diffuse-field equalized), but I think the '81 uses different microphone elements etc, though arguably, I have no experience with the differences between the two units as I have only (in terms of Neumann mannequins) worked with KU 100's. I think the ears may have also been changed, but I don't remember the whole pedigree right now.

I'm curious to read it (I only skimmed it) because there are some juried tests in the text you cited - those really interest me as getting juried results to make sense and be valid can be tricky. You would not think that a simple thing is so complex...but it is (or it can be). I have been involved in a few juried tests (designing and running them) as well as functioning as a juror, so it will be interesting to see the methodology.

I'm also curious if methodology used is based on the Bradley-Terry-Luce models (and incidentally, if you are interested in juried test methodology for listening studies, web-search Norman Otto (formerly of Ford) as well as Scott Lake (formerly of G.M.) and their S.A.E. / I.N.C.E. papers).

Anyway, as far as 'dismantling' the Holophonics system, you're right - there really is / was no malice on my part. My issue is with that which seems to border on false advertising (if not outright); if you claim it is an active system, demonstrate it as such and let others verify your approach with independent measurements.

Like I said, if Zuccarelli had said something akin to "this is binaural, but it's different than the other binaural mannequins because I have paid closer attention to the [fill in the blank with measurable properties system / assembly / mathematical construct]..." or something like that I honestly would not have any issues with it.

This paper reminded me of a post by nature recordist Bernie Krause on the nature recordists yahoo group. It is an interesting anecdote.

If one's looking for a reasonable binaural set up, just take two omni
lavaliere mics (I use Sony ECM 55Bs), tie a rope around the trunk of
a tree with the diameter equivalent to the distance between ears on a
human head, attach the mics opposite each other, and voila! No
hassle. No great expense (about $700.00 should do it). And they will
last longer under humid condition than other types.

Once, on a project, we were given an Aachen head (with Schoeps
capsules) implanted. Costs $25,000USD and comes in a 90lb. monster
Anvil case. Our client had us drag the damned thing to a Costa Rican
jungle where we set it up. The mics failed in 15 minutes (humidity).
We implanted another spare set of mics and those failed, too.
Finally, we tied the 55s to a nearby tree, recorded, and sent the
binaural stuff to our client, who swore those were the best (Aachen)
recordings he had ever heard (we never spilled the beans). Those
same recordings are currently featured in the binaural hearing room
at Disney's Animal Kingdom in Florida. So much for the claims of the
fancy stuff.

Bernie

Well...

I hate to say it, but this (link) isn't really anything in which I would put a lot of faith - I'm not sure where to start, so I will start with the listening tests (I might as well). I am not slamming / praising OKM, nor am I slamming / praising Neumann, I am however taking issue with the methodology of this paper. That's a shame, because with proper technical controls and merit, the paper would (with some additions) be of greater merit.

There is exceedingly little detail in the description of how the listening tests were carried out. There is no overt declaration as to whether these tests were done blind, double-blind, calibrated loudness (i.e. 1:1 playback), were they normalized in terms of loudness (between microphone type), and if so, what method etc. How were the votes tallied? Did they raise their hands, use keypads? Could they see one another? Were their votes checked for consistency via the method of circular triads? Scaling methods - Google Books

Others have pointed out to me that this is only a discussion forum and therefore should not be treated as if it were a peer-reveiwed journal. However, I disagree with that when it comes to citing data. You see, for all I know, this test followed a classical paired comparison test, using the method of circular triads (to screen jurors for consistent votes, lest they be 'noise' factors in the analysis of the data), but how would we know this?

Again, I don't want to come off as heavy-handed in this, but I can assure you that this paper could not be submitted to a peer reviewed journal because the validity of the experiments are all in questions for one simple reason: The author failed to disclose any of the particulars about how these data were derived, as well as a host of other matters not discussed pertaining to the listening tests.

Also, let me make this perfectly clear: I am not saying that the OKMs can't make good recordings that sound very convincing. I have said of others' in-ear binaural works that they are very good-sounding (check out some of Dallas Simpson's works) and seem very convincing to me because they are getting what I consider the most salient aspects of the boundary conditions either met or closely approximated; correct spacing, correct shadowing, and the pinna (and to some extent the cavum). The single biggest difference between in-ear and mannequin seems to be the ear canal effects. In the mannequin, they never vary, but with an in-ear binaural approach, each recordings can be slightly different.

If you check out the Best Bets for Binaural thread you'll see that there has been some wrangling going on there over what constitutes binaural, but for me, I think that in-ear and mannequin are most highly similar (and what I think constitute binaural) and are closely aligned due to the similarity in boundary conditions that each approach provides.

Now mind you, I truly am NOT saying that one system is 'better' than the other...but what I am saying is that ironically, this text's lack of structure and documentation with regard to test methodology, juror consistency checks (were they even employed?) and so on does more harm than good in terms of promoting a dialogue on why any of the preferences observed are what they are. All of the statistics and preferences presented would have real 'weight' if the listening test methodology was fully described.

