Big Sound 2015 Bob Katz EQing Headphones with the Harman Target Response

Today I performed a long awaited experiment of creating an EQ for two pairs of headphones from my understanding of the raw frequency response measurements and the Harman target response curve. The acid test was whether or not Bob Katz felt the EQ good while listening to tracks he's intimately familiar with.

I will write a more detailed explanation in future, but the basic process was this:

First I looked at the Harmen target response curve, which is the black curve on the plot below.

<# some text #>

I then printed out a large version of the frequency response plots for the headphone we were trying to EQ. In the case below, it's the Sennheiser HD 800.

BigSound2015_BobKatz2_Photo_HD800Plot Because the Harman response curve is measured at the ear drum, we compare it to the Raw frequency response plots, which are the plots at the bottom of the graph. You can ignore the compensated plot at the top of the graph.

I then superimpose the Harman plot (by hand) over the HD 800 raw plots, aligning the two graphs at about 150Hz. Once the plot is drawn satisfactorily, I then measure the difference between the two plots at numerous points along the plot. It is these differences that make up the compensating EQ for the headphone.

Bob used the parametric equalizer in the J-River software suite. Using a parametric EQ is a bit of a trick, and it took us about an hour to develop the basic compensation curve for the two headphones we adjusted—an Audeze LCD-X and Sennheiser HD 800. After the initial stab at EQ using the Harman curve, Bob would spend about 15-20 minutes gently tweaking the curve using various pieces of music with which he was very, very familiar. In general, his tweaks were on the order of 2dB or less, and were generally reducing the offset suggested by the Harman curve. However, the general shape and overall Harman correction was retained.

One particular area of interest to me was the gentle rise from 200Hz to 1kHz. This rise is rarely found on headphones. It exists due to the acoustic gain of the head, neck, and torso when in a sound field.

<# some text #>

The plot above shows the ear drum response when in a flat sound field. The total response is the black curve (somewhat miss-) labeled, "Ear Resonance." If you look carefully, you can mentally sum the head gain (Plot 1), torso and neck gain (Plot 2), and a bit of the first ear canal resonance (Plot 5), and you will see that it results in this rise from 200Hz to 1kHz.

When I have heard this rise on a pair of headphones (NAD VISO HP50, and Focal Spirit Pro) I felt it did a much better job of presenting the human voice with all its overtones, and chest, throat, and nasal resonances and sounds. This better balance of the voice causes it to be very stable in image depth (not smeared in depth) and removes any sense of veil or excess distance to the voice.

Well, I simply couldn't be more pleased that Bob Katz agreed. He found the EQ a truly excellent starting point for further tweaking by ear, and he had never even thought of including this rise and felt it did, indeed, improve the tonal balance.

More, of course, of Bob's feelings can be heard in the video.

Equipment List

Front End
NAS - Synology DS414 ($479)
Renderer - Aurender W20 (~$17,600);
Digital Distribution Amps - Four ATI DMM100 Digital Matchmakers
and one DDA212-XLR digital audio distribution amp ($1450).

Power Conditioning
PS Audio, two P10 power regeneration station ($4999) and four DecTet conditioned plug strips ($499).

DAC/Amps
AURALiC Vega DAC ($3499) and Taurus MkII headphone amp ($1899)
Simaudio MOON Neo 430 HA ($4300 w/DAC).
HeadAmp GS-X Mk2 ($2800)
Schiit Ragnarok ($1699) and Yggdrasil ($2299)
Burson Audio Conductor Virtuoso ($1495 w/PCM1793; $1995 w/ESS1908)
Woo Audio WA-234 ($15,900)
Antelope Audio Zodiac Platinum DSD DAC, Voltikus Power Supply, and 10M Rubidium Atomic Clock. ($13,045)
Apex High Fi Audio (TTVJ) Teton ($5000)
Eddie Current Black Widow ($1248)
Violectric V281 ($2299)
Bakoon HPA-21 ($2995) current output headphone amplifier.
KGSSSRE (Kevin Gilmore Solid State Special Reviewer's Edition E-Stat Amp ($Unobtanium)

Headphones
Sennheiser HD 800 ($1599)
Audeze LCD-3 ($1945) and LCD-X ($1699)
JPS Labs Abyss AB-1266 ($5495)
Stax SR-009 ($4450) and SR-007 ($2350)
HIFIMAN HE-1000 ($3000)
Mr. Speakers Ether ($1499)
Enigmacoustics Dharma (~$1200)
Audio Zenith PMx2 ($1398)

Cables
Digital cables by AudioQuest.
Cable complements for wiring entire systems will be from: Nordost; JPS Labs; WyWires; Cable Pro; AudioQuest, and Cardas.

Accessories
Headphone stands by Klutz Designs

COMMENTS
tony's picture

I've been working with the Harmon Curve, Tyll's Measurements, my Audiologist's Test Results and a Vintage BSR consumer 10 Band Eq.

I though that I'd discovered or uncovered something special, something like a rare treasure.

Then I learned that the people like Phonak do this stuff All-the-Time!, for them it's a Standard Operating Procedure.

Now the Two Professors are Piling-on.

Guess I'll order some HD800s, what am I waiting for ?

Annnnnnd a EQ from JR River.

Who'd though Summer would be a great thing for Headphones, it's turning out that way for me.

I'll tell my wife it's for my College Courses!, she'll never know any different, the HD800s kinda have that School look of intelligence, not that flashy Celebrity look of the Beats stuff. Hmm, wonder how School Lectures on Egyptology will sound on the big Sennheisers?

