Katz's Corner: Episode One

Editor's Note: Famed recordist Bob Katz is a headphone enthusiast...who knew?

Bob recently offered to write a series of articles for InnerFidelity readers on his recent adventures with headphones. I just couldn't turn down an opportunity for a peek at our favorite head-worn audio device through his venerable ears.

Please help me welcome Bob during his stay here at InnerFidelity in the comments below!

(For more info on Bob check out his website.)

Bob Katz's Corner: Episode One

Single or plural?

Should we say, "The Sennheiser", or "the Sennheisers"? "My headphone", or "my headphones"? One headphone or a pair? I'm sure Tyll has the language all straight in his head, but I'm still struggling with subject and verb matching. Let's just have some fun and enjoy the grammatical flukes.

The Great Headphone Shootout: Part 1
I became a headphone enthusiast at the same time I turned professional recording engineer. 44 years ago, I built a custom audio mixer in a suitcase and for reference monitors, took around my precious set of Stax SRD-3 phones powered by a souped-up Stax SRA-12S preamplifier. Although these phones require the standard bias voltage, I dared to exceed that voltage and got about 3 dB more headphone sensitivity and effective headroom from the amplifier. I turned the voltage up until just below the point where the electrostatic field made the hairs on my arms stand up—which brings new meaning to the expression "those speakers give me goosebumps!" But I never lost a set of cans (or any humans) that way; clearly Stax has conservatively rated their headphones. Plus, there's a high value resistor in series with the bias voltage on Stax phones which protects the user in case of a frayed wire or a short. The stator impedance and the diaphragm impedance are so high that you can parallel multiple Stax phones on the same amplifier without loading or losing level.


Here's a photo of me at a 70s recording session.

Later on I upgraded these to SRD-5 capsules. I owned three SRD-5 phones with a Stax splitter so three people can audition a recording at once with Stax fidelity, but in recent years I sold off two of those cans along with some passive energizers so I could afford to move up to pro models. In 2010 I purchased a set of new phones with gold-edition SRD-5 shells which someone had equipped with pro capsules, cables and connectors. We use these in the Studio B mix room as a secondary reference to the Genelecs, with a little bit of useful bass boost added in a preamplifier to assist the Stax's low frequency response. Even the well-cared-for forty-year old Stax phones sound and look like new, electrostatic cans don't seem to wear out as fast as dynamics. The high end just sweetens a bit with age as the coating slowly comes off the diaphragm. Just replace the earpads when they get hard.

I have dreamed of owning Stax Omegas ever since I heard my first pair while visiting fellow recording engineer Todd Garfinkle, and in 2010 purchased a custom-built KGSS amplifier from HeadAmp along with a new pair of Stax SR-007s, aka the "Omega Mark II". So the dream goes on: Stax has been my secondary monitor reference and listening pleasure center for over 40 years. I've also modified these Omegas by plugging up the port to tighten the bass as per Spritzer's recommendations at Head-Fi. I built a custom 1 dB/step attenuator for this amp as well.

My love for headphones parallels a love for good speakers. Studio A, the mastering room, has a state-of-the-art loudspeaker system whose response measures flat (within 2 dB) to below 20 Hz, including a target rolloff above 1 kHz. Even though the system sounds terrific without any aid, it sounds even better using Acourate Convolver, which is the only digital room and loudspeaker correction system that I have found that sounds completely transparent and improves sound with no compromise.


Above is a before/after graph of the fantastic frequency response improvement with Acourate convolver. This is the left front speaker measured using Acourate's psychoacoustic weighting. You can see that the high frequency rolloff is well controlled to a target I set by ear that has a little bit less high end than the native high frequency response of my Revel Ultra Gems. This high frequency curve results in a group of about 50 of my best reference recordings having a good-sounding high end that is neither too bright nor too dull.

Note how a formerly ragged response has been smoothed out to near-sonic perfection. And it sounds as good as it looks. As you can imagine, listening in this room is a pleasure, which I do for hours each day. The system is accurate, musical, impacting, mesmerizing, captivating...I expect the same performance from headphones, which I always compare with the sound of live music and to that of my reference loudspeakers.

I never liked the sound of the Stax models with rectangular diaphragms, like the Lambdas, they always seem to color the sound with a boxlike resonance that I think is due to the parallel surfaces. The large (circumaural), round Omegas throw a luscious, slightly u-shaped soundstage—you can actually get some imaging towards the front of the head, in contrast to smaller headphones which inevitably image at the sides or towards the rear.

