CanJam at RMAF2017 People I Love - Paul Barton

Paul Barton is a big fan of headphone measurements, so much so that he's currently working at Canada's National Research Council to improve headphone measurements with more anatomically correct ear canals. He thinks measurements can be made more accurate above 8kHz in that manner.

You may remember the Sony MDR-Z1R measurement discrepancy between Jude's rig and mine. You'll recall the biggest discrepancy was at around 10kHz. Jude uses the G.R.A.S. KEMAR head and torso simulator with anthropometric pinnae that includes anatomically correct ear canals; the ear canals in my head are perfectly cylindrical. I've been thinking about it quite a bit, and I've got to think the anatomically correct ear canals have lower-Q (more damped) resonances than mine. We may not be used to seeing the curves as presented on Jude's rig, but they may actually be more representative of what we hear.

I asked Paul about it in the video and his response stimulates my brain. Check it out.

View on YouTube.

Rthomas's picture

Hi Tyll,

Have you ever asked Paul whether he's working on a purist, open headphone? I wonder what Paul could do if the only objective was sound quality without having to accomodate a bunch of smartphone and travel related features. A NAD or PSB HD6XX/HD800 competitor would be very welcome

Tyll Hertsens's picture
I'd love to see that as well. I'll try to remember to metion it when I talk with him next.
Beagle's picture

In his interview with Jude a couple of years ago, he mentioned he was working on a premium flagship model, or it was I the works at least.

100VoltTube's picture

Mr barton talks about 3D printing ear canals, out of plastic, I presume. Real ear canals are not made of plastic, (duh) but does that matter? I would have thought that plastics, being more rigid, would resonate more, and present a higher acoustic impedance than human flesh.

Tyll Hertsens's picture
Totally agree...but you gotta start somewhere. Just playing with the geometry is enlightening evidently.
JMB's picture

Most human tissue has much more viscoelastic properties and therefore has much better damping than plastics which behave more elastic i.e. low damping.
Why should the ear canal be considered in headphone measurements?I In ear phones shorten the ear canal and how much will depend on on the exact form and length. So the normal effect of it is eliminated and should be compensated in voicing and measurements of IEMs. (Maybe length and form of ear canal would be the reason why someone likes or hates IEMs)
But with circumaural headphones the ear canal is not altered and contributes to hearing in the same way it does with normal hearing. The same should be true then for artificial head measurements and if that measurement head is linearized in free space (like our brain does with our ears) it should work also with head phones independent of the exact artificial ear canal if the free air compensation is correct.
The outer ear has influence on hearing and especially on ear phone will alter that. I wonder how off the ear headphones (like the new Myshere 3.1) would measure differently to open circumaurals?

DonGateley's picture

The eardrum itself with the tympanic cavity and bones behind it should significantly affect the overall transfer function of the ear at the tympanic window where motion enters the cochlea. That is the only location that really means anything when considering measurement and it is rather difficult to get to. For this reason I consider all current methods and devices to measure 'phone response to be largely invalid.

Such incomplete plots have and can have no absolute value nor can they be interpreted in any way other than to compare the plot of one phone relative to another phone (convolutionally divide the impulse responses.) The ratio can be meaningful even if the actual plot isn't.

The region above 8k responds to the most minor factors. Simply taking the 'phones off and putting them back on radically effects the measurements I take right at my eardrum with a .1" Knowles mic. I have found the repeatability of such measurements to be most elusive.

jude's picture

At the 2016 AES International Conference on Headphone Technology in Aalborg, Denmark (which Tyll also attended), Andrew Bright (then of Goertek, now the Director of Audio Systems Engineering at Apple) gave a very interesting talk called "Headphones, known-knowns and unknown-unknowns," and, among other things, he discussed this.

In his summary he said, "For me, when I undertook this, I was kind of expecting to conclude that we should try to measure something in the cochlea, but I was surprised that measuring the eardrum pressure seems to be sufficient. This was a surprise to me."

Among other things during the talk, he cited papers by Hudde et al. titled "A Full-Size Physical model of the Human Middle Ear" and "Measuring and Modeling Basic Properties of the Human Middle Ear and Ear Canal. Part III: Eardrum Impedances, Transfer Functions and Model Calculations," in which Bright was discussing the question of whether or not we should be measuring something more directly analogous to pressure into the cochlea, and that maybe doing so would give us something more consistent when measuring at high frequency. From these and other studies he concluded that the answer is that it's "probably not necessary."

Among the findings from Hudde et al., Bright said "at high frequencies (>3 kHz) the eardrum presents a largely reflective surface," and that "individual characteristics of the eardrum, middle, and inner are unlikely causes of individual differences in headphone response." He further said, "The transfer function from eardrum pressure to vestible (cochlea) pressure, while not flat, does not exhibit the wide variations seen in person-to-person measurements. This is an unlikely cause of individual differences in headphone response. It is thus likely that pressure at the eardrum is a good measure of hearing."

So, in his summary, he stated that eardrum pressure is likely a good measure of hearing. He also said that differences in a headphone's response between individuals are caused primarily by variations in ear canal length and area profile (area vs. length), which he did go into more detail about during his talk.

I'm abbreviating a very compelling hour-long talk by Bright, and perhaps not well enough. Given your post, though, I thought you'd find this part interesting.

DonGateley's picture

Thank you for that. Upon further pondering I too find it highly likely that the eardrum is stiff and reflective. Sufficiently stiff to attenuate the motion through the bones and to the cochlea's oval window (which is 1/18 the area of the eardrum.) The eardrum is conically shaped (for stiffness?) with the tip of the cone connecting to the malleus.

There is also a muscle which stiffens it further in the presence of high sound pressures. So, yep. It is probably pretty nearly completely reflective such that the inner ear contributes little to the response and pressure at the eardrum is what counts.

I am still at a loss as to why I was unable to get even near repeatability of measurements above 8k at my own ear drum. The slightest change to 'phone position caused large changes in response as did removal and replacement of the stiff, low physical volume jig I made to get the wee little microphone placed repeatably near my eardrum. Ah, well, there are things I just wasn't meant to understand. :-)

munnalul's picture

Pixel Gun 3D is among one of the most popular video games in the world of online video games and this is simply at no cost. Many thanks to the group of hackers & developers that handled to locate some technicalities & manipulate the server of Pixel Gun 3d.