Tony in Michigan

Bob Katz's picture

If i get any older, I'm going to have to include my audiologist's measurements, too ;-). Yeah, Tony, I'm a genuine fourth generation baby boomer. Anyway, personally I recommend you stay away from the HD800. I could never make an EQ that would satisfy. Either it would sound harsh on some material with brass or it would sound muffled after the brass-fix eq. This is because a static equalizer can't really fix a high Q resonance circa 6 kHz. It's a bandaid, not a cure. Let's find you a more neutral dynamic headphone to start with. But I guess not an Audeze because you don't like it's devil weight :-(. Those "enigma" phones... the ones I put 3rd in my eval on Saturday, look like a good starting point for an EQ for you, and I'm sure Tyll may have already published it. But get rid of that BSR eq... it's too crude to capture the fine curves that you need to apply. It took me an hour with a really versatile parametric. I couldn't even have come close with a 1 octave graphic. Why not use a player that can accept plugins, like the JRiver?

tony's picture

Well, Ok!

When I heard the 800s they were better than anything I'd heard. That was months ago.

I've since fallen in love with what I already own and am nearly ready to make that long commitment.

But I became a bit insecure about DACs.

I do love my little Odac.

That Antique Eq., it's a brief loaner from an old Professor friend of mine who is now Stroked out. He wants it back for it's little blinking lights ( lovely light show ).

I dug into the internet about the Plug-in you mentioned.

The Plug-in stuff is cheap. ( low dollars )

Of course it's Computer based, which I suppose is ok. I'm adjusted to Boxes with power switches, lights and such. But change can be a good thing ( I was able to switch from 78s to 33s without difficulty, I can do it again, despite what my Wife might say )

If all I'm doing is connecting the dots, the Eq dot is the only one missing.

I read about all these folks chasing gear, endlessly chasing, from piece to piece to piece.

I'm happy with my stuff, I just have to properly Eq.

You saved people plenty of grief if they can understand the concepts.

Tony in Michigan

ps. are you about to reveal "things" about the DIY gear?

Wildcatsare1's picture

Gents,

Attempting to match your settings with my stock HD800, Taurus Mark 2, Monarchy M22B (recapped), Amarra sQ+, and Tidal. While not able to eq 3kHz to +12 db, sounds a bit tinny, a boost of 6.3, with +4.5 at 29 Hz, and -3.0 DB at 6 kHz, sounds pretty damn good.

Looking forward to modding my HD800s and adding a balanced set of Norne Audio Draug 2. I would be interested in seeing various settings for modded and un-modded HD800s.

Down the road I would love to see a shootout featuring your wonderful array of DACs similar to what you are doing with amps in this installment. Perhaps throwing in a vintage PCM63 or 1704 based DAC.

Thanks again Tyl for your leadership in the advancement of all things Headphone related!

Best,

Kevin

castleofargh's picture

why would you want to add +12db@3khz?
the curve is to match with a raw measurement of whatever headphone you have(assuming it's calibrated correctly). it's not in any way what your EQ must look like. ;)

@Tyll
the EQ used in the video is DMG equilibrium. it may be used as a vst inside jriver, but people expecting it when getting jriver will be mighty disappointed ^_^.

I've been doing what you did for the last month with all my cheap IEMs using the vibro veritas little mic and harman's target.
when adding +1db somewhere doesn't seem to equate to +1db in measurement on an IEM(resonance or whatever I'm too much of a noob to rationalize it all), I go with Room Eq Wizard's EQ simulation, and the simulated curve does fit a lot better than my simple curve subtractions as an EQ.
all in all IDK if I get better results than what I did by ear before because also tweak a little in the end for my own preference, but what's for sure is how much faster I get to those results. that's what harman's curve brought to me.

anyway you've turned "big sound 2015" around for me, I was expecting not much else than FOTM testimonies diluted in hidden advertising. in the end I'm enjoying it big time(yeah I'm saying I was wrong but don't repeat it).

and thanks BOb for being yourself.

Rillion's picture

Were the annotations to the Harman curve (+4 dB, +12 dB, 200 Hz) ignored for the purposes of this study? Those numbers look off just slightly from where the black curve lies.

Rillion's picture

You can ignore my question above. I see that the annotations are just indicating a constant offset in dB. The 200 Hz line is still puzzling but that should not matter much.

Bob Katz's picture

Hi guys. Just flew back from Bozeman, and boy are my arms tired! But seriously I'm just recovering from the trip and in perusing Tyll's copious and comprehensive report I want to set the record straight that I used 3rd party eq plugin for the HArman response curve: Equilibrium from DMG audio. Highly recommended. In JRiver player. And iZotope ozone to dither the entire output down to 24 bit to feed my Prism Dac, which then went to three different amps which we could then compare. More later. :-)

I only have read a tiny bit of this copious report! amazing stuff

Rillion's picture

I just dialed in this Harman target into my AKG K702 (using Tyll's K702 raw graphs instead of the Senn. HD800) and it certainly adds some clarity to vocals. It does some other things that I'm not too happy about. I am not going to say more since I don't want to bias anyone's options before Bob and Tyll spend some more time with it and lend their expert assessments.

Hear The Difference's picture

I really appreciate the time and care that Tyll and Bob took in porting, setting up (and teaching me that "plumber's butt" doesn't just happen to plumbers), listening, documenting, and reporting the "Big Sound 2015" comparisons.

With so much good work done, I am the first to admit that some critic on the sidelines can easily envision what more could be done (and then complain).

So here goes:

* How were these particular headphones chosen (I suspect the answer lies in the early descriptions from Tyll);
* Why was Bob Katz chosen (nothing against him, but I know and trust Tyll... would liked to have had Tyll's take - I never heard of Bob Katz! I know... wash my mouth out with soap... EVERYONE (else) has heard of Bob)
* Could we have a similar test with a listening *panel*, not just one person, as a future step?
* It was said that not all headphones went through the same amps... that is fine, but it would be good to have understood which amps each headphone was tested with... was it the one they were sitting upon? How was that amp chosen for that headphone?
* Long initial discussion about R2R vs. noise-shaped delta-sigma DACs... preference for delta-sigma seemed to persist, based on sound stage, but it was commented that measurements of a difference had not been made... I'm intrigued by and in search of measurements that correlate with sound stage... what are they?
* Efforts to get to the Harmon Target Curve with EQ was great, as was the minimum phase argument (i.e., any minimum phase headphone can be equalized to match the frequency and approach the sound quality of any other minimum phase headphone, and nearly all headphones are minimum phase)
* Statement was made early on that unless you match frequency response ("tonality"), you need not bother about other specifications such as transient response and sound stage/imaging - I would have thought that frequency response would dictate transient response, and I don't know what measurements govern size/height/depth of soundstage (but would love to learn!)
* Statement was made by Bob that "none (of the headpones) met my standard," with subsequent discussion suggesting that his standard required the application of EQ to the headphone, and that he was going to take back the EQ curve created by differencing the headphone frequency response with the Harmon Target Curve and comparing it to his own more simply derived EQ curve. Looking forward to that.

