How Do Measurements Sound? by Audeze

Editor's note: After my review of the LCD-4, Audeze emailed me concerning discrepancies between my measurements and theirs. Numerous fairly technical emails were exchanged, I sent a number of Excel spreadsheets of my measurements for their use in comparisons, and there certainly does seem to be a discrepancy between measurements.

The folks at Audeze asked if they could write an article about these measurement observations and their response to my LCD-4 review. I said, "Of course, I've got plenty to learn on this subject." This article is Audeze's response.

How much do you trust measurements to determine how a headphone will sound? What if measurements from different systems vary significantly, which set of measurement would you use to judge how a headphone will sound?

Headphone measurements are tricky. Interpreting them and relating to what you would likely hear is a bigger problem.

Before we get into the recently published review and measurements on the LCD-4, let's begin with an example. Here's a comparison of the EL-8 measured on Tyll's setup and a calibrated GRAS 43AG ear and cheek simulator at Audeze. As you can see, Tyll's measurements shows a significant drop between 5 and 8 kHz, but you don't see it in the GRAS 43AG. How will you know which of the two measurements reflect what you will hear?

160413_Blog_AudezeMeasurementsExplaination_Photo_Img1

The differences we see in measurements -- not just on Tyll's system but across multiple systems -- are due to many factors, including how the inner ear is modeled, the differences in pinna and concha geometry, and how the headphone and earpad interact with the pinna. Adding to the confusion are measurements published by different users.

Measurements of the same headphone on multiple systems yield significant, if not dramatic, variations in the plotted frequency response, especially above 2kHz. It's not easy to interpolate and go from one system to another, let alone interpret them.

Measurements are useful if their limitations are understood and they do play a key role in our design process.

Audeze Design Philosophy
Audeze designs its headphones to sound as natural and as close to the original performance as possible. We always strive to accurately reproduce instruments and vocals, each in their own space. Our goal is to make them sound like a good pair of speakers in a well treated room.

We accomplish these goals by evaluating our headphones in a number of ways:

  1. Our engineers have their own set of measurements made with probe microphones near the eardrum and we measure the response of our headphones and our reference stereo speakers. Then we measure using dummy heads and ear simulators. Over time we've accumulated tens of thousands of these measurements in our database.
  2. We compare test tracks played through our headphones to those played through a pair of reference stereo speakers in a well-treated room.
  3. We conducted dozens of live recording sessions with a grammy award winning recording engineer and recorded vocals and instruments in controlled environments including position of the artists, microphone orientation, etc. We listen carefully to how our headphones reproduce the music. We're proud of the accurate imaging and natural tonal balance of our headphones.
  4. We have developed DSP tools that allow us to modify the frequency response and impulse response of our headphones without introducing artifacts and without the need to physically modify our headphones. These tools help us evaluate how changes and tweaks to the headphone's response affect what we ultimately hear.
  5. We use several sound engineers to beta test them in studios.

LCD-4 Measurements
The LCD-4 is the results of years of research and development to ensure the most natural and transparent sound possible. That's why it was a shock to read Tyll's LCD-4 review. We respect Tyll's subjective opinions about the sound and we'll leave it at that. But the measurements published on Innerfidelity and their interpretation contradicts our own.

After Tyll's comments on the measurements, we reviewed everything objectively; the summary of our findings and our view of the measurements are provided below. We also measured the LCD-4 at a third-party facility and shared this information with Tyll.

The Issue
Here's what Tyll's article said about the frequency response of the LCD-4 he reviewed:

(Quote: A) "Most all LCD headphones I've measured in the past have a distinct drop in treble response starting at 4kHz and going up to about 8kHz."

(Quote: B) "You can see that the LCD-4 is essentially flat from 4kHz to 20kHz when it should be a falling response."

We simply do not see this in any of our measurements. And we do not see this in the audio quality descriptions we get from the vast majority of LCD-4 owners.

More measurements, Checking and Double-Checking.
We measured the LCD-4 on:

a. Artificial head: GRAS KEMAR RA0045 Ear Simulator, KB0066 and KB0065 Pinnae in IEC 60318-4 (previously 60711) configuration.

b. Ear and Cheek Simulator: GRAS 43AG in IEC 60318-4 configuration.

c. Headphone test fixture: GRAS 45CA, KB0071 pinnae, IEC 60711 Ear Simulator. (This setup is similar to the one used in in this paper on Listener Preferences for Different Headphone Target Response Curves by Sean Olive et al. We also used a Sennheiser HD800 as a control to verify that they measured similarly in this setup to those shown in Seal Olive et al paper.

d. Off-site by a third party: To validate our measurements, we had the headphones measured at a 3rd party facility, THX. Their measurements were done using 45CA, the ear simulator was a 43BB IEC 60318-4, (low-noise version using a large pinna).

e. In ear probes on real people: Etymotic Research ER-7C Probe Mic System, series B and customized plugged ear probes.

Comparative Measurements
An average of 10 measurements were used and a 12th-octave smoothing was applied. All measurements are raw and no compensation curve was applied. Measurements were at DRP (Drum Reference Point).

160413_Blog_AudezeMeasurementsExplaination_Photo_Img2

Graph of LCD-4 on various setups at Audeze vs Tyll's HATS

As you can see from the different graphs, there is significant variation in measurements after 2kHz. You can also see around 6kHz, there is almost a 5dB difference. After 6kHz, you can see the response slanting downwards.

The characteristic dip between in the 4kHz-5kHz region is due to the earpad shape and prevents our LCD series from appearing bright.

Below we have plotted a graph of LCD-4 vs HD 800 on a GRAS 45CA (control measurement). We did this to make sure our measurement setup yielded similar results to what was published in the paper by Sean Olive et al.

160413_Blog_AudezeMeasurementsExplaination_Photo_Img3

As another comparison, we have measured LCD-4 using in-ear probes at DRP and also measured HD800 at DRP (as control, red dotted line) and we have plotted the average of 5 measurements each made on the left ear. Note that we have included only the range 250Hz to 10kHz as per the probe specification.

160413_Blog_AudezeMeasurementsExplaination_Photo_Img4

Sometimes removing the ear canal resonance gives a different perspective on the measurements. Below we have plotted the in-ear measurements we made at the ear canal entrance of one of our engineers for the LCD-4 and another planar headphone. Compared to the other headphone we do not see significant deficiencies in the treble region here either.

The Audeze Sound

  • Unparalleled low-frequency extension due to our large planar drivers and the good seal established by our headband and earpads.
  • Smooth through the midrange due to excellent control we have on our diaphragm.
  • Close to ideal mid and upper-treble extension starting at 6kHz, slanting downwards from there to 40kHz.

The LCD-4 fits the above sound signature. As we mentioned earlier, our goal is to design headphones that emulate a good sounding speaker in a well treated room. For this reason we've not included any diffused field or free field compensation curves and have stuck with averaged raw responses in this article.

Conclusion
Measurements across multiple systems are significantly different due to the differences in the pinna, concha geometry, inner ear model and how the earpads and headphone interact.

