Binaural VS ORTF

[bobkeiser@ieee.o]

Originally Posted by MBishopSFX

I love this mic and am amazed at all the uses I've found for ol' Fritz over the years.
.

Yes, and Frau Blucher misses him immensely...so experienced, so worldly...so...binaural.

[Mark A. Jay]

Originally Posted by bobkeiser@ieee.o

Yes, and Frau Blucher misses him immensely...so experienced, so worldly...so...binaural.

That's some pretty funny sh*&...and a reference to "Young Frankenstein" as well? Manna from heaven...

[Mark A. Jay]

Originally Posted by Broken

1. Binaural only works well for loud speaker if you secure the listeners head in the right place. It is totally useless for anything other than lab listening tests.

It's ideal for headphones.

2. Binaual is great if you want to capture the elevation and azimuth. It isn't perfect and localisation errors happen around specific angles. This is due to the fact that we rely a lot on head movements to pin-point a sound source.

Head tracking helps but this technology isn't widely used.....yet.

point 1: To say that binaural is 'totally useless for anything other than lab listening tests' is very dismissive. While I will agree (as I have stated elsewhere) that the most-accurate and realistic approach to binaural is via headphones, playing it over speakers has its place - even very experienced producers (i.e. Tchad Blake et al) have spoken about binaural's 'depth' over speakers and how it differs from conventional stereo.

I myself have used binaural as part of a conventional stereo (that is, two-channel) mix where the signal(s) from the mannequin(s) were hard panned left and right, and other conventional signals were mixed. Others use binaural as part of a 5.1 or 7.1 mix. Is their work (when incorporating binaural as part of the 5.1 / 7.1 process) also 'useless' because it is to be played over speakers?

Again, with regard to point number 1, I have had clients and acquaintances alike listen to binaural recordings on their stereo systems and have heard some very interesting comments. For some, they tell me tthat it sounds 'different' and they can't quite place it, but they find themselves liking the way it sounds. For others, they don't like the sound over speakers at all.

This seems to be largely an issue involving one thing in two places: the boundary conditions under which the recording was made, and the boundary conditions under which the listener experiences it.

In other words (and this is related to your 2nd point), if a binaural recording is made in a comparatively diffuse space then that recording will sound very different than that same musical entity recorded in a much less reverberant space. That's logical and in fact, would be the case also over headphones.

However, some of those who have listened to binaural over speakers have (without necessarily their knowledge) a sort of LE/DE room in their house, and the speakers have usually been 'parked' in the DE portion of the room. Additionally, in at least one of these scenarios, the listewner-to-speaker disctance is small. In such boundary conditions, the speaker / human / room interface starts to approximate (not achieve but approximate) headphone playback; the listener experiences primarily direct sounds from each speaker, although to be certain cross-talk and other reflections would and do exist. Granted, I would say that the vast majority of listening environments do not approximate a sort of giant pair of headphones as I have described, but rather, a much more diffuse environment - and under such playback boundafry conditions, the cross-talk un-does much of the accuracy of the signal, but once more, this does not rtender binaural 'useless'.

The point that I am making is that how 'well' binaural assimilates itself to a conventional two-channel speaker playback chain depends a great deal on the boundary conditions under which one listens. Interestingly enough, one could make the very same assertion about just about every conventionally recorded two-channel recording out there, albeit not to such a great effect; every listening room alters (ultimate playback of) the mix and presents it to the listener in a very different way than it was engineered, mixed, and mastered; all of those control rooms affect what goes into the final mix.

Until the day comes when someone provides the impulse reponses (from the monitors in the control room) for the recording studio engineer's chair, from the mastering engineer's chair (the impulse reponse through his / her monitors in his / her room), and makes the available to the public (and every audio system is capable of successfully and transparently convolving these elemnts) so that they can convolve those impulse reponses with the impulse response of their own room, then truly nothing we listen to in any environment is technically correct. Should we then say that because we don't replicate all of those boundary conditions heard by the recoring / mastering engineers that the music is somehow 'useless' unless we account for all of those room impulse responses?

Hmmm...now that I think about it, I kind of wonder what microphone one would use to capture the impulse responses of the environments - my money would be on a binaural mannequin.