In short, an intellectually stimulating read (or actually... video - would be great if someone did transcribe it as a read, though!).

Thanks!

Bob Katz's picture

Hear the Difference: Thanks for your kind words and perspecuity!

1) I don't know how these headphones were chosen. I'm sure it was a democratic vote between Tyll and Tyll :-). Listen, this is not a democracy. You don't get that from Stereophile or any other magazine. You do get a cumulative slow trend aggregation of the objective and subjective observations of reviewers like Tyll (and maybe me) coupled with the influence of enthusiasts like you. In other words, more like real life. Which isn't fair, but is a lot fairer than Congress since no lobbiest is there to influence the writer, maybe being a little bit gentle on advertisers, but Tyll told me to "let 'er rip" so I didn't mince any words. What you heard (and what you read) from me is my uncensored opinion.

2) You never heard of Bob Katz? Let me google that for you. http://bfy.tw/1hDF
I've been a headphone hobbiest since 1972, when I bought my first pair of Stax cans and modified my first Stax amplifier. Sunday I was able to evaluate Tyll's proposed equalizer using reference recordings which I personally have recorded or mastered and which I am intimately familiar with in some cases for 20 to 40 years. No other reviewer currently writing about headphones has the perspective of knowing intimately how the recordings which he uses for evaluation actually sound.

3) Listening panel? Would be nice. I'm sure Tyll will get to it in his copious free time :-). It's 5 hours by plane from Orlando to Bozeman, want to come?

4) R2R and sound stage. My reasoning for suspecting a soundstage and depth issue is more related to suspicions of monotonicity issues than stereo separation. There is also normally with R2R a frequency response issue which Schiit has reportedly compensated for with their own filter, so the Schiit R2R is a special case. Measurements? It would be nice. Some day. Some day. We are still it seems in 1950 when it comes to correlating measurements with actual perception and maybe figuratively in 1960 when it comes to headphone design, to be frank, so your desire for measurements is a little optimistic. All you would be able to do after seeing the measurements but not hearing the DACs would be to speculate. And that's what some of us reviewers do well --- even so, we are in the dark ages about which measurements make which sonic difference!

5) Which headphone mated with which amp. We moved headphones around, sometimes. The permutations are endless. Sometimes I suspect the choices were arbitrary. About the only thing I came out with after being at Tyll's regarding matching is "don't match a bright headphone with a bright amp!" And you need to know that: I felt that only three amps in the available selection were neutral and transparent. The rest were muffled and "fake warm" to my ears. You'll have to wait for my Big Sound article to find out which ones :-)

6) Yes, frequency response and transient response are related. Said Fourier. But I can show you two amplifiers which have equally flat frequency response, but have distinctly different apparent tonality (at matched output levels) and have distinctly different APPARENT frequency response. I don't know the reasons why and thus I have to depend on my ears for the final judgment.

7) Yeah, I'm looking forward to utilizing the Harmon curve myself! So much time, so little to do :-)

Enjoy!

Bob K.

Hear The Difference's picture

Thanks for describing the fact that unlike near any other reviewer, you actually knew what the test recording material sounded like at the recording session and had made many such recordings, so you had more of a basis than most for evaluating the headphone's reproduction of it. Thanks also for your detailed and thoughtful answers, and again, for the time and care you and Tyll took in preparing this report.

tony's picture

All good questions!

However the Larger Point is that "we" just got the green light on Eq.

When have we ever been allowed to ponder the possibilities of Eq?

Here we are, seeing, for the first time in my memory ( spanning back to the 1950s ) top Audio Professionals exploring the correction of Consumer Transducers.

We were only ever allowed to explore differences in Cables, Phono Carts., Pre-Amps, Amps, Loudspeakers, etc. all fresh out of the Box, no modding!!

We've just been empowered !

From now on, we'll be discussing our own Eq preference when we discuss our various gear. Our horizons have been widened.

Professors Bob & Tyll just performed an Eq demonstration on two of the finest headphones, they demonstrated (for us) the devices needed, their initial discoveries are only beginning to be revealed. This is profound!

The possibilities are that we can reach back into our closet for a back-up headphone and Eq the darn thing to learn how good we can make it.

I Eq'ed my personal HD580 Sennheisers to the point of not having a flaw I can detect. I can by-pass the Eq with a push of a button, the before vs. after is astounding! My wireless RS120 headphones Eq'ed surpass my HD580s without Eq. ( by a goodly margin ).

Geez, this is a total Win/Win/Win!

I was worried that my little Schiit Asgard 2 Class A Amp wasn't quite up to quality Amp levels, I was wrong, it's superb, maybe beyond superb. ( I'll learn more when my HD800s arrive ) .

Bob & Tyll didn't cover tons of stuff I'd of liked them to address but
how darn much can be covered properly on an initial visit of only a couple of days? I think they covered plenty for me to sink my teeth into and what they covered revealed questions I never would've had before.

Hold on, this is getting exciting.

Tony in Michigan

Bob Katz's picture

Flat is good when flat is needed. EQ is good when EQ is needed. I've heard some horrid sounding EQs in my lifetime when it was better not to EQ than to use the horrid one. If the equalizer you choose is not transparent enough, then the losses by adding the EQ will not be worth the addition of the EQ. I recommend using Equlibrium and to dither JRiver to 24 bits with Izotope Ozone. That will be the most transparent, accurate solution I can recommend. You can fiddle with freeware all you like, but I can't guarantee it if I haven't tested it or used it in the first place. And I can vouch for JRiver with Equilibrium EQ and ozone for the dither.