Measurements are valuable tools to get a rough idea of how a headphone may sound, but they are not a substitute for listening. We see significant differences in measurements made on similar systems and also in our in-ear measurements (especially in the treble region). We encourage Innerfidelity readers to audition the headphones for the best possible results.

Thank You

Audeze thanks Tyll and Innerfidelity for being gracious enough to allow us to post this response and share our thoughts with you. We have offered Tyll access to all our measurement systems, labs and resources if he choses to visit Audeze and compare his measurement system to ours. Our goal is encourage more conversation around this topic and contribute back to the community.

Editor's Note: This subject is of great concern to me. My recent survey with a series of polls was intended to get a coarse read on InnerFidelity reader's ability to utilize plots. My take on the polls and reader comments is that the measurements from my system does provide utility, which is reassuring.

However, the above article does show that measurements from one system to the next will yield differing results, and that doesn't make me very happy. I've contacted G.R.A.S. and B&K and have requested a short loan of their head and torso simulators to take measurements here with my rig.

I intend to measure about five headphones on each head, and will look for general differences between the heads. I will also look at how the various heads measure the differences between the headphones to see if the interaction between headphones and the head is consistant. I've got a couple of other ideas, but need to make arrangements before I can tell you about them.

I will continue to use my measurement head for headphone measurements—switching to another head would invalidate comparisons to previous measurements, and....well...I don't have a spare $25,000 to pop for a new one. But it's possible experience with other heads may lead to valuable observations for future compensation curves.

I continue to feel the LCD-4 is too elevated in the top octave and stand by the subjective opinion expressed in my review, but I am deeply thankful to Audeze for taking the time to write their opinion and comparative measurement results. It will lead me, and InnerFidelity readers, to an interesting learning opportunity. Thank you.

I will be appending a copy of this page to the LCD-4 review as a manufacturers comment.

Very much looking forward to reader's comments on this issue.

COMMENTS
jackork's picture

In the end it's all about what you personally hear and not about measurements. Final judgement should be made with listening, so if you want to buy a headphone then don't rely too much on measurements... at least i think that, I know that a lot of people will disagree.

lip_lip's picture

Well if the whole article is about measurements, and measurements being part of design process, even Audeze would like you to know that measurements are important. I guess they disagree too, quietly. How else would you find good headphones?

I believe what they have tried to say is give the headphones a chance. Any measurement above 10k has less repeatable measurements on different equipment. And don't judge headphone quality on objective data only. Like Audeze, graphs help you pick what to audition, then get your hands on them before the action of buying them..

Johan B's picture

Plenty of your measurements do not show the dramatic dip between 5Khz and 8Khz. The LCD2 does not, The EL8 does, HD800S does not. So it is not systematic. So if it is an error than it is an intermittant one. Easy check is to re-measure a headphone that does not show the dip and then after that re-measure the LCD4. Next step is to test a different LCD4. Done.

Tyll Hertsens's picture
I've measured two different EL8 Opens and they measured the same basically.
Johan B's picture

So if you now measure a headphone without the dip ... then it proves ability to register signal between 5Khz-8Khz. Then this may indicate that this is not a bad measurement. For the sake of consistency don't change the basis of measurement. It cuts both ways. The Audeze measurement may be the reason for the way their headphones sound.

Audeze_R's picture

Johan,
There is a bigger point we are trying to make here. In addition to measurements on Audeze GRAS Kemar head we also measured with a completely different system from GRAS and also measured using a 3rd party system at THX. There are significant differences between the systems. We also measured with in-ear probes that sit near the eardrum and correlated them to our Kemar measurements

Johan B's picture

Indeed there are choices to be made to see which system of measurement hits the accuracy tick box. On the other hand if signal is allmost absent in a measurement then this is quite a big thing. Is there is nothing to measure at this specific interval. As a drummer I like listening to many Pat Metheny albums (Kin <--> is a great example) ... the Cymbals are recorded so well .. compared to live cymbals .. they are in the everywhere in the frequency range (yes even the sub range) ... but dominant in 5Khz-10Khz interval and I would hate to lose information there. You can argue about model responses .. the sound pressure of cymbals is enormous.

Johan B's picture

http://www.drummerworld.com/forums/showthread.php?t=66957

Is there any instrument covering such a wide range? Lets use the right source to assess the right response.

detlev24's picture

Yes, this is a widely spread issue. As I posted short on your "CanJam at RMAF 2015: Pioneer SE Master 1 Flagship Headphones" topic, also AKG has similar concerns regarding your K812 measurements.

Evaluation of loudspeakers is a lot easier; you need an anechoic chamber and as a manufacturer (should) strive for a neutral - flat - frequency response. Of course, neutral is not natural and this is one reason why there is calibration to the B&K 1974 target curve or similar (see HARMAN, Bob Katz etc.) at the listening position (which is highly influenced by the surroundings/room).

To cause even more confusion: Sonarworks, e.g., translate their measurements into a "Perceived acoustic power frequency response" graph; on their individual calibration report. I got one from a (not my) Audeze LCD-2, which looks interesting. Frequency response curves on their plugin, which get corrected to the B&K 1974 target, look different again. Sure, Sonarworks protect their know-how so the scientific interest in the tools they apply will lead to a dead end sooner or later. :)

Tyll, if you are interested in the LCD-2 chart: please let me know how I can pass you the link/file directly.

Best regards

detlev24's picture

Maybe measuring at the ear drum is not the way to go, since individual anatomic conditions will alter frequency response already before. Not to mention further individual processing up to and in the brains.

It could be more transparent to imitate the same acoustic space calibrated studio monitors create in a treated listening room - and try to measure this equivalent before it hits human anatomy. As far as I know, loudspeaker measurements are not performed at ear drum level either; the natural and accurate calibration result is nevertheless assured and accepted as a reference by most audio professionals.

ivanhagberg's picture

I really think this is the way to go. I have never understood why measurements of headphones are made inside the ear. Why not outside, as for loudspeakers. Imagine the result if you measured all loudspeakers at the eardrum - what a mix of "standards" would not have emerged then. But now there are some pretty trustworthy standards of how to get good loudspeaker sounds. The confusion for loudspeakers are gone, as far as I know.

If headphones are supposed to "sound like a pair of good speakers in an acoustically treated room" - why is measuring done inside the ear drum and not outside?

detlev24's picture

Consequently, this also means that there will probably never be a standard for in-ear monitors, since in this case measurement at the ear drum makes sense. Therefore, the anatomy of as many human ears as possible should be reconstructed digitally (with thousands of samples) and merged to an average dummy profile which then could be used industry wide for measurements; and constantly be improved. In this case the result might be natural and accurate for those whose anatomy is near to the created average - and for everybody else, individual EQ could be the only way to go. Of course, customizable DSP in portable devices would help a lot.

As for over-the-ear headphones: the acoustic space that is not yet influenced by individual anatomy should be measured, I think!