Thus, your generalization that binaural, when not heard over headphones is 'totally useless for anything other than lab listening tests' is cavalier as well as highly dismissive of the creative / aesthetic effects. Again, I concur that binaural over headphones is the most technically accurate way to reproduce binaural (and I would contend recorded sound in general) but that doesn't mean that someone can't embrace the effect it achieves over speakers.

Point 2: I spoke about this issue (related to head-tracking) in another post, but in a nutshell, having the HRTF change in real time based on the movements of a person's head could be a very, very interesting thing.

The issues that I see (personally - just my opinion here) is that in spaces that are rather reverberant, the differences in the HRTFs would be minor - but as Dallas and others have pointed out (and I agree based on experience) as the boundary conditions approach a highly or fully reverberant space, then the diffuse field makes localization very difficult, and mannequin recordings in such a space have poor localization.

Indeed they do - and they should because this is a core virtue of binaural mannequin technology - temporal and spatial accuracy, probably the most accurate of all techniques out there. However, accuracy doesn't always translate well to the aesthetic, but I am merely making the point that binaural mannequin recordings (with or without head tracking software and real time estimates of the HRTFs) in reverberant conditions would sound as undefined over headphones as would the sound we as humans would experience in such a space.

On the other hand, as the boundary conditions get closer and closer to free field / anechoic conditions, then the differences in HRTFs would be extremely important, especially so for a small source-to-listener ear (i.e. mannequin ear) distance. In this case, the ability to morph one HRTF to another and another and another in real time would probably just about as 'real' sounding as is possible with current technology.

[Mark A. Jay]

Originally Posted by Big_Bang

You got me... I didnt read the whole thread, but I promiss I will.

And THANK YOU for the tips on them books! Been looking to get some!! thankssss!!

I am extremely interested in binaurals, but for completely non-musical applications. I am also extensively knoweledgeable on the psycho-acoustic phenomena resulting. right now its the pure engineering of the recording process. I have extensively heard binaural recordings to an extent that my mixing skills have gone throught the roof! I have much more perception correlating drum, playing with phase on guitars, vocal and fx etc..

Me want to DIY a dummy

But this is what I am getting at. If the mic is inside the dummyhead, and there is but 1 point of entry to the mic's membrane, plus the polar pattern is skewed by the fake ear and cannal... what the point of having and omni?

Isnt it logical to use a hyper-cardioid condenser?

Whats left over is mechanical vibration, that properly mounted will certainly hit the membrane. Different timings to each microphone would even increase "delta T" perception.

BUT (this is my logic working up on me) ...

If its omni, the preassure gradient on the negative X (membrane facing inside dummy) axis would cancel the opposing positive axis (ear canal membrane), thus canceling (to an extent) the mic's response to that frequency. If you take into account the opposing microfone, the effect could be higher.

If it were hypercardioid, this would be by far less prominent, and whatever inter-microfone cancelations would highten perception...

wooof, I tired myself with my reasoning hope its not too confusing...

All of my replies are now [bracketed]. I changed the text to black font as I was unaware that red font is problematic.

"Me want to DIY a dummy "\


[I highly encourage people to build their own mannequin heads. As long as you have the skills, the patience, and a modest budget, you can make your own and realize about 90% of the benefits for around 10% of the cost (of buying new). Of course, there are some real advantages to buying a well-engineered mannequin head microphone, but then again, to just get used to the technology it's a good idea. Plus, if you build one and don't like it, you can take it apart and use the omni's once more as regular microphones - the risk is pretty low.

My advice here would be to spend most of your money on two low-noise omni elements with flat frequency response. In that way, you would know exactly what the change in frequency response would be due to the head, the ears, and the ear canal. If you use an ultra-cheap microphone element, it gets a bit tougher. Of course, you can always come up with your own correction factor - it's not that difficult.]

"But this is what I am getting at. If the mic is inside the dummyhead, and there is but 1 point of entry to the mic's membrane, plus the polar pattern is skewed by the fake ear and cannal... what the point of having and omni? Isnt it logical to use a hyper-cardioid condenser?"

[Many people make this assumption - that the microphone elements should be cardiod (or hyper-cardiod) but that is not the case. The omni is acting as a pressure transducer, and this is indeed what is desired for an in-the-ear microphone. The head shape, pinna shape, and ear canal (as a minimum) define the complex polar pattern (the binaural polar pattern is very different and much more complex than any polar pattern alone).