Fellow headphone fans know that the state of the art in headphones has taken a tremendous leap with the proliferation of Planar Magnetic technology. Probably the best-known player is Audeze. I was intrigued to read initial reviews showing the large physical size of the Audeze diaphragm and praising its sonics, so I decided I should look into this new headphone. My friend, producer and musician Charlie Bertini (Applejazz Records) needed a new headphone, and was going to purchase the Sennheiser 650, but I convinced him he should go for the Audeze LCD-2, even though it exceeded his budget. Charlie is thrilled with his phones; when I checked them out at his home, they turned my head, even powered by his bargain Schiit Magni amp, which I also recommended. So I decided that I too should become an Audeze user. I would perform a shootout of the three main Audeze models and purchase the best one for myself, as a complement to my Stax, while performing a friendly contest to see which phone is the winner, the Stax or one of the Planar-Magnetic Audeze.

Getting Ready for Headphone Slutz!
It's very difficult for anyone to assemble the world's best headphones and enough amps to drive them for a shootout, but I set out to do it. The Cable Company has a wonderful deal. You can get a loaner of any headphone, cable, or amplifier for about 10 days and make your decision. The rental rate is 5% of the sale price, which is fully applied towards any purchase, so it's a win-win scenario if you're buying. You then send back the loaner and get a completely new piece. I rented the LCD-3 and LCD-X and borrowed Charlie's own LCD-2, which I planned to shoot out against my venerable Stax Omegas. In addition, I assembled a bunch of dynamic headphone amplifiers and a couple of premium DACs so that nearly all of us could audition at once. Seven of my friends and colleagues joined in the fun, all audio industry professionals—musicians, producers, recording, mixing and mastering engineers, acousticians and speaker designers. I call this group "The Headphone Slutz", and this series is about our great headphone shootout adventure.

Science or Art?
Our ears are wired for "louder". It's part of the way the ear/brain works. Since the louder of two devices almost always sounds better on an instant A/B comparison, it's a common sales gimmick to make the DAC or amplifier you are trying to sell be just a little bit louder than its competitor. Some DAC manufacturers join in the fray by setting the levels of their DACs a bit higher than the 2 volt standard established long ago by Sony/Philips. So, caveat emptor! Even 0.1 dB level difference can influence our perception of an amplifier or DAC. I always match the levels of every amplifier or DAC before listening.

The same issue applies with loudspeakers and headphones. Wouldn't it make sense to match headphones to the same sound pressure level (SPL) with a test rig before comparing them? Or would it? The fact is that when two transducers differ in frequency response (as most headphone models do), it is technically impossible to match their perceived loudness by any meter. Because SPL is not loudness! Let me repeat: SPL is not loudness. SPL is an objective measurement of sound pressure, but it does not indicate how the ear/brain interprets that pressure. When some frequency ranges in one headphone model are emphasized and some reduced in the other, as is always the case, all we can do is use our ears to match them as closely as possible.

The conundrum can easily be shown by this test: subjectively match the loudness of one headphone model to another using a single piece of music. Then change the music. You will discover that on the next selection, headphone A may be a bit louder than B, but on a third piece of music, B sounds a bit louder than A. This is simply because different frequencies are emphasized in different pieces of music, which may favor one headphone over another. So it is actually unscientific to match two headphone models by SPL and the only practical thing to do is try to get them audibly as close as possible to one another and do minor tweaks by ear during the testing. Then listen to a wide variety of musical selections and arrive at your conclusions over time.

In my next column I'll tell you about the objective (scientific) equipment testing I performed on the headphone amplifiers and then invite you to our listening party.

sfoclt's picture

Started reading just for a taste and read all the way through. Looking forward to more.

Seth195208's picture

This will be fun. Weeeeeeee!

veggieboy2001's picture

Thanks for the interesting article...I'm curious as to how it turns out!!

zobel's picture

I have been enjoying his website, and some you tube posts he has up. It is so great to get updated with the latest ideas and information on recording, mixing, and mastering! I am going to get his book "Mastering Audio: The Art and The Science". As a project studio hobbyist, and one old enough to have used wire recorders, I will really enjoy that read, and learning how the art and science has progressed.