Bob Katz's picture

You might find another plugin that does dither quite well. Here's one that's not expensive: Try Psychodither from Stillwell audio. Set it to a very very small shaping or flat shaping, and to 24 bits.

tony's picture

Got it, I'll be internet digging tonight.

I'm not into that "Free" business, always comes with hidden costs.

I was looking at your LapTop display, I though it was exciting to see such capability.

Will I become an Audio God?, the power I'd have!

Tony in Michigan

ps. or worse yet, will I become addicted?

IgorC's picture

I've tried OT EQ on my HD800. But with one exception. 6kHz@-6dB was too much for my taste. I actually like a bit of crisspness. So I left 6kHz@-3dB.

EQed HD800 sound more like my studio monitors (JBL LSR5").
The sound has more body, but still more neutral/balanced comparing to original no EQ. That a bit annoying 6kHz peak has gone. Now rock/metal is much more enjoyable.

First impression: I like it

Need more hearing tests.

IgorC's picture

P.S. The source was Benchmark DAC1 USB with Zout close to 0 ohms

Rillion's picture

Yeah, 6 kHz is a trouble spot for me too. My comparisons with external speakers say that I need to cut it on my full-size cans. However, I'm on the fence whether or not it improves anything. Leaving the 6 kHz peak in adds a bit more "sparkle" to cymbals which I like. Also, listening to external speakers, I've found my hearing at 6 kHz to be extremely position-dependent. Of course, with headphones we don't have the luxury of moving our heads slightly to change the sound.

Bob Katz's picture

Igor, your reactions mirror mine with the HD800, even after the mod. What it is that after the mod when material with lots of energy circa 6 khz comes around, it's "less annoying". That's about the best I can say about the modified HD800. The worst I can say is it is schizophrenic: It has two personalities, depending on the dynamics and frequency contents of the music. So I never arrived at a satisfactory EQ for me, just something that allowed me to tolerate its issues a bit more. Listeners who are a bit more laissez-faire and just want a clear sounding headphone that's enjoyable and don't mind a little harshness now and then might get away with my first basic EQ before I started getting annoyed. I can publish that. Give me a little time to recover from Bozeman first :-)

Hifihedgehog's picture

I just want to say thank you to you, Tyll and Bob, for the educational remarks shared in these last two articles with their adjoining videos. I learned much new technical terminology, plentiful facets busting common audiophile myths and proper approach to critical listening.

I am curious to know if you would be willing to post Bob's equalization charts of the modified Harmon curves. I am very interested in seeing what an Anax-modded HD800 "should" sound like because Bob's observations coincided much with my own. If you could post these, it would be much appreciated.

I am also curious about the matter of crossfeed, as to which implementations you feel would be best, and how the headphone frequency response curve should be formed when listening with a crossfeed filter enabled. I personally prefer listening with a crossfeed circuit enabled whenever possible and I cannot imagine listening without one since hard-panned material--or, for that matter, any lower frequency sounds in the side channels--sound peculiar and off without it.

johnjen's picture

I too have played with EQ using the parametric EQ in JRiver.
I never did find a satisfying result until I sopped trying to 'fix' the peaks and valleys and instead adjusted only the low frequencies.

Currently I have added 3dB (which I just dropped down from 4dB, after other tweaks settled in) at 16Hz with a very broad Q of 0.17.
This EQ brings up all the bass from the subharmonic region to ≈100Hz in a very gradual and gentle way. It's also very easy to implement using the parametric EQ found within the DSP portion of JRiver.

And now I'm gunna go and add that 200 to 1KHZ emphasis to see how I can blend it in as well.

More fun science experiments… :thumb

Thanks for more food for thought…

JJ

detlev24's picture

It seems like pressure will s l o w l y rise on the DAP industry, to implement more sophisticated EQs and maybe crossfeed capabilities, as well. (In the future: + DSP)

I am curious to see who will be the "pioneer" and deliver a firmware update for present models and/or release a new approach of a DAP.

Let us see whether this topic helps to put aside discussions about cable sound etc... Well, maybe there is a need to EQ cables, as well! ;P

sgrossklass's picture

Actually DAP firmware with a parametric EQ and (meanwhile) two different crossfeed algorithms has been out for many years - it's called Rockbox and has graced my wee little Clip+' for about 3 years now. Its main advantage probably is a portable codebase (à la Linux), whereas manufacturers often tend to start from scratch once the underlying hardware base has changed significantly.

Right now the Rockbox devs could probably use a bit of help as major DAP manufacturers have recently fled the world of ARM processor architecture for (royalty-free) MIPS. Cheapskates. There's a fair few bits of ARM-specific assembler code for commonly-used functions which would have to be replicated in order to reach a similar kind of efficiency, and other such fun stuff. Interested end-users are advised to "come back in a year or two and see how far things have progressed".

mir's picture

I've had my HD800 custom calibrated by Sonarworks.

Lately they've also included an average HD800 calibration profile for their awesome software.

http://sonarworks.com/headphones/overview/

http://sonarworks.com/2015/07/krk-focal-and-sennheiser-headphone-users-r...