So let's turn it around - is it possible that two systems perform as comparably as the paper asserts? Yes, it is entirely possible, however, since there is such poor documenation as to the listening tests, we can't really say for certain.

I do not wish to come off as sounding condescending here, but I have written papers on signal processing techniques, noise control, sound quality, etc, and all but a few magazine articles have been for peer-reviewed journals (S.A.E., I.N.C.E., I.M.A.C. et al). I've reviewed other engineers' papers to be published that were (quite literally) orders of magnitude more thorough, documented, and better written than this paper is, and even those papers need to be re-written several times.

Am I being overly harsh? No, because if you are going to conduct a test (and a listening study is a test) you had better document pretty much every aspect of it; to not do so is to relegate ones work to be dismissed and not taken seriously.

You see, peer-reviewed publications help separate the bad from the good science, and while you may be saying "Mark, this isn't about science...it's about the art of music", the fact is that this paper presents conclusions based on methodology and controls, purported scientific methods that are not really described in the paper.

While papers like these make for interesting discussion, they don't really serve the dialogue of how we get to the bottom of perceived differences. Like it or not, that takes some very well-controlled juried tests that bear little if anything with what's been published here. Look...the science of juried tests is a very dry and boring subject (at least it is to me), and kudos to all those out there who found the ability to develop the methodology that allows us to say "yes, these sounds are judged to be statistically different as judged by a pool of jurors whose votes were deemed reliable and coinsistent by several consistency check methods"

So, if you just can't sleep and really want to know about how juried tests are done, here's a great place to start: AES E-Library: Listening Test Methods for Automotive Sound Quality

I agree that this was not a good examination of which systems works best. I just hadn't seen any real comparisons done (scientific or otherwise) between the various "binaural mics" and thought it would fire some more debate. It doesn't help validate anything when I found the paper listed on the very site selling the less expensive model either.


BTW Can one join the AES without being a pro? I've been getting teased with all of these papers and videos that I can't view. It looks like I need to pony up the cash but I'm worried I will spend days going over all of the papers and videos and not get any work done heh. I saw some AES videos that Bob Schulein did about binaural and they seem to be very interesting also.

I would LOVE to compare binaural mic to binaural mic, or mix version 'a' to mix version 'b' etc, but, without the proper experiment hardware, I would not want to publish the data.

This is why I find so many of the listening tests that I see (here and in other forums) as grist for the mill. To really do this properly is tricky, but trust me, this method is highly developed as well as decidedly entrenched in industry (most people probably have no idea how many listening tests are performed to find the 'best' door closure sound, or the best sunroof sound, or the best washing machine sound quality - these tests are used to control how your appliances sound). I'm dead serious - this is a methodology that really got defined in the early 1990's (for sounds comparisons) and has become as commonplace as algebra

As far as AES menbership goes, there are different types, and I am not sure which best suits your situation / plans. However, here is the link:

Membership Types

Frankly though, I think the membership is worth it. I mean, papers are availabe deeply discounted, and in my opinion, places like AES, SAE, INCE etc, because they are peer-reviewed arenas (for publications), I feel much more confident about the quality of the research. Reading others papers is something that I find to be very enlightening, and a great use of my time.

Yes, you can.

It's the Audio Engineering Society, not a professional institute.





Posted via the Gearslutz iPhone app

In my quest to find any material that can help me better understand the principals involved in inexpensive binaural recording I came across this powerpoint presentation. It's pretty interesting.

http://www.gbcasa.org/cms/audio/Grie...Headphones.ppt

Tut: Thanks for the powerpoint presentation. I just started to skim over it, but will give it more attention later on.

I noticed that Stephan Peus' name was in the presentation regarding the KU-100; Stephan and a colleague and I co-wrote a paper once for an ATA conference in Bologna, Italy back in the 1990's (it was a technical paper on the development of a surrogate human being with a binaural mannequin head - paper 01A1019). I had no idea he was such a force for the creation of the KU 100 (and its more anthropormorphic ears and ear canal-based microphones).

Anyway, as for the presentation that you found, as I said, I will defintily review it later as it looks like it could make for some very interesting reading. Thanks for having un-earthed this one.

I was doing some other research about the author of the last paper and went to his website. There are some other binaural related powerpoint presentations as well as some interesting technical info that I'm slowly going through.

Here's his website

http://www.davidgriesinger.com/

And here is a link to an audio presentation his delivers

GBCASA Meeting Featuring Leo Beranek and David Griesinger

It might take me awhile to wrap my head around some of this though

I'm reading this thread with real curiosity. Recently, on Freesound Project I found something interesting regarding this topic. Do you think, they might be right about the "head interior" concept? Did anyone tried to follow such approach?