Best regards

Tyll Hertsens's picture
The ears of IEC spec heads are designed from sampled OF MANY, MANY Ears to make it the shape of the human average ear. The same is true of the head size. This actually leads to a little problem in that the head is slightly smaller than a male head as female heads are smaller in general. Many times headphones are made more for a male head size, and the dummy is a bit too small, sometimes leading to fit issues.
Tyll Hertsens's picture
It's a good thought, but practically speaking, I also measure IEMs with the rig so I need one with an ear canal. I could use a separate rig, but I'd rather use the same rig for everything. I also do isolation tests, so IEMs need to be in a head for that.
detlev24's picture

Tyll, thank you for your replies. I was speaking about how I would expect measurements to be made in general (with in-ear and over-the-ear headphones separately treated), not pointing at your work. As long as there is no standard, please continue at least as good as usual, since your work is highly appreciated! It helped a few times to choose the right gear and of course, listening tests were the final instance before buying [not necessarily needed for loudspeakers^^].

I do not think Audeze, AKG or other manufacturer's measurements are better than yours. I never was able to hear what is advertised; being used to calibrated studio sound. I found Sonarworks to do a great job on my HE-500, although my loudspeakers are not calibrated 100% to the B&K 1974 curve, but to the variation Bob Katz and others use.

As for AUDEZE, if your goal "is to make them sound like a good pair of speakers in a well treated room", why not send a few LCD-4 headphones to, e.g., Sonarworks to see what their individual calibration results will be? You would get a correction of one - amongst many audio professionals - respected third party and this certainly would help on further improvements.

Btw, there is a new review of the K812 with measurements by Sonarworks; which basically reflects Tyll's findings: http://sonarworks.com/2016/04/test-of-the-akg-k812-studio-headphones/

Best regards

detlev24's picture

"Frequency response curves on their plugin, which get corrected to the B&K 1974 target, look different again." is to be ignored!

There was an error in creation of the individual calibration report. After correction: individual calibration report = frequency response curve on their plugin. [Both in terms of "Perceived acoustic power frequency response".]

ivanhagberg's picture

In the end it is all about what YOU hear. Close your eyes, don't look at the measurements for a while. Just enjoy the music, the fantastic ability we have to transform sound waves to a conscious sensation that music really is. A truly miraculous gift. It is rewarding to see that manufacturers get serious about how to make better headphones. Either you focus on the technic and enjoy that pathway or you trust some valued reviewers subjective statements, like Tyll's, understanding that there are tricky deviations regarding measurements and finally enjoy your expensive headphones and gears. Don't forget to listen to the music!

TMoney's picture

It seems to me like Audeze has been getting the same criticism on their treble on each release since the LCD-2 and have not made solving those issues a priority.

Perhaps Audeze should hire some golden-eared consultants or take advantage of the extremely experienced enthusiast community in the LA area and beta-test as other SoCal companies have been doing.

Beagle's picture

That measurements are the biggest cause of listener fatigue?

aamefford's picture

I would be very curious to see tests of the same headphone on 2 or 3 identical systems. They should report identical results, but I wonder.

Tyll Hertsens's picture
I am going to try this, and I expect there to be difference, especially in the treble.
Aufdemaury deus ex machina's picture

The real question is which measurement equipment should we use to get the most accurate results if this is at all possible to actually identify the most accurate gear to use.

To my knowledge the Audeze El8's are the only headphones that dip that significantly in the 5-8khz,
other headphones such as AKG offerings don't show this at all, so I don't see much of a problem , along with that I had an opportunity to demo them before being privy to tyll's measurements and I already thought that they sounded dark and had a steep notch up top.

I have no doubt that there will be differences in measurements across different facilities and gear,
though I'm hesitant to think Tyll measurements could make such faulty errors, that being a colossal 20db + notch in the treble.

Tyll's Hd800 measurements are quite consistent with the graphs Sennhieser provides to customers,
his LCD3F sr 27xxx are also consistent with my own personal Audeze SR 27xxx graphs that I received from Audeze themselves. I can provide both of these to anyone inclined enough to ask (Via Email)

Regarding the actual shape of the concha, pinna and inner ear. In all cases, person to person these all very any way in IRL, so I fail to see the point Audeze tries to make as if Tyll's headgear is inaccurate, because there are already variables. As long as his head is modelled to best match the average/majority of people, I and many others will most likely be content with this, though I'm not saying that there are no improvements to be made.

Any who, lets not point fingers and be patient for Tyll's test when he experiments with his loaned
GRAS and B&k gear. The point I want to emphasize is that despite the differences across systems,
Tyll's measurement are well accurate enough and are very credible, particularly because he is a third party as well. In my eyes we shouldn't waver his Database as irrelevant.

ultrabike's picture

that different couplers and measurement rigs may need their own compensation curve.

Has there been any consideration to an approach similar to what was done by Stereophile in 1991? (free space measurements of mids and treble to emulate anechoic conditions):

http://www.stereophile.com/content/grado-hp-1-headphones-measurements#4f...

Perhaps a coupler that minimizes echo and reverberation. Like a mini-anechoic chamber?

gandhisfist's picture

Own a pair of EL-8 closed headphone and using a tone generator I can definitely hear the big null around 7kHz on the headphone. Doesn't appear there is anything wrong with Tyll's rig to me. This was a pair that I RMA'd in October last year and have confirmed is the newer revision of the EL-8 Closed back.

Phoniac's picture

Finally someone used his brain! Indeed the best way to check all this is to use your own ears while playing back sine waves up and down. I used this method, described in a headfi-org thread, where the author even offers a set of such files for easy playback, and after a few minutes found it to be quite easy to adjust my EQ to make the phone very linear. The result is a mirror of the measurements found online, but still adapted to my own ears/hearing. Everyone should do this and get away from the 'no EQ' mantra, as that is biological nonsense. Everyones ears hear different and need individual EQ!

GumbyDammit223's picture

Can you provide a link to the headfi tread?

GumbyDammit223's picture
Phoniac's picture

Actually it was this one:
http://www.head-fi.org/t/796791/the-most-reliable-easiest-way-to-eq-head...
The whole thing is much too complicated in description. Simply load the stepped sine waves (download link in the first post) into whatever player (home, car...) and start jumping through them. One quickly gets a feeling of too loud or too low level regions. Sweep and pink noise are not usuable with the same efficiency for me.

Phoniac's picture

I am amazed by the pics at the top of this post, because most of them will fail with phones that have pads formed after the natural skull. Menas they have a higher height behind the ear at the lower position (sometimes known as 'angled'), and a reduced height at the front area (cheek bone?). You won't see this very often but that is how it should be like. So with both fitted, providing a perfect seal for a human listener, the heads and ears in the pictures will fail in bass when measuring lower frequencies, as the seal willbe not perfect. Even with the head top left, as the raised cheekbone is too far away from the ear.