Thanks,Tyll for bringing this uber-pro, three time Grammy winning engineer to my attention. Much fun!

tony's picture

Mr. Katz,

I just ordered one of your CDs .

I collect music for home playback, I rate each, I give the best 5 Stars, 4 Stars are good, 3 Stars are keepers for some important reason, 2 & 1 Star will get deleted at some point.

I'd like to ask why a great recording system can result in a 3 Star CD?, why bother doing middling work? (not that you do).

I don't think I'm alone in appraising my music, the Hifi stores will like to demonstrate on their "select" music.

Movie Industry soundtracks all seem to be the best of 5 Stars however the same song on the music CD is not quite as wonderful.

In the 1980s Sheffield Labs & Reference Recordings quality levels seemed up with the playback performance capabilities of a superb home systems, the majority of Vinyl seemed barely playable by comparison.

Today, headphones can resolve levels of 5 Star music ( I can go from 5- thru 5+ in easy to grasp steps).

Buying music was always risky business.

On the matter of these high level headphones:

I just had a good listen to a wide range of $1,000 and up headphones at the Ann Arbor headphone meet.

I couldn't cope with the big Audeze 3 Fazor , it sounded wonderful but was way to heavy, I was happy to get them off my head, phew, it was a bit of a let-down as I'd heard such wonderful things about their musicality. The Audeze "new" 8s in Open were wonderful and comfortable to wear for a while, I don't know how they'd be for extended periods.
I discovered that I can wear a 10oz. headphone all day and that's what I end up doing with my Sennheisers.

And I discovered that a 5 Star CD will be better on my inexpensive Sennheiser RS120 wireless FM headphones than a 4 Star CD on my super expensive wired Headphone system.

As a result: I'm chasing well recorded music much more than I'm interested in a $5,000 headphone system. I might be alone in this summary as I see headphone people's collection of gear is typically vast.

Can you comment on using a dbx 31 band eq to flatten out headphone response curves, please? This seems to be a Taboo among the headphone community but not at all in the Pro-Audio community.

Thank you,

Tony in Michigan

ps. I'll have that Paquito d'Rivera CD in a couple of days

Bob Katz's picture

Dear Tony and group: I'm very pleased to see all your comments and whet your interest. In the next few episodes you will eventually learn what headphones I picked and what I considered to be the winner. But there are ups and downs and it's going to be a roller coaster ride. Tony, you will learn what I will feel about using a DBX graphic equalizer to equalize headphones as the episodes move on. Somewhere around episode 5 of Katz's Corner! Sorry to make you wait, but I don't want to spoil the suspense. Thanks for getting Paquito D'Rivera's Havana Cafe album. The 1644 CD is a little bit bright, I prefer listening to the 2444 original which I have on my hard disk server. Unfortunately, HD Tracks has not made that available. Put in a request! I do think it sounds better on a set of speakers, which really can throw that soundstage. But you will not be disappointed in the realism that large diaphragm headphones can produce with this recording. But if you're unhappy with the sound of my recording, it's a matter of taste, and you're allowed to give it only 3 stars. :-)

tony's picture

It just arrived today, definitely an "A" set of music.

I was writing a letter to the Ayre guy, Mr.Hanson but I kept getting paused by the music pulling my attention away from my writings, this is quite a stunning performance.

I'm reading this to be a Chesky, 1996, 128x oversampling.

Well, if this is your calling card I'm calling!

I'll start a search for your work at Chesky, did you say you did 150 projects for them? , I'll start buying them up.

Funny thing is that I'm not a fan of this type of music but I can't stop listening to it, it's beautiful. Thank you!

Tony in Michigan

Bob Katz's picture

I meantioned Havana Cafe, which is only available in 1644. But Portraits of Cuba is avaiable at 2496 at HD tracks. I don't remember using a 96k converter to make this so I suspect the 2496 comes from the analog Dolby SR backup I also made at the session. But I didn't do the tape transfer on this like I did for Rebekka Pidgeon's 15th anniversary album, so I can't "certify" the quality of the 2496 of Portraits of Cuba at HD tracks until I hear it. Someone who has both the 1644 CD and the 2496 from HD tracks could comment. Remember, I was the recording engineer, so I know what it really sounds like :-) But I'd like to hear your thoughts!