You see, what you are attempting to do is to imitate that physical system so that the signals imparted to the microphones are very similar to those that would be imparted to your eardrums if you were in the space occupied by the binauaral mannequin head. The dimensions of the head, the ears, their geometry, and so on all contribute to the very complex polar pattern.]

"Whats left over is mechanical vibration, that properly mounted will certainly hit the membrane. Different timings to each microphone would even increase "delta T" perception."

[Microphone elements are much less sensitive to shear forces than they are forces normal to the diaphragm(s). Thus, if vibration is imparted to a properly-constructed mannequin head, then the microphones will be (for the most part) insensitive to vibration imparted in the X-axis (front to back motion) and the Z-axis (up and down motion). Note: I am assuming right-hand coordinates.

The Y-axis (lateral or if you prefer, left to right) would be most sensitive to mechanical vibration because those forces would be normal to the surface of the diaphragms. However, I have not noticed this in my binaural work, but then again I have not imparted any substantial vibration to the Y axis of my mannequin head.

Anyway, any microphone is prone to picking up vibration, but in pretty much every case I can think of, the sensitivity of the mic to vibration is highest for vibration normal to the diaphragm.

As far as the delta T...the ear-to-ear dimensions of the mannequin heads have been chosen to be close to a specific percentile of the human population (most often, 50th percentile).]

"If its omni, the preassure gradient on the negative X (membrane facing inside dummy) axis would cancel the opposing positive axis (ear canal membrane), thus canceling (to an extent) the mic's response to that frequency. If you take into account the opposing microfone, the effect could be higher.

"If it were hypercardioid, this would be by far less prominent, and whatever inter-microfone cancelations would highten perception..."

[There is no gradient being measured in an omni. Cardiods (and other variants) operate on a pressure gradient principle but omni's constitute a pressure transducer, not a gradient-based transducer. As such, the diaphragm responds to pressure that is incident upon it, and this pressure can only be normal or at various angles to the surface, but not to the back side of the diaphragm. Thus, the omni senses only the pressure that is incident upon the ear / ear canal.

You can argue that there is a gradient that we experience (and this is true) when we hear, but so too would a mannequin head microphone; the gradient needs to be measured by both transducers, just as we sense a gradient only when we have the use of both ears. That is, it is fair and correct to say that a pressure gradient is measured when considering the signals from both ears, but one ear alone cannot measure a gradient.]

[Does this help?]

[NorseHorse]

Originally Posted by Mark A. Jay

Does this help?

That helps. Using the GS "quote" boxes is also helpful. You can copy / paste / delete the portions you need. Like this:

Originally Posted by Mark A. Jay

Does

Originally Posted by Mark A. Jay

this

Originally Posted by Mark A. Jay

help?

[Mark A. Jay]

Thanks...I'll change the font color.

[jacktadoussac]

This is one of the best threads I've seen here at Gearslutz -- in terms of motivating people to do stuff or try a new approach.

From what I've gleaned from this, an enthusiast could very well take a pair of quality omnidirectional mics and install them in a homemade head, and you're on your way to sonic ecstasy.

For some reason I can imagine carving a wooden head perhaps with a softwood like white pine. Cherrywood and oak are obviously much harder but I don't think I would be daunted my softwood -- of course my head is much harder stuff...yada yada.

Has anyone out there carved a wooden head and installed a pair of DPA 4006 mics?

[Mark A. Jay]

Originally Posted by jacktadoussac

This is one of the best threads I've seen here at Gearslutz -- in terms of motivating people to do stuff or try a new approach.

From what I've gleaned from this, an enthusiast could very well take a pair of quality omnidirectional mics and install them in a homemade head, and you're on your way to sonic ecstasy.

For some reason I can imagine carving a wooden head perhaps with a softwood like white pine. Cherrywood and oak are obviously much harder but I don't think I would be daunted my softwood -- of course my head is much harder stuff...yada yada.

Has anyone out there carved a wooden head and installed a pair of DPA 4006 mics?

...and probably to pretty decent success. That's my guess anyway. I'm not sure what the mass would be (not knowing the densities of various hardwoods off the top of my head), and I'd have to think about the damping / mobility of the material.

However, my guess is that for all but insanely-high SPL levels, the mobility would not be an issue and thus, not color the mics due to structure-borne vibration exciting the mic diaphragms.

Frankly, it sounds like a cool idea...even from an aesthetic point of view.