How comes the measurement industry ignores the reality completely? If heads would better represent a typical listener's head maybe more manufacturers already would offer phones with perfect, 'anatomic' sealing?

hanshopf's picture

If I'd be a manufacturer of headphones I'd ask myself which headphones on the market, past or present, already came nearest to my goal of accurateness. I'd measure this headphone and take that curve as a starting point to work with. Let's chose HD600, which has proven over time to be the most widely accepted headphones by audio pros. This surly has been for a reason, despite them all having different ear canals and head shapes - and minds... . Of course every LCD is technically far superior to the outdated HD600. But my LCD-X makes a violoncello sound like a double bass and makes it hard to keep apart both groups in an orchestral recording. In real life both groups can be easily distinguished. Therefore: HD600's tonality is still more accurate than any of the LCD's (I listened to all of them including LCD4 with my own music in a shop where I also could compare to Stax 009, 007, L700, L500, 005, 002, Ether, Hifiman He1000, He500, He560, Oppo PM3 etc etc).
Yes, I finally bought a LCD-X, because they were very convincing despite their failures tonality wise compared to HD600 (let me say that I found the Staxes to be less accurate, all of them, as well as the Hifiman's, all of them as well...).
I then sent my LCD-X to Sonarworks for measurement and calibration. Their target curve is nearest to HD600. Want to have your headphones sound like HD600 without that one's flaws? Send them there. Unfortunately their plugin has some deficits either. I found that it sounds only really good when bypassing their "avoid clipping" mechanism and use Izotope in Audirvana or even better let this make the DAC (Hugo in my case). Furthermore I had to add a very slight subsonic filter to the Sonarworks calibration (10.0Hz, 0.98Q, 18db), because bass was too strong for my ears. Nevertheless I believe that their target curve is extremely accurate (more than anything I ever heard on headphones), but only with maximum latency.
Anyway: I learned that it must be extremely difficult to build accurate headphones. But why then not go ahead with work from a point where others already got it right?

BaggedMilk's picture

http://ko.goldenears.net/board/index.php?mid=GR_Headphones&search_target...

Another rig that has the same dip as Tyll's rig.

BaggedMilk's picture

Albeit on this rig the canyon is somewhat covered up by the excessive smoothing Goldenears uses, it's still quite apparent it measured a big ol' notch out of the FR there.

Audeze_R's picture

Thanks for the link. Yes we have looked at those too. And this is yet another example of how another system measures differently. Sometimes it is easier to notice differences if they are overlaid. Here is a link (https://goo.gl/MiflEf) where we have overlaid Goldenears measurements on Tylls and our own (a bit crude, but the best we can do without access to the actual data).

Depending on the measurement setup used, notch near 7-8kHz by itself is quite common in planar magnetic drivers and you can see those in many measurements from different manufacturers at innerfidelity and in our own measurements of LCD-4 and our in-ear measurements too. As we have shown with our LCD-4 measurements, it is the region between 4-8kHz where we see a lot of difference between different setups.

We have stressed multiple times, our goal is to show different setups measure differently (including our own) and it is very hard look at a frequency response and know how a headphone would sound in the treble region. We are not trying to show that one system is right or another is wrong.

There are some exceptions to being able to use measurements while comparing headphones in the treble region. For example, if the earpad design remains essentially the same, and you are familiar with the sound signature of one of the headphones (say one from the LCD series), then you could use the relative differences between frequency responses made on the same system to better judge how the tonal balance would differ. This is something we use a ourselves during our design process to identify regions that need improvement.

castleofargh's picture

it's good to repeat again and again how different measurement tools do not give the same results. people keep on making silly comparisons between innerfidelity, headroom, golden ears, etc. taking one headphone from one website and another from the other website. people what R U doin? stahp!!!

@audeze, if I understand it right, you're saying that Tyll's measurements sometimes show a dip and sometimes don't? if so it would of course be great to find the cause.
could it be as simple as a positioning thing? as some headphones change a lot and others don't when moving around.

also you talk about a convolver of sort, you developed to simulate things when researching for a new headphone. why don't you guys develop a simplified consumer tool like that? with a few settings, maybe a few room simulations too(I hate headphone panning). no headphone is going to be perfect, and that would add so much value to a headphone, including some objective improvements I'm guessing.

Audeze_R's picture

Yes we are trying to show different setups yield different results and there is no right or wrong measurements.

The reason for the dip on El-8s has a lot to do with the earpad shape and the how it interacts with the pinna. Tyll's measurments have consistently shown the dip with our El-8 series in spite of repeated measurements and across different samples. It is the way the El-8s measure on his setup and there is nothing wrong with the measurement.

Though we were expecting to see some differences in measurements both ourselves and Tyll were surprised by the extent of variation across systems.

Regarding your question about developing a user friendly 'convovler', yes we are working on one and you can find a review of the system we are developing here (http://goo.gl/SBZ6ru)

Tyll Hertsens's picture
I do 5 different FR measurements at different positions on the head, and I rarely see to much variation in the 4-8kHz area. LCD4 plots are quite stable in the 4-8kHz area with movement. So I don't think that's the problem. I do agree with Audeze's statement that it's probably the ear canal and mic diaphragm impedance etc. that causes the differences seen in differing heads.
neo's picture

I would love to hear about Audeze's poor QC too and their failures: LCD-3/2. Surely, they would like to 'respond'

GumbyDammit223's picture

We have some pretty cool tools to make real pretty pictures with all sorts of numbers. They are, in my opinion, the first filter to use when trying to find a headphone to purchase (assuming cost is not a major contributor, or has already been applied). After that, it's finding samples to personally audition, or the opinions from people you trust. If you're lucky, you have a local store you can go to and listen to your potential future purchase. More likely though, you go to places like Inner Fidelity to find reviews of what you're interested in to help sway your choice one way or another.

I respect Tyll's opinion since he's been doing this for so long and has so much experience with so many different products. The measurements, while appreciated, I take with a huge boulder of salt, because there is such huge variability in them. I WANT to read about his personal opinions on the headphones! That's why I'm already thinking about taking the plunge into a pair of Ether C's.

Unless we're all the same model robots made by the Acme Robot Company, I'm going rely on what Tyll and others SAY about a product and not put too much credence into the computer printouts, no matter how pretty they look.

My 2¢.

Tyll Hertsens's picture
I think my take-away point from Audeze's article is that it's important not to become biased by measurements. I ALWAYS compare what I hear to measurements weighting what I hear as most important. I do regularly find headphones that differ in sound from the measured indications. At that point I always weight what I hear as the primary impression.
markus's picture

I use narrow band pink noise to find lowered and accentuated regions in the treble area. Download it here
http://www.moshier.net/pink.html

Audeze_R's picture

Here is the link to the measurements published in this article in pdf format (https://goo.gl/4fOUD7), some of the wordings look fuzzy in our images.