RudeWolf's picture

Hey Katz,

Great article! A company I work for has developed a calibration software which, according to our users, works wonders for recording studios. We do multiple measurements around the listening spot and then correct any room induced deficiencies. Check it out, you might like it!


Also one more thing - it works on headphones as well! We do our own headphone measurements and then generate calibration curves which the user can use on our plug-in. The effect is really beneficial to studio guys who so far have been struggling to get good translation between studio headphones and speakers. Calibrate both and you'll get the same tonality!


Give that a whirl as well!


tony's picture

Thank you for writing back.

You're presenting a dimension of music we rarely get to learn about, of course we'll ( I ) be patient.

I hope this attention from out in the audience doesn't put you off.

I had a look-see at the images of you that Google put-up, hmm, are you actually using a leather sofa as a desk chair?

I've already learned a couple of things from you so this series will be helpful, thanks for sharing.

Tony in Michigan

eatapc's picture

Great article, Bob! As far as matching levels, isn't a LUFS or LKFS measurement supposed to be the standard for subjective loudness? When comparing audio files, I like to throw them in some software that'll give me a LUFS measurement so I can adjust the levels on the tracks to match each other. Is that applicable to headphones? Here's one:

Take care -- Mark B

Azteca X's picture

Bob, I'm thrilled to have you posting on IF. I have the latest version of your book in arms reach in my office. Thanks for being such a great resource and all of the great work you've put out over the years. I look forward to reading about your adventures in headphone-dom!

Three Toes of Fury's picture

You had me at "I built a custom audio mixer in a suitcase"!

Welcome Bob...thanks bunches for sharing your experiences, gear, and test rig...cant wait to read more of your adventures in audio!

Peace .n. Living in Stereo


Three Toes of Fury's picture

one question and/or request: will your articles discuss songs or music styles that you used when testing gear? Im always interested in learning more about the music that headphone/amp enthusiasts use when testing gear. I know this is heavily influenced by personal music tastes and preferences however its a great way to be exposed to new stuff.

Bob Katz's picture

I'm thrilled to be here, too. Thanks for the comments. I'm feeling giddy :-).

As for sonarworks, which I had not heard of, I was put off by the "results in 15 minutes" claim on their site, because I know what it takes to analyze a room and correct ONLY what is needed and useful and no more. I know how long it took me to get my room calibrated to my satisfaction, with a combination of several pieces of software including the wetware that's between my ears! It took months to make me happy, not 15 minutes!

In addition, the resolution of any correction method has to be very high in order to meet my standards of sonic transparency. I've auditioned and tested several of the best correction systems on the market, and I haven't found anything better than Acourate in terms of transparency. Acourate was the only system that matched the transparency of pure analog with no DSP in between. Part of the reason for Acourate's transparency is that it takes in signal at 64 bit floating point and puts out the outputs of its filters dithered accurately to 24 bits for the DAC. The rest of the transparency is due to Dr. Brueggemann's attention to detail, method of avoiding overcorrection and use of a psychoacoustic weighting based on Jim Johnston's published research, which I believe to be accurate. So perhaps Sonarworks is sonically transparent but given the statistical failure of all but one of the systems I've tried, I'll remain skeptical if you don't mind.

I'm also intrigued by Sonarworks' claim of headphone response correction. First of all, they would have to measure my pair because manufacturing variances exist and my ears are good enough to hear them. Next, as Tyll has pointed out at Innerfidelity, headphone response measurements are in their infancy. It seems to me that determining the amount of smoothing necessary to reject the radical extremes of peaks and dips that headphone measurements show is also in its infancy. We should take baby steps when performing headphone correction or it will result in overcorrection that literally sucks. With the correction introducing comb filters instead of correcting for them. So again I'm skeptical. I also feel uneasy about the company's big claims with no technical papers to support their claims. I would need to know details about the measurement window(s) they use, whether their correction is IIR, FIR or a combination of the two, the impulse length they use for measurement and filter correction, whether they incorporate any sample rate conversion, the number of FFT steps used, the psychoacoustic weighting method, and the precision and internal resolution of their processor. In other words, I feel that any company making claims about room correction should offer more than an artsy paragraph about their proprietary software. I need a white paper, more than an inkling of the nature of the methods that they employ. I am not impressed by technical mysteries or endorsements by people I do not know or trust. There are always compromises involved in room correction. Overcorrection is a common disease. What do they do to avoid overcorrection? Averaging multiple microphone positions has its proponents, but it does not result in an optimum response anywhere, only an "average" response.