Tyll and Bob should definitely try this!

hanshopf's picture

I also sent my HD800 to Sonarworks in Latvia for measurement and calibration after having tried to eq them myself after the Harman Curve. My own results did not get even near to the breathtaking result of that guys from Sonarworks.
To the folks who are now experimenting with eq for the first time: never use a cheap tool like the one which come with JRiver! The quality is far away from professional standards, really. You will never get good results with something like this.
I found that my plugin from Sonarworks sounds clearly better when used in Audirvana. In JRiver it did not sound as brilliant. I have no idea why and tried many different settings. Another thing I found out: the calibration by Sonarworks was not perfekt in the sub-bass region. I had to implement a Subsonic filter because my ears startet hurting due to slight bass overload in the subsonic. Probably my HD800 was the first one they ever calibrated and did not listen to it long enough to consider the subsonic problem with these highly resolving drivers.
Therefore I bought DMG Audio's excellent Equalizer "Equality" and created a subsonic filter with these specs: Freq 10 Hz, Q 0.98, 18db. This fully solves the subsonic problem and touches the calibration by Sonarworks ever so slightly (you will only hear a difference after very careful comparisons).
My conclusion: HD 800's measured and calibrated by Sonarworks with added Subsonic Filter by DMG Equality and both plugins integrated in Audirvana Plus gives you an unbelievably clear result, which is miles or even lightyears ahead of any standard headphone. It's real studio quality!

mir's picture

Interesting, thanks for your reply.

I didn't need to tweak anything, nor had any problems with my calibrated HD800.

Guys at Sonarworks did a great job!

Bob Katz's picture

Hanshopf: I'm so glad that you're enjoying your Sonarworks EQ. I do intend to evaluate their EQs in the near future versus curves that I got from the master: Tyll himself.

Jriver EQ: That was a mistaken impression on Tyll's. I use Equlibrium. Equality can do what Equlibirium can do if you set the FIR impulse size to 32768 or higher. Unfortunately, Equality is a CPU resource hog and Equilibrium is much more CPU friendly. Dave Gamble never got to the point of applying that optimization to Equality. And since 65K impulse size sounds even more accurate, and it might slow down Equality to a crawl, I recommend you buy Equilibrium. Maybe DAve will give you a tradup allowance.

hanshopf's picture

... for your helpful comment! I had already increased the IR size up to 128K (thinking the more the merrier...). I did not notice any slowing down but will lock into that!

I was not happy with Sonarworks' calibration of my HD 800 from the start. I needed some weeks of trial and error to find out that due to reasons unknown and unexplainable to me it seems to sound clearly better when integrated in Audirvana than in JRiver.
Furthermore I do not run it at 100% of their ideal curve but ended up reducing their correction to 99% (ok, one percent less sounds silly, but this small difference changed a lot for me).
But my biggest problem was with bass. I needed a long time to identify the issue, which luckily was just a problem with sub-bass and nothing else and therefore could be easily dealt with. But I am so happy that I took the trouble to look into the matter, because I believe that I ended up in finally achieving studio-like accurateness in my home!
I am eager to read about your impressions when you'll get into this matter. Unfortunately due to tax reasons it will not be as easy to send headphones from US to Latvia for measurement.
Thanks for all your write-up's here. I read them with greatest interest and enjoy them very much. Big thanks also to Tyll to make this possible!

theoctavist's picture

"never use cheap tools like comes with jriver"

when I hear nonsense like the sentence above, it serves as a poignant reminder of just why the "HiFI" industry has become a farce, a laughing stock. That old tired warhorse/battle cry about price-performance correlation.. just no. no. you can't pick your own facts. One should consider whether or not his/her claims are bolstered by *DATA* and *SCIENTIFIC EVIDENCE* before making them.

audiofly5's picture

I am bothered by the fact that EQ'ing always looses significant information and coherence in the region being EQed. I never liked HD800 EQd. I have had positive results with EQing IEMs but not in high-end audio.

hanshopf's picture

This is not a fact... . You either did not use an highest quality Equalizer or did not make a decent correction or both. Fact is that without EQ you have significant loss of information in headphones, because overemphasized frequencies mask others and bring the sound out of focus. This problem is especially pronounced with the HD 800. With a good EQ you actually get much more information out of them than in their natural state. Believe me. My measured and calibrated HD 800's (by Sonarworks) are proof to that.

audiofly5's picture

The less 'schiit' in your audio chain the better. Using an EQ in your source is adding another punch of filters to the system. And filtering always has losses. The more bands you have the more problems you add.

If we look at the problem from the HP side then what you are doing with equalizer is forcing the system(driver and housing - assuming Proper amplification) do produce sounds that it does not want to do.
Examples.
1) If the HP has driver resonance at 2-3 kHz( common 2-3k region peak) then adding an EQ to this region is Not going to fix the problem. The drivers is still going to be resonant and uncontrolled in this region.
2) Boosting bass.
3) HD800 mods treat the cup reflections and some resonace tha causes the brightness and i doubt you can EQ this out. (otherwise we wouldn't have HD800 mods..)

I am not inherently against EQ'ing, just i havent heard great results with EQing. I have heard significant positive changes going from amp to amp, from a above average DS DAC to good R2R DAC. EQing IME has not added anything siginificant to the realism of the sound. It has only made it more unnatural.

hanshopf's picture

Theoretically at least I think you are correct. But in real life headphones today are far away from the standards in quality achieved with studio monitors. That means: headphones are simply not good enough.
But with high quality EQ and calibration after measurement of your individual headphones the gains in quality are much, much bigger than supposed losses by "adding another punch of filters" as you say. Personally I think that nowadays professional digital Equalizers like the one I use from DMG do not create any perceivable loss in quality. Really, they are that good! Wish I could demonstrate you my calibrated HD800...

audiofly5's picture

''But in real life headphones today are far away from the standards in quality achieved with studio monitors.''

It depends what you mean. Yes, studio monitors have full scale, life like soundstage. They seem to project a more full image, that can make you feel that speakers can be more resolving. Generally i would agree that speakers can provide a more enjoyable, immersive experience. But, then again i think that some values that HD800 has, are still unique and unbeatable. The HD800s can still do most life like trasparent 'weightless' sound. Faster transients. These characteristic mean that you don't hear the 'driver sound'. I think that is because speakers are hard to make that fast, and i am yet to hear a speaker that is More detailed than HD800. (i don't have any experience with other TOTL HPs..)