Seth195208's picture
Audeze_R's picture

Thanks for the link. This one we have not seen before. And here is the overlaid version. Green is Tyll's. (https://goo.gl/hVIkUz)

sludgeogre's picture

As a chemical engineer I deal with this all the time. It is not important for a system used to evaluate many devices to give "true" measurements, only "consistent" measurements that can be used as an apples to apples comparison, which is why you saw some utility with the measurement interpretations that users gave. The measurements will make sense as a whole, but individual measurements compared to other measurement systems will vary and many times cannot be compared at all. Your measurements need to be evaluated against all of your own other measurements on other headphones, not against any other measurement system unless a correlation between the two has been proven or if both systems have a well characterized measurement capability as measured against a true standard, which is something you can't do in audio.

tony's picture

I began my GM Corp career in the Instrumentation Department, I was responsible for calibrating & maintaing the Torque and Pressure Measurement Equipment, some of my other lads maintained the Audio frequency Gear ( Bruel & Kjaer ).
I understand the difficulties in getting reliable and repeatable test results. ( this is one of the reasons I was so impressed by your Big-Book of headphone data that you passed around at RMAF a few years back ). I've seen Audio test data results like this over the years.
I'm not at all worried about your accuracy levels ( isolated as you are up in Montana ), even if you were in our GM Tech. Center your results would and could be challanged by various Engineering points of view.
The critical element is the interpretive skill of the observer, of which we have a "substantively acute" person in you!
We also have the "part time" efforts of Senior Katz.
A proof of both of you is in how you disagree on things. You two are right out of the Research Labs in the way you observe.
Probably, (for you to advance) you'd need to join-up with someone like AtomicBob up in Seattle area or accept a position in one of other Audio Research Laboratories and surround yourself with like-minded co-reasearchers, you don't have the money to invest in the "Standards" like Bruel & Kjaer all by yourself.
Still, I find you entirely useful and a bit of a Prodigy, guys like you don't exist in most hobby areas.
I would encourage you have a look at a few Alex Dykes reviews, this man is the A+, top Reviewer, of all time, notice the absence of measurement gear and the high levels of accuracy he achieves.
I value your measurements, I'd even buy the printed book
version and have you Autograph it. Each year updated. Very much like the Ametuer Radio Handbook the ARRL have been putting out since the 1950s.

Your data and thoughts are gonna disagree with other's thought and data : "when two people agree, only one is doing the thinking" !

Tony in Michigan

zobel's picture

When two people agree, they both can be doing the thinking. :)

tony's picture

You might be referring to "Consensus" averaging or getting close is close enough.

Going into details will bring out vast disagreements.

Even the Periodic Table is still being debated.

Subjective agreements change like Women's Fashions.

A funny off-topic concept: We now have the two leading Presidential Primary Candidates who are both "Independents", neither one is accepting PAC money. Geez, can't we agree on anything?

Nice hearing from y'all,

Tony in NY, soon back to the Sunset Pig in Colliefornia.

the Bronx is Bern'n

zobel's picture

I totally agree! (can you tell I'm pulling your leg here?)
As far as I know we both were thinking too... :)

tony's picture

I didn't realize that you were pulling my leg.

I called a GM Research Audio guy who kinda suggests that the only Consumer Audio folks with capable measurement staff are : Shure, Harmon, AKG and Sennheiser with an emphasis on Shure and Sennheiser!, don't know about any Asians but it's likely that some exist. Atomic Bob up in Seattle should be included but he's not Consumer Audio.

I suspect that we shouldn't expect our smallish Manufacturers or capable reviewers ( like Tyll ) to have the careful measurement instrumentation. We can hope that our Tyll can provide deliberate interpretations of what he suspects from data he manages to collect and corelate.

We got lucky with Bob Katz who "owns" Calibration Standards to support his work and findings.

Consumer Products Manufacturing outfits understand that the most popular products are not the ones that have the best measurements. (probably applies to wives, too!)

And despite a Capable Emissions Testing Industry, VW still managed to fool everyone with their published Diesel emissions fake data.

I accept the concept that "Accuracy" in measurements can only half-state a problem but it does help us create a better result.

This debate has been on-going since the 1500AD Italian fella starting developing the science of Physics, nearly got him killed.

Measurements has had the useful result of reducing the "range of acceptable error" in our lives.

Tony in Michigan

ps. those 4K Samsung TV's look beautiful

aamefford's picture

I think slidgeogre hit the nail almost on the head. Actually right on the head. I can't wait to see the results of 2 or 3 "identical" systems measuring the same pair of headphones. I think the results may show pretty significant variations. This would mean direct comparison can only be done on the same measurement apparatus. Comparisons on different ostensibly identical rigs would not really even be valid. Hmmmmm. As I said, I am looking forward to Tyll's results of this test.

Audeze_R's picture

There seems to be some level of consensus that headphones measurements made on the same system can be compared objectively, but how do we know for sure?

If you measured headphones A and B on a setup-1 and also measured them on another setup-2, will A-B (the difference) be the same on the two setups? could it be almost the same for some headphones but not others? What would one choose as a reference headphone?

This is the reason why in our measurement examples of LCD-4, we choose HD800 as a control to see if the difference (between FRs) we see in our setup is the same as what is seen on other setups. We feel, that the difference may not be the same, but our analysis is by no means complete. We look forward to Tyll's comparisons and analysis.

gatucho's picture

I this post hits the nail exactly. I work in scientific research and it is widely known that different experimental setups will produce different results; EVEN if they are capable of detecting the same specific phenomena. For instance, equipment X may detect event A as high gain (10dB), whereas equipment Y as a mid gain (5dB) (I use this example to speak in terms we can relate). This situation brings out a deceitful conclusion. "If I adjust equipment Y by +5dB then I can compare its measurements with equipment X"

Wrong, you cant and many years and dollars have been lost by this fallacious assumption. The thing is that measurements are a map from the set of the physical world phenomena to a simplified quantitative representation of these phenomena. However, the inverse is not true, unless proven theoretically or experimentally, a measurement WILL NOT map to a specific physical event. This is something like the following: someone takes a 2d photo of a 3d world. If perfectly accurate, then only one 2d photo is possible given a specific perspective. However,note that other 3d world configurations may be reduced to the same 2d photo. Thus the 2d photo does not represents any specific 3d world.

In sound scientists and engineers use several theoretical tools, such as SN theorem, to make this bridge between frequency-domain measurements and real world phenomena. However, it is clear for me that given the complexity of sound interaction with human body elements, a single (or a pair) of measurements will not reveal all the complete history.

Moreover, as Audezeze puts it, it has not been properly verified that a single data point adjustment would be enough to calibrate Tylls and others measuring rigs so that equivalent interpretations can be obtained from both data sets.

It is very risky to attempt to making extrapolations from known phenomena to unknown phenomena. For instance, as Audeze is telling, if you know the sound signature of a headphone very by experience, then the measurements or change of these could be use to predict specific changes in the actual sound of the device. However, to expect that the same sound quality change occurs when a similar measurement feature is found in a different headphone would be ill advised. Both measurements, although apparently similar, could be very well measuring entirely different physical events, with entirely different subjectively perceived effects. Even more, if someone makes this assumption this could, and will, effectively create a "expectation bias", so than when the said hypothetical listener actually hears the headphones he could very well end up "hearing" the fallacious thing.