Next, regarding LUFS or LKFS measurement.... LUFS is useful for matching program material within a few dB, but not for the degree of precision that would be necessary for matching up two distinct models of headphones or loudspeakers. There's nothing that can do that, I believe. The ear detects the frequency response differences and doesn't come up with a single "loudness number"... but rather an integration of many different frequency bands which have the partial loudness. It's more appropriate to talk about a given frequency band and its weighted level when talking about the perceived loudness, so if headphone 1 has band 1 strong and band 2 weak and headphone 2 is the opposite, it would be a matter of personal taste when the two match in subjective loudness. The ear perceives those differences, which are just weighted out and ignored with the LUFS measurement.

So LUFS, LKFS or even SPL, doesn't tell you enough about the subjective loudness. The R-128 standard is a weighting filter, a simplified approximation of the ear's performance. Yes, it's a BIG improvement over the wasteland that came before, but LUFS is not going to tell you how to match up two headphones to the 0.1 dB of perceived loudness that would be needed to do a machine-matched-to-perception loudness comparison. But LUFS is useful. I do use LUFS measurements in my mastering work, and they are a good guide as to how things will come out within the real world, within I would guess about + or - 2 dB of the perceived loudness, which is "good enough for professional application." But not good enough for precise scientific judgment of "which headphone is better", I'm afraid. I would love to hear about any experiment you perform to adjust headphone response with an LUFS meter and a measurement rig (which is very expensive and which I do not own) and let me know how closely you perceive the loudness of two different headphone models that you put through the adjustment. I think the frequency response differences would be so dominant and obvious to the ears that the meaning of "matched loudness" would just go out the door. For example, my Stax have a natural high frequency extension and a bit of an HF rise. My Audeze are more robust and strong in the bass and much warmer. I would struggle with attempting to match them every time I switch between. It's just not possible to satisfy some instant matched loudness with those kind of sonic differences. So that's why the method I suggest in Katz's Corner Episode 1.

RudeWolf's picture

I wouldn't have expected anything less from you, Bob - our claims should look silly to anyone who can grab a mic, run REW and go Rambo on a room. We are also not saying that everyone should forfeit physical room treatment - things like late reflections, resonances and others fall outside what we can achieve. However the problem is that not many people nowadays can afford or want to do complete room treating. With the lower cost entry ticket into music production we wanted to make good sound available to everyone. Here you can see what we can do in a typical bedroom setup -


The problem with whitepapers is that we can't tell much before our patents are approved. When that's done, expect to see a few publications in AES.

Despite this we're not two guys in a garage. It took about three years and about 1M$ in R&D to get where we are. I agree with both you and Tyll that headphone measurements direly lack a standard for FR measurements. After trying out the usual set of measurement head assemblies we set out to make our own - one that's specially suited for FR and THD measurements. Then there was the question developing a headphone target curve. Took us a few hundred evenings of relentless tweaking, measuring and blind testing, but we think it works. Our clients tend to agree and what's more important - their clients agree as well. Also it's not easy getting endorsed by Grammy winning engineers, these guys are very careful! You should know that!

Anyway - the proof of the pudding is in the tasting! I'd be delighted to hear what you think of how it sounds! If then you'd think we're a bunch of phonies, make an episode and tell people what mumbo jumbo to avoid!


Bob Katz's picture

Now you're cooking with gas. I think that before your patent is granted, you should write something that in essence reflects your above answer and approach but in more tantalizing detail. In other words, a bit more modesty will still attract the garage studios ;-).

In my copious free time I'll investigate your product!

RudeWolf's picture

Now, regarding the actual article - it will be interesting to hear your thoughts about the headphones mention earlier. We have measured a multitude of various cans to get an idea of their FR, THD as well as sample to sample spec spread.

Sometimes it's surprising how different are well regarded studio headphones. So far we have steered clear of the audiophile market due to its quirks, but it's quite evident that even the Pro market has its share of quirks as well.

Also do tell what you thought of the LCD-2 headphone. I use one at home for listening. The improvement after calibration wasn't subtle - mostly the higher end got much more pronounced. Here's how it measured to us -


Any deviation seen in this graphic should be viewed as a deviation from our target curve, not a perfectly flat FR measurement. Smart smoothing is applied with about 1/12 oct. across most of the range and less smoothing in bass for better correction accuracy.