Bottom line is, great HP like HD800 are a beast of their own. They can eat some'phonitors' for breakfast imo.

hanshopf's picture

I understand what you mean! Thanks for your comments! What I liked to express was essentially that it seems to be much easier to build accurate speakers than accurate headphones. For the time being I still wait for headphones which are out of the box as accurate as studio monitors. Probably this will never be possible without EQ. But for me this problem has now been solved. I recommend anybody to have a try and let your headphones be measured and calibrated with a reference EQ curve like the one which Sonarworks developed.

hanshopf's picture

Theoretically at least I think you are correct. But in real life headphones today are far away from the standards in quality achieved with studio monitors. That means: headphones are simply not good enough.
But with high quality EQ and calibration after measurement of your individual headphones the gains in quality are much, much bigger than supposed losses by "adding another punch of filters" as you say. Personally I think that nowadays professional digital Equalizers like the one I use from DMG do not create any perceivable loss in quality. Really, they are that good! Wish I could demonstrate you my calibrated HD800...

tony's picture

Using Pro gear like DMG suggests you to be a Sound Professional.

I wish I could hear your HD800s.

I have it's orphan little brother the HD580s. I've applied Eq and am delighted with the results.

I am now contemplating the HD800 as a possible replacement for my old Sennheisers, however I'm so delighted now that I wonder what improvement would result.

For me, transducers have a "Singing" voice, I like the Sennheiser Voice ( best of all ), the HD580s sing like Pavarati, corrected my HD580s are also superb Baratone.

How much better are the HD800s?, how much better could they be?

Tony in Michigan

hanshopf's picture

no I am not a sound professional but a filmmaker having to supervise also the mixes of the music tracks in the studio. But I studied in a music school and believe I know quite well how music sounds before it meets the microphones. I just want to get this experience back as accurate as possible. As I have to live without a home studio at the moment I was eager to make my headphones sound really studio-like.
I was disappointed by the HD800 before using EQ and preferred HD600. But there is no doubt about the remarkable qualities of HD800. This wide soundstage in combination with an equalized accurate frequency response makes me really happy with listening at home for the first time in my life. So yes, they are much better than HD580 or HD600. But I would recommend to use them the way I described in a post above.
Eric (in Berlin)

tony's picture

Your English is excellent.

Still, German is your phrasing which I love and admire.

Thank you for the explanation.

I'm thinking about the Sennheiser Wireless Stage devices to use instead of my Wireless RS headphones.

Thank you for helping the Syrians.

We're probably at fault by our Iraq actions, seems we broke-it but don't seem able to accept responsibility.

Are you owning the matching 800 Amp/DAC?

I love the little Schiit Asgard 2 ( super cheap ) Class A but would consider the HD800 and it's matching electronics.

Thank you for writing back.

Congradulations on VW,

I'm GM and now in 3rd. Place. I didn't think anyone could top us but Toyota did.

Now years after the little VW Sedan arrived in the 1950s, VW tops everyone.

I owned a '68 Beetle and a '69 Westphalia Camper bus. Loved em both.

Tony in Michigan

hanshopf's picture

No, I used to drive my HD800 with the Phonitor 2 but ended up in plugging them just into the headphone output of my sensational Dac "HUGO" from Chordelectronics. I believe to get the most accurate sound this way and that adding an external amp in case of remarkable HUGO just adds colorations.

TMRaven's picture

That's not true at all. Even using the most simplistic and craptastic of EQs like iTune's built-in EQ will not do much to degrade audio quality. If you're hearing anything, you're being persuaded by confirmation bias. The worst thing that can happen with minimalistic graphics EQs is that they won't be as accurate as they could be.

There is no dramatic loss of information like you describe.

tony's picture

all good points.

Not be as accurate?, how could anyone know unless they created what they're listening to.

Loss of information is nonsense, at least for me.

I've increased the Freq.Response capabilities of my devices, I can hear things like Rumble that a non-Equalized headphone can't resolve.

Eq. is a total win for me ( a non-professional ), the Professionals have all embraced Eq. decades ago, it's a necesity in their world.

It's only the "Audiophile Nervosa" branch that poo-poo Eq.

Tony in Michigan

TMRaven's picture

Not as accurate as in you don't get fine control over which frequency is affected which way, like you would with a parametric EQ. In a a limited band EQ, EQ'ing down the 2khz band will do more than just affect 2khz.

Bob Katz's picture

Which equalizer did you use? Either it was :

a bad one or

You are having a problem with equalization. Then don't use it.

audiofly5's picture

I tried ToneBooster plugins. Isone(it has a built in EQ) and EQ plugin.

Tritone Hydratone, Electric-Q, foobar EQ and several others.

idesign's picture

Very ittle has been said about the Burson Conductor Virtuoso and I'm curious about your thoughts and how it works with the HD800?

Audiologist's picture

After reviewing both videos and the articles, I wanted to thank Bob and Tyll for all the time and effort. At the end, some questions come to mind. I hope Bob and/or Tyll will be able to clarify those.
Although, I really enjoyed listening to you both, I got a bit confused about Bob’s top headphone picks. At the beginning of the first video he said, “If you don’t get the tonality right, nothing else matters.” Yet Bob chooses Sennheiser HD 800 as one of his top picks. He then takes this headphone for equalization in the second video bringing the tonality closer to a Harman target curve, which took an enormous +/- 6 dB of correction within the most sensitive range of human hearing. At the end, Bob tells us that heavily EQ corrected HD 800 now started to resemble his personal reference sound. One can clearly see from corrected/uncorrected graphs that both original and equalized one will have to have COMPLETELY different tonalities. So the question that comes to mind is this. If Bob was concentrated on tonality in the first place, why did he choose the uncorrected HD 800 as one of his top picks?
Another question that I have is about Bob’s number 1 and 2 top picks that have frequency response plots posted on Innerfidelity also varying quite dramatically, and neither appears to have a tonality of Harman target response (confirmed by Harman white papers as well). In addition, in Bob’s latest article on DIY M3 amplifier, he wrote that when used with this amplifier, Audeze LCD-X headphone did not have to be equalized anymore, except for some very gentle rise above 10Khz. He considered this to be his reference sound then. At the same time, the equalized Sennheiser HD 800 brought to a Harman target curve sounded like Bob’s reference to him a week later. Well, when one takes the frequency response graphs for both Audeze LCD-X and the equalized Sennheiser HD 800, the curves would look quite a bit different. So the question is how could they both represent his reference sound.
I completely agree with Bob that there is a lot of subjectivity involved in the process. In my opinion, not only that, but also other factors. One of them, I believe is worth mentioning. That brings me to my third question.
Do you think that listening to a headphone after a headphone in a row without a break could affect the perception of the sound coming from latter headphones? I do think so. It works the same way for all of our senses.
Smell: We have to smell coffee beans between perfumes at a perfume store in order to fully experience the second of the following perfumes.
Taste: We are given ginger to eat between sushi dishes in order to fully experience the flavor of every dish.
Touch: If we hold ice in our hands, and then touch something hot, it will not appear as hot, unless we take the time to readjust and bring our hand temperature back to normal before touching the hot item.
Sight: If one looks at a TV set in a DEMO mode where both the brightness and the contrast maxed out, and right after switches to a calibrated neutral, the latter will look dull and unpleasant.
Sound: If one listens to a bright sounding headphone and then to a more neutral one, the sound of the latter headphone will be affected by the sound of the former.