Lastly, I should add that Im a measurement and objectivity believer (I must); however, I have also learned that the world is sufficiently complex to also being overly critic of any conclusion supposedly "supported" by measurements, engineering, science, etc, because there is always certain level of human factor when assessing the objective data and we humans are very easily drawn into our own dogma.

PD pls excuse my poor english skills

Visigoth's picture

Tyll, I would love to see a new graph plotting the Fostex TH900 with this new "head". It may vindicate what many have been saying all along, in that it's one of the best headphones out there.

Stefraki's picture

I've been making this point for a while, specifically regarding IEMs where I feel the effect is amplified hugely - the differences in our ear canals should make a significant difference to how we perceive sound.

Everyone understands that the same speakers in a small room will have more bass and in a large room will have less. In a headphone system, the 'room' is the space from the baffle to the ear drum. The engineers can control the baffle and the ear pad, they can't control the size of the ears, the oiliness of the skin, the amount of hair that will be in that space, or the size of the ear canal. All that will make a difference.

I think that with full size headphones the differences are probably, to borrow Audeze's term 'significant but not dramatic', but with IEMs? I am quietly confident that they are both significant and dramatic.

If the 'room' is essentially the ear canal, and as you can see from tip sizes - one person's 'room' can be up to twice the size of another's, we should be seeing very large changes in perception.

The IE800 is a good example. I have huge ear canals, many large tips are just too small for me. I used to perceive the IE800s bass as pretty much ruler flat and got a thousand people telling me I was wrong, that it was hugely boosted. But my ears don't lie. I heard flat bass while people with smaller ear canals heard way too much bass.

I think we need to accept that measurements are useful only in reference to our experience. If I see an IEM with 'flat bass' on the measurements, I know full well it is going to sound like lean bass to me. Someone else might know it will sound flat to them. There is no use in the head fi discussions where people brandish measurements to 'prove' a model has too much or too little bass. We are likely all subjectively perceiving something quite different due to our quite different anatomies.

zobel's picture

By this I mean; Take several subjectively flat sounding headphones, noting perceived differences in frequency response, and then build a measurement device that will produce a representation of those responses? A flat line from 20Hz to 20kHz would be the target 'curve' for the best headphones. No craziness above 2kHz, just a nice smooth, flat line. Since measurements below 1kHz now are good with the dummies, the work would be in coming up with something that can measure the frequencies above that in a meaningful way as well, perhaps taking the model ear completely out of the picture?

Due to the variation from person to person, (as seen from dummy to dummy) unfortunately, this too, would represent an average SPL/Freq. graph. YMMV.

hanshopf's picture

If measurements in general showed to be reliable enough to help improve the frequency response of headphones by EQ, doesn't this mean that the measurements already at least should be good enough for manufacturers to tune their headphones more accurately than we are used to?

sunnydaler's picture

Tyll's measurements in treble area match well with my impressions of HPs I've tried. At least, the over-ear ones.

Nomad_Soul's picture

I wonder, how much does a GRAS 43AG cost? Well out of my price range for sure, but now I'm curious.

Nomad_Soul's picture

Woops, just finished reading and noticed the mention of 25k down at the bottom. Definitely pricey.

thefitz's picture

I wonder how these variances affect the reviews. One might run their listening suite, make their observations, put together those charts, notice some strange behavior, and "go back and listen" to see if they can hear those peaks and notches (which now they magically will). Hopefully the reviews are written in their entirety before measurements are taken.

thefitz's picture

Based on some of the comments above, wouldn't it be wild to compensate the readings for the frequency response of the HD600, instead of the Harman headphone response target curve? That way, any "variation" imparted by the measurement tools would be present in the HD600 curve, which should mitigate the differences between systems. A pair of cans that matches the Harman target curve 100% accurately might not read that way on Tyll's systems.

Bob Katz's picture

I think that presenting headphone responses as deviations of another known phone's responses could be educational. It could be very useful. Spreadsheets make that possible.

derbigpr500's picture

...as people think. The brain compensates the peaks and dips that might be introduced. It's impossible to explain this here, because the readers need extensive medical knowledge or hearing neurophysiology, but the brain is "calibrated" to hear things a certain way, and as a person grows up an as the brain adapts to the physical world around it, the shape/size of the outer ear included, it learns to compensate. It's evolutionary crucial as well to be able to hear things correctly. It might not be precise down to 1-2 dB, but it's precise enough that you don't really have to worry that one person is going to perceive the sound of a headphone drastically differently than another person, given that both are healthy and don't have some major deformations of the outer ear.

Audeze_R's picture

Good points on the effect of brain smoothing out and of the reason for 3rd or 6th octave smoothing is often applied to get an idea of what one may actually hear, also the reason why Sean Olive's research showed some level of consensus on a preferred target response and also why many of the IEC standards are defined for 3rd octave smoothed responses.

Something that may need some research however is the effect of headphones on audio perception and variation of perception over a diverse population. Our ears and brains have evolved and are tuned to being in a open space. Audio put out through headphones are not 'natural' in that sense and are restricted to a far more constrained space where and where the outer ear can come in contact with the headphones or be squished or physically modified.

HeaMe's picture

Hi, Guys, I'm big fan of LCD-collection from Korea.
Tyll,Audeze,Golden Ears and Speakerphone's rig follows different protocol(Itu-t and IEC 60318-7, ITU-t, IEC 60268-7). So it is natural that the measurement shows some differences. I think LCD-4 measurements' difference is not that critical(I bet Tyll would not have changed his opinion, using the other rigs - as all of thoese LCD-4 measurements shows similar features.) Over 10kHz is not guaranteed area and 2~10K looks quite differnet and of course, it looks the difference goes beyond the tolerance but Audeze's QC is not that excellent. So i think there can be effect both effect of different rigs and product variations between LCD-4s.

But the weird thing is Audeze's EL-8 measurement with GRAS 43AG. EL-8's 7K Notch is obvious. Most EL-8 users' ears including me and 3 measuring web site (GoldenEars, Speakerphone, Innerfidelity. How it can be? It;s Mystery.

HeaMe's picture

I made some typing error. How can I edit it?

Audeze_R's picture

We are not denying an existence of a notch at 7kHz on other measurement systems or that they are wrong, many headphones have it to varying degree. And we are glad it got this much attention. As you can see, with LCD-4 measurements, the differences are less dramatic yet significant. El-8s measured on our GRAS KEMAR head does look different (from 43AG), with a notch at 7kHz.
If you try to play tones at different frequencies and make a mental picture of FR, you would be surprised how much they can differ from some of the measurements and also how much variations you could notice beyond 7kHZ with adjusting the position of your headphones (extent of variations depend a lot on the earpad geometry and position of the drivers etc.)