MRC01's picture

That LCD-2 FR graph doesn't look like anything like the measurements that Tyll and others have done. The only thing that matches others is the 4.5k dip and the 9-10k peak. My ears subjectively agree those are real too. Nobody else is measuring the 400 Hz plateau or -7 dB @ 20 Hz.

Perhaps your target curve is different from a standard HRTF. For example some think the deep bass needs to be boosted more with headphones to give the same subjective impression as loudspeakers, and such a target might show the -7 dB @ 20 that you measured but nobody else has.

Bob Katz's picture

In future episodes, yes I do discuss various pieces of music that I used to evaluate the cans!

Bob Katz's picture

I was using a sofa as a listening seat for the mastering room for a period of time. The sofa had a low back to avoid comb filtering off the back, and it comfortably seated three. But the gotcha was that it turned out to be a Helmholz resonator and I should have known better. It has been replaced with those ergonomic chairs with the mesh which is very sonically transparent and therefore have a much smaller negative impact on the room acoustics. Pretty neutral impact I would say. Start with one human body in this room and add a second and we lose about 0.4 dB in the bottom end. It's audible and measurable.

tony's picture

I'm lookin at your room and loving that Sofa, I can take the loss of Bass rattling my wife's blissful world, besides I'll never be able to install a system to your accuracy standards, I think!

I love those equipment controls sitting just in front ( maybe instead of a Coffee table ) , in a nice rack-box mounting.

That Sofa would be a nice place to lay aside the various books and business literature I have to read, I can imagine a dozen reads sitting there waiting and a nice small dog, hmmmmm, I think I'm seeing "Quality of Life" here, thanks for the view.

And thanks for sacrificing your sofa for the product quality you create for "my" enjoyments. I think I like you!, keep up the hard work.

Tony in Michigan

ps. where's Tyll's pitching in here?, poor Tyll is still living with a Shelf System for his Music stuff ( very industrial, if you ask me, more in keeping for an old school technical serviceman ) .

Beagle's picture

..to have Bob Katz on board. Someone who's on our side, the side of good sound. Really enjoyed his articles on the loudness wars.

Looking forward to the follow-ups!

tony's picture


I just spent 2 or 3 hours watching the Videos you've done, SAE school, Mastering, loudness, your history, your workings and various music related tie-bits.

Wow, you explained many things I've wondered about.

I accept ( and like ) your A,B,C rating system.

I guess I accept your statement of CD being dead with 96/24 the standard. I've been a 44/16 fan all these years.

I noticed your GENELECs, can I have a pair ?

Apple setting the Standards cleared-up my future hopes.

I've been buying and loving the niche european artists, not knowing why they sound so wonderful, I think I now realize why: they sell to sensible european music lovers that appreciate subtleness of the musical presentation, they're old school analog lovers but today digital but definitely not Loudness people.

You are bringing into focus a good bit of long out of focus concepts. We left Vinyl years ago and got lost in digital. However the Converters are quite good today and we're getting a good listen using our headphone systems. I feel like my Sennheiser Systems are a 10X magnifier compared to the Loudspeakers & Electronics I've owned. ( are those genelecs as good as my Sennhiesers ) ?

Thank you for being helpful, you haven't inspired me to record but you gave me hope that A & B music will be available from an ever widening group of sources.

Tony in Michigan

John Grandberg's picture

Bob, I can't tell you how excited I am to have you contributing here with us. I too have enjoyed your output immensely over the years - both the music and the writing. I've learned a ton that I otherwise would never have discovered, as there just aren't many folks with your knowledge who are so willing to share it. So thanks!

Wait a minute.... *I* write for InnerFidelity. *You* write for InnerFidelity. This gig is technically considered "work".

Yep, from now on I'm telling people "I work with Bob Katz".

Tyll Hertsens's picture
He he he he, I'm giddy too.
Bob Katz's picture

My head is now officially in the clouds....


Bob Katz's picture

-6 below 50 Hz? Not according to my ears.

RudeWolf's picture

The graph is logarithmic, so 50Hz runs about -1dB and -6dB is around 20Hz. Have seen better, but orthodynamics can take the extra push needed to straighten them out. Also sample to sample inconsistencies for Audeze headphones are well documented. We've seen worse with Hifiman - I guess it's still some time until they get as good as Sennheiser in this regard.