So, Tyll, did Bob have breaks between listening to the headphones to fully experience the sound of each one? Also, Bob mentioned that he adjusted the volume of the headphones to his preferred listening level. Does it mean that your reference volume matched settings were not used? Considering how slightest mismatch in volume could affect our perception of overall sound quality do you think this could have affected objectivity of Bob’s judgments?
Similar questions could surely be addressed to Big Sound’s previous guests but they simply were not so hyper critical with judgments.
I once again thank both of you for taking the time!

Bob Katz's picture

Why did I put the uncorrected HD800 as one of my top picks? Because with most material that does not excite the 6k resonance, a MODIFIED HD800 sounds relatively decent and not too edgy or bright. I was surprised after reading negative reports of the unmodified HD800 that it sounded reasonably "flat", after which Tyll revealed to me that the HD800 I was listening to was modified.

Next, please note that I put the HD800, a modified one, in fourth place. That doesn't speak very well for the other headphone choices, I'm afraid. It was either feast or famine over at Big Sound. As far as I'm concerned it was a beauty contest with only four contestants. NOT TEN. The rest of the cans sounded like dirt to me :-(

What you are discovering about my comments about EQ are that simply: There is good, and there is better. The torso, head and neck response is what I did not compensate for in my own EQ. As Tyll said in the second video of my visit, almost no headphone has that compensation built in. And furthermore, almost no one is aware of what it can do. The effect of adding that portion of the Harman curve to the response is subtle, but important as the last touch in the EQ. In the future I'll try to include the Torso, neck and head response compensations and compare them with my previous super EQ and give you my comments. We are talking about subtlety at that degree and I will be the first to admit that my original simple EQ could possibly be bettered, now that I've heard the results when employing the head/torso compensation over at Big Sound. I did reduce the amount of it, but did not remove it completely. Also, since we applied the EQ manually by estimating an amount of difference on a pencilled graph, there is definitely wiggle room to tweak the EQ while listening to reference recordings. To repeat, I did not remove the torso compensation completely and was quite impressed with Tyll's/Harman's scientific contribution.

Rest: Rest assured there was plenty of rest between listenings.

Also note that as an experienced listener, I am very well versed in the subjective effect of listening to "dull after bright" for example, to color our perceptions. That was properly dealt with. O ye of little faith. :-)

EQs: As a professional recording engineer, I once did a comparison of various fine microphones as scientifically as possible using my clarinet and a friend. We matched the microphones in level, and as carefully as possible placed them exactly. The result was that the similarities between the microphones were quite surprising. I came away with the thought that when you start with very fine microphones which differ in only small ways from one another, their similarities will outweigh their differences. Same with EQ.... in the overall picture, small changes in certain ranges, if performed with a low q curve, will be very subtle. When levels are then matched thereafter, the ear/brain's ability to accomodate will be very strong.

In a future episode I will try to AB compare as fairly as possible my hand-tuned LCD-X EQ versus the Harman curve modified by me at their best. I know that the visual differences between the two EQs will appear large, but I suspect the sonic differences will be small and perceived as refinements, not as earth-shaking. That's the way you should look at it.

As for the level matching: I thought I made it clear in a very early episode of Katz's Corner (what, you didn't read that :-)) that it is impossible to loudness match two different models of headphones by any sort of measurement. All you can do is set it by ear. Tyll has written of his attempts to pink noise match them but I saw him hem and haw and ultimately come down to a first approximation. Only the ear can loudness match dissimilar headphones and even then it is subjective: If one headphone has a litle peak at 250 Hz and a dip at 10k and the other headphone is the opposite, how do you loudness match the two cans? By ear of course, and then it has to be subjective, since you know that the ear has distinctly different loudness sensitivity at 10 k than at 250. The science has not caught up with the art at that critical point.

So, the amps were originally matched for equal output, but when comparing two brands of headphones, they had to be manually offset from that. When comparing two different amplifiers, we endeavored to use the same headphone and patch from one amp to the other. So in those cases, as much as we had control, the amps were matched. I must tell you that it took me 15 minutes to level match two amps with a Fluke meter, so given that there are only so many hours in a day, after a while I took some shortcuts, and I'm sure that Tyll did as well. When a more formal shootout took place, however, we took extreme pains to match and extreme pains to compare the amps at matched levels with the same pair of cans. In the two days we had in common, that shootout only occurred once. Again, so much time, so little to do :-). We did the best we could.

Reader's expectations of reviewers being supermen who can work 24/7 are greatly exaggerated :-)

Hope this helps,

Bob

Earspeakers's picture

I don't know the backstory on Harmon curve you used, but are you guys familiar with diffuse field eq? This is what Stax used in the ED-1, and the Pawel HP-1. Some searching will provide details or I can dig them up and post.