Bob Katz's picture

I took Tyll's painfully-derived measurements of my personal LCD-X phones and derived a Harman-style EQ for these phones. The result is excellent. HOWEVER, and this is a big however. I went through a bunch of iterative steps which involved my ears as a final judge. Eventually I arrived at my interpretation of the Harman curve which I applied. But how do I know that the result I obtained is a) the truth (that is, what the Harman curve should be), or b) how Tyll's measurement system responds and therefore not applicable to the general public or c) a degree of personal taste.

I think that for the foreseeable future, what designers of headphone equalizers should do is something like Audeze has done, that is, the same as I did, to try to make the headphone follow the Harman curve as closely as possible. Note that they also tried to verify their measurement system's response by seeing how closely it followed the Harman results with a set of HD800's. Then follow the Harman compensation curve with a general overall baxandall bass and treble filter to make the final tweak.

My system can send a Dirac pulse through any digital equalization system and replicate the amplitude and phase response, so in the end I have a single net curve that I apply to my LCD-X cans. The systemI recommend which you can use is: Acourate and Acourate Convolver. Together you can design a filter set that you can apply in playback either with Acourate Convolver, or with the convolution engine available in JRiver. So you can be a happy camper, with your own custom filter. One that works FOR YOU, not an approximation.

Bob Katz's picture

Now what do I suggest that Tyll do in his measurement system? Given that I did have tremendous success designing a filter using Tyll's measurements, but only after quite a bit of iteration, I suggest that
Tyll present his measurements, in addition to the raw ones:
a) with 1/6 octave filtering (for psychoacoustic reasons)
b) showing deviation from the Harman curve as a difference function
c) showing the same deviation inverted as a filter
d) making the filter coefficients available to users (at extra cost of course!) in an Excel spreadsheet form

Also remember that unit-to-unit variations are still quite large for the majority of headphones that are out there. This not only complicates and possibly explains the differences between Audeze's observations and Tylls, and also points out that your mileage can really vary. That's why I suggest adding a Baxandall-style equalizer as a final tweak to any store-bought compensation EQ.

Audeze_R's picture

Bob, thanks for those very good suggestions. We may do this ourselves for our headphones to help our customers with EQ.

Regarding unit-to-unit variations, we measured the review sample on our KEMAR HATS before we shipped it to Tyll and we have shared those measurements with Tyll too. All the measurements we have shown in the article are from a different sample. So, to help with the analysis, we have plotted the the sample we used to Tyll's review sample here (https://goo.gl/ls1PZT).

tony's picture

This is as far over my head as the Space Shuttle Orbiting.

I am glad there are folks, like you, doing this level of work, gives me hope of even better transduce devices and improved music reproduction systems.

I'm home on break from pouring gas on the "Bern Baby Bern" movement. I'll be at the Detroit Symphony Orchestra & Lise de la Salle performance of Rachmaninoff Piano Concerto #1, this Saturday ( being Broadcast "Live" on the DSO Live internet site ). Then I'm heading to Ft.Meyers & Venice Fl., on a hunt for a Wintering home. Should I order Sunglasses?

I could see you taking a Teaching Position at a University.

Hmm, Dr.Katz has a nice "sound" to it, doncha think?

Tony in Michigan

detlev24's picture

Could you do a quick comparison of your EQ to the one by Sonarworks? The LCD-X model is now available in their free trial; I know it is an average curve and probably not accurate enough but it certainly would be interesting to read your opinion.

Thank you for your many valuable contributions!

Best regards

Bob Katz's picture

Come and visit me in Orlando! Anyway, I never got my final degree... I went to work in the industry for the next 40+ years! Nearly Everything I learned about audio from books, from mentors and hard knocks!

tony's picture

Hello Bob,
Thank you for the offer. I'll be on an loosely organized tour with one of my brothers, we'll be exploring Naples to Tampa, two weeks, including the inter coastal waters. It's our first get-together in over 50 years.
I'd love to experience the world you live in, although Audio is only a companion for me now-a-days. Funny thing is that I wasn't unhappy with the 78s of my youth, the 33s were much better and I was delighted with them and the moving coil phono carts. Now, my modest 16/44 gear is completely delightful, leaving me with slight motivations for improved reproduction capability. Even portable stuff is wonderful to my failing hearing.
Still, a tuned Revel system could be draw enough but I'd be happy to only hang out and philosophize on World histories.
Formal Education hasn't much significance . Real World experience ( disciplined work ) and personal study is what yields success ( and maybe greatness ). I imagine you as a qualified mentor to young minds yearning to get some earned confidence in Audio Recording. I think a College would cherish what you bring to the Classroom & Lab. Of course, Colleges actually pay a living wage and have a certain amount of prestige. You might even have access to an endless stream of "free" interns and a possibility of "Contract" work from the local Bureaucracies and a designated "Parking Space" with your name on it.

Well, thanks again for that offer!, Tues in NY will reveal how much energy my group will deliver in California and the rest of the Primaries, if NY is a success we'll be turning on the after-burners and going "full-out", I'm tired now and need some time off, by Convention time I'll be exhausted and Bern'd-Out!

Tony in Michigan

sunnydaler's picture

Maybe the very thin diaphragm is the culprit. doesn't it make the driver more sensitive to acoustic load and also hard to control tension (resulting unit-to-unit variations)?

sunnydaler's picture

very thin diaphragm is also prone to make popping, crackling noises (pastic bubble wrap sound) because it isn't stable enough to endure the chamber pressure.

arnaud's picture

Unless there is a significant issue with Tyll's compensation curves (e.g. If the compensation curves he received from the maker aren't truly matching his actual head), I don't see the point to criticize his measuring gear.

Raw data is extremely affected by the coupler / head geometry and skin impedance. Calibrated measurements should be much less sensitive as pointed earlier in the thread. As such, I actually do not see the use for publishing results out of couplers like gras's because there is absolutely no way to properly calibrate them... They're great for convenience of use and comparing designs, not to present absolute levels imo.

As for the LCD4 measured response, why not address the elephant in the room? That is any ortho is bound to ring like mad somewhere between 8 and 12kHz which forces the maker to severely absorb high frequency energy of the driver (hence the typical dip relative to other phones in the octave below that wall).

To me, while I appreciate Audeze response, it does not add much information, feels more like an attempt to blur the lines. E.g you just compared apples and oranges with showing non-compensated curves from a bunch of rigs. Luckily, some of them don't show the mid-highs ortho dip as much as others (although the comparison to the hd800 makes it obvious), what does that prove exactly?

The main positive I see here though is discussion on the topic of headphone measurements :-).

Arnaud

Audeze_R's picture

Here is the DF compensated measurements for the sample Tyll used for review (https://goo.gl/53qSqr). The DF compensation curve for Tyll's rig and for our KEMAR HATS have been applied. Tyll shared the DF compensated measurements with us and we used the one provided to us by GRAS.

Most of the article was about the extent of variation between measurement rigs in the frequency range 4kHz- 8kHz and that even if a compensation is applied, the measurements do not line up (true within systems we use too).

Yes we wanted a discussion started on measurements and interpreting measurements, the intent was not about the correctness of different rigs. We are thankful for the overwhelming response this article has received.