As for what to correct and what to leave be - there's an engineer in charge of the process and he makes the decisions about more troublesome spots. Sometimes trying to overcorrect excites very nasty resonant effects - it's up to him to do fine tuning. Our reference is a corrected HD650 headphone which has been fine tuned to precisely mimic the tonality of flat sounding speakers in an excellent [treated] studio. Software does the majority of the work, but as you noted - fine tuning usually is crucial. There have been some headphones which sound fine by just inverting the measurement curve, but there are others which have proven to be very unruly. Especially some closed back on-ear headphones which change tonality drastically on even slight placement changes.

As for everyone else's (p)reference - our plug-in allows adjusting the target curve and the magnitude of correction, so everyone can see what's good for them. With that said - 'flat' (no adjustment) is default and usually our clients find it agreeable. One of the guys over at SoundOnSound forum said that he didn't care that much about whether our sound is 'correct' as long as he is able to hear more. Take away peaks to prevent masking and wide dips that result in lost information and you can achieve that.

I very much agree that headphone measurements are still in their diapers, however being in diapers still beats crawling around naked.

Bob Katz's picture

Anyway, this is the dark ages of headphone response correction. I doubt that I would like the sound of an LCD-2 corrected to the degree you show in your above graph. In general I would describe the LCD-2 with the Fazor as "very nice, but closed in on top". Notice I did not say "dark" or "rolled off" because am not certain that frequency response is the full answer. There's so much comb filtering in the measurements that the ear rejects, but there's other anomalies that the ear detects and notices... I think the science is not strong enough to know what to count and what not to count. But I admire your efforts! I base my judgments of what's right on acoustic recordings that I know intimately because I was the recording engineer and I was there. Recordings that I have reproduced on a wide variety of loudspeakers in many locations over years and years and lately my Acourate-corrected pair here at Studio A at Digital Domain. Recordings which I have made which I attempt to make sound like the live performance, as much as humanly possible. But I won't spoil the suspense. You'll hear about my conclusions in episodes 2- and up. I already heard privately from one fellow whose reactions to the Stax Lambdas versus the Omegas are the very opposite of mine. He thinks the Omegas are very dark and he likes the "openness" of the Lambdas. While I've always found the Lambdas to be colored. So, bottom line: different strokes for different folks. I know what I like, I know what I think sounds right and natural ----- BUT---- I don't pretend to know all the answers. My opinion is that I would be very cautious about correcting the LCD-2's with the degree of fine response changes you are proposing in the above comment. Skeptically, I'd like to say, since the measurements are so raggged, how do you know what part of the measurements to believe? How do you know what parts to reject? What is your reference for accuracy? How do you know that you are actually correcting for errors in the headphones (versus your idea of what they should sound like, using recordings that you may have no idea what they should sound like). So many of us adjust equalizers for our taste, but how many of us know what a recording actually sounds like and was intended to sound like? As I said, we are in the infancy period when it comes to headphone response analysis and correction. This is a fascinating subject. It would be nice to get rid of the chicken and the egg question entirely... I try to be as objective in my subjective judgments as possible.

Bob Katz's picture

Diapers it is. You will hear about my very empirical headphone EQ in a near future episode.

spritzer's picture

Great article Bob!! I'd like to put on my Stax Mafia hat though as the phones were called SR-3 and SR-5, SRD-3 and SRD-5 were the transformer based energizers Stax sold with the phones.

These sound truly great on modern amps. The earpads compromise the headstage a bit but there are plenty of modern earpads which can fix that.

PDC3's picture

Mr. Katz, seems as though you'll be sharing the InnerFidelity aesthetic of "some science, some soul" and I've been pleased with what I've learned on this site in the past six months. Looking forward to more of your thoughts and experiments.

Bob Katz's picture

Thanks, Stax mafia! I did make a mistake in nomenclature. Sorry about that. Tell us about the modern earpads that increase the soundstage. But today I find the SR-5s to sound thin and bright with the KGSS amp. They definitely need some EQ. The old transformer energizers were a nice complement to the too-bright Stax. But the best combination is a solid state amp with EQ or perhaps a tube amp with some EQ.

tony's picture

How can we EQ headphones? I'm ready to get into this.

Tony in Michigan