I have several ED-1's (unfortunately no longer manufactured) and with Stax they're a game changer. I won't listen without one in the chain. The effect is most noticeable with say a cello, with it the instrument is focused, without it smears across the soundstage. It turns the Stax super-reality field to 11.

Though probably since I'm a classical musician I'm exclusively a Stax listener, I'd be interested to see diffuse field eq applied with other headphones.

castleofargh's picture

the harman curve was obtained through several tests that had the only purpose to find if there was a general preference for a signature. the starting point was a HD800 EQed to their own diffuse field(they picked a size for the room...). then people played around with EQ to find what they preferred. the result could be called a tilted diffuse field, it still very much follows the general tendencies, but is warmer.

elmura's picture

I've long stood by Equalising by ear using test tones. Rather than attempting to assimilate a generalised curve (Harman etc). Your own ears and some effort are the answer to getting a balanced, natural sound, for YOU.

As Tyll and science have stated, what we hear is somewhat unique. No generic synthetic model / dummy / microphone can replicate what you hear. Your ear shape, your head shape, your body size & shape, your age, your hair, and the condition of your hearing all affect what you hear.

By using test tones at a wide range of frequencies, along with an equalizer, you can compensate for both the headphones or speakers and your individual hearing.

It takes time and often takes a few attempts spaced by days to allow your hearing to "reset" but the end result is a customised curve specific to the headphones AND your hearing.

One caveat, don't EQ above 12.5kHz - it's unreliable, and your hearing sensitivity is usually not as great.

With my <1 year old Senn HD800, the EQ curve has a rising bass, and a depression in the 2.5kHz to 10kHz range. I later fine tuned it to limit how much boost was given to bass as the HD800 produced audible distortion and muddying of midrange. I maxed out +4dB at 20-25Hz and -4dB at 6kHz. The upper bass and midband were untouched (100Hz - 2kHz), as was above 10kHz.

Note that this is specific to my hearing and my sample of the Senn HD800.

equity's picture

It would be awesome if there were an EQ database with recommended settings to achieve the Harmon curve (maybe there already is?).

Jim.Austin's picture

Bob, simple question: Assuming you're going to EQ, what headphone would you start with? Clearly not the HD-800 (which I've only heard on the showroom floor at B&H and liked; I especially liked their spacious sound. I didn't listen long enough or seriously enough to detect the 6kHz problem)). Maybe the LCD-3 or -X, although you seemed to have some reservations there, too. Are there any 'phones out there you recommend with more enthusiasm? (I own HD-650s. Can I do better?)

Surprising that there aren't 'phones out there that aim for the Harman curve. AKG maybe?

Thanks,

Jim Austin (occasional writer for Stereophile).

Jim.Austin's picture

Bob - after writing my previous post, I discovered your "Katz's Corner" posts, which seemed to answer my questions. It seems you weren't as pleased with the LCD-X you brought home as you were with the loaner from The Cable Company. But you still seemed to like it a lot--and better than the other LCD's. (Maybe about the same as the Staxx, but that's out of range for me.)

I'm wondering, though, how you feel about the X's size/weight/comfort now that you've used them for a while.

Thanks,

Jim

Realsoundisgood's picture

Since the raw FR of the DT1350 is virtually flat, I tried applying the Harman curve and found it worked out well. I set the 3k level to +8dB rather than +12dB, but otherwise mimicked the curve in the article. All I've used is an iPad and Onkyo's HF Player app for my test. Hardly high end. Thanks for your hard work guys. I ran a small recording studio once upon a time, kept most of the control room audio system, and used my original masters for this little experiment. The result in the headphones is ~so~ very close to my masters played over my studio monitor system. I'm amazed.

chlyhne's picture

Hi Tyll. Many people would love to have their headphones equalized, but they don't know the response curve. Ofcourse innerfidelity comes to the rescue witht the measurement sheets, and that is fantastic. However, they are quite cumbersome to substract from the Harman target response curve. So i propose that you relase the data behind the curves in whatever format you have it in. I guarantee you that someone will soon have made a script that makes an equalization curve for the whole lot, and then you don't have to do the work. Best regards, and thanks for this! This was golden!

itsastickup's picture

The harmon curves are a)an average, b)have doubt cast on them. Wouldn't it be better to EQ by ear?

Meanwhile, NameckPiccolo on head-fi had a method that used continuous sine tones with a slider to run up and down the frequencies. I eventually realised that this method simply doesn't work because the tone is so intense that you have to EQ at a much lower volume, and end with completely wrong EQ for normal listenng volume.

Instead I've used fast-decay sine-tone midi-style piano. that way the brief note doesn't become so intense as to require lowering the volume. I calilbrate my listening volume with the hard-limter Spotify and a selection of music.

With the continuous sine tone method my bass was way to high, as the curves would suggest if volume was lower than listening volume. With this method it's presumably flat all the way along and is really amazing. the higher bass is no longer obscuring the lower bass.

itsastickup's picture

To ilustrate the above:

These are APO equalizer settings for windows PCs.

Filter 1: ON PK Fc 40 Hz Gain 12 dB BW Oct 1
Filter 2: ON PK Fc 587 Hz Gain 3 dB BW Oct .25
Filter 3: ON PK Fc 5000 Hz Gain 3 dB BW Oct .5
Filter 4: ON PK Fc 1500 Hz Gain 3 dB BW Oct .5
Filter 5: ON PK Fc 5510 Hz Gain -5 dB BW Oct .1
Filter 6: ON PK Fc 8168 Hz Gain -5 dB BW Oct .2
Filter 7: ON PK Fc 15000 Hz Gain -5 dB BW Oct 1

The piano I used is limited to about 3000Khz. The figures above that I guessed with use of a continuous tone and less dbs than suggested by that tmethod to account for the higher volume.

Augustus's picture

I would like to try it myself - the final EQ, if you wouldn't mind uploading it. Thanks!

X