In an earlier response, we also posted a link to an article that compares HATS from different manufacturers (http://goo.gl/CY3zz8) and how much tolerance the IEC standards allow.

Dreyka's picture

You should of kept to the science because a large chunk of this innerfidelity post reads like PR damage control.

The LCD-4 is the results of years of research and development to ensure the most natural and transparent sound possible.

Poor move to include this garbage when it is only distracting from the point you were making.

arnaud's picture

Audeze_R, thank you for the instructive link to the univ. paper with very clean HRTF measurements and comparisons for various heads. My take from this is that band averaged DF curves are actually not that different (Fig. 32) and actually even free field HRTFs as specific heading aren't that different (Figs. 37-40) below 6kHz (frequency and / or test samples smoothing is mandatory above that range due to high sensitivity of response, Tyll already does this since the beginning).

Most of the gain effects in the HRTFs in 2-6kHz range are due to concha / ear canal geometry as discussed by Tyll previously ( http://www.innerfidelity.com/content/headphone-measurements-explained-fr... ). These gain effects get partly normalised out by the compensation curve used for the head when looking at compensated headphone measurements. For over-the-ear headphones at least, although the HRTF gain is heading specific, I do not expect very drastic differences in compensated measured headphone response curves between similar shape heads as done in the paper.

Actually, in your link above (https://goo.gl/53qSqr) comparing Tyll's DF compensated measurement to yours, it actually seems to prove the point: besides the few dBs variability between measuring heads, the trend is still obvious. There is a clear 15 dB notch in the headphone response at 4-5kHz with a very strong resonance at 7-8kHz which is more or less visible depending on the head used due to directivity and what not. It seems that correlated with what Tyll heard and he expressed that in his review.

Please correct me if I am wrong, but this peak at 8kHz is not the result of uniform motion of the diaphragm but an illustration of "pocket modes" occurring in any ortho headphone in this frequency range. e.g. it's either a bunch of "small" pockets on the diaphragm that all start to sing together at around the same frequency as the bending wavelength equals double the traces spacing. And/or it's the equivalent in the acoustic space inside the magnets slots. e.g. you are acoustic standing waves forming in the thickness of the magnets which dramatically amplify the radiated noise at that frequency. This is how I understand the "ortho wall" and what makes ortho drivers requires very careful damping (unfortunately often resulting in severe roll off of the response below / above that peak because it's hard to acoustic treat a narrow frequency range).

cheers,
arnaud

Audeze_R's picture

Our drivers have a near ideal response when measured using an IEC baffle. There is no reason for us to suppress a frequency range on LCD-4. The dip seen in the 4-5kHz range is due to the geometry of our earpads.

arnaud's picture

Thanks again for the comment. I read before (it was a panel at a socal event or something) that you use IEC baffle to design the transducer and then the enclosure / earpad system around (very good engineering approach I feel). Would that be too sensitive then to share some of these IEC baffle measurements?

On a marketing standpoint, if I go by your comment above, I assume you could show reasonably flat near field SPL response of your drivers which isn't a bad thing to promote. On my side, regarding the thick magnet layering in front the diaphragm, I have hard time imagining there is no such acoustic resonance effect (that justifies the design of phasors...) so you're proving your point with these extra measurements is, to me at least, much more effective than asserting that the LCD4 is the fruit of several man years of R&D without any other details.

Anyhow, much appreciate being to exchange thoughts :-)

arnaud

Audeze_R's picture

Arnaud, you have a valid point. Why not show the measurements made using IEC baffle?

We are unable to publish the measurements due to their propriety nature. Just publishing the measurements will not be of much use unless it is accompanied with details explaining the physics behind the measurements and how they look.

We understand it would be hard for anyone to just trust our words without evidence to support said claims, so we have shared these measurements with Tyll and requested him to respond.

tony's picture

For the second or third time you have me intrigued with your designs.

I'm fascinated with your Cipher Cable! A cable with a DAC and EQ system seems an wonderfully useful solution.

I'm accepting Tyll's acceptance and approval of the device's performance which is probably enough to justify buying.

I wonder if you will offer this cable for general use? and what you might say about using it with other outfits headphones.

Tony in Michigan

Audeze_R's picture

Tony, We are glad you are interested in our designs. Having access to a powerful DAC with a tiny foot print provides us with an opportunity to do EQ and possibly other DSP functions in the future.

However, we do not plan to make them available for general use in the near future.

castleofargh's picture

maybe to exclude the possibility, Tyll could send you back the headphone for you to check if nothing changed? (transportation hazard?). it's a long shot, but still another possibility.

ADU's picture

I've been doin my best to follow all the different threads in this discussion while finishing up my taxes.

When Tyll originally posted his LCD-4 review, I asked whether he thought the response in the treble was "by design", or the result of some limitation in the technology. I'm still not sure what the answer is. :) But it's fairly clear to me from looking at the Audeze graphs that there is at least an _element_ of design coming into play in the LCDs' treble response.

Audeze appears to be targeting a constant slope in logged space in between 3 kHz and 20 kHz, which is not the typical treble response for a neutral or audiophile headphone. The treble tends to get progressively steeper (on average) in the higher frequencies on most of the better sounding headphones, rather than following a constant slope in logged space. This is undoubtedly why Audeze and similar planars come across somewhat "depressed" in the mid-treble, and a bit over-emphasized in the upper trouble to reviewers who are used to a more curve-shaped falloff (changing/steepening slope) in the treble.

The Inner Fidelity and Golden Ears graphs may be exaggerating the mid-treble depression though (at least in Audeze's eyes), because the HATS systems they use for in-ear measurements have a natural depression or notch in their treble response between the resonances at ~3 kHz and 9-10 kHz, which is around the same area being depressed by Audeze's constant slope approach.

I'll try to post some graphs in the next few days to illustrate a little better what I'm talking about. But hopefully most of the above makes sense.

Kudos to both IF and Audeze for having the cojones to bring these issues out into the open for some discussion, and also to the members of this site for providing such excellent feedback. These are the kinds of things which make this site such a pleasure to be part of.

Maybe some other mfrs will following Audeze's lead on this, and it will result in even better-engineered and better-sounding products for us all to enjoy (even those of us who can only currently afford $150 cans :) ).

ADU's picture

"Upper trouble" should've been "upper treble" above btw.

thefitz's picture

https://www.audeze.com/sites/default/files/images/EL8_Open_KU100_FR.jpg

There it is, that 6kHz dip on Audeze's own website. What are they posting here?

To be honest, if you're going to have a dip somewhere, 6kHz is one of the best spots. I'd take a dip over the 6kHz stab found in the HD700 any day of the week.

pf-THX's picture

Hello all. pf here from THX. We also have been researching measurements on different heads and different microphones. We also have seen similar results.

We measure with:
B&K Hats.
Kemar.
GRAS 45CA.
Custom Flat plate with B&K 4191.
Our own ears with insert microphone at DRP.

We are also very interested in what is going on.

X