Thirty-Five Miles for a dB: Choosing Headphones for Mastering
Which headphones can one trust with mastering decisions in the bottom octaves? I visited Dan Clark Audio and Audeze to find out.

At a Glance
I have included a brief summary ahead of what is surely the most comprehensive resource I have ever attempted to compile.
Headphones can handle most of the work
I primarily use headphones for revisions and quality control, but headphones can serve as a primary monitor for mastering if one learns how to interpret bass from pressure alone. Still, a well-tuned room offers bodily cues that speed up decisions about the size of the low end as a whole and any overlapping bass elements, like a kick and bass.
A revealing low-end test
Listen to a kick drum over a sustained bass line. Does the kick stand apart from the bass as much as it does in a well-treated room and then clear out? Low-frequency extension from measurements will not tell one whether a headphone passes that test. What people call “slow” bass is generally the balance between neighboring octaves, which EQ can typically address.
Choose the headphones based on things EQ can’t supply
EQ drastically improved the headphones that I used uncorrected for years. Select a pair of headphones based on things that EQ can’t correct. Specifically, look for a seal that stands up to reseating and personal features (glasses and hair including beards & sideburns) as well as consistent behavior as level rises. Ideally, select a headphone that does not require narrow EQ adjustments to suit taste or anatomy.
Correct, then keep level consistent
A published EQ preset targeting the Harman curve is a reasonable starting point for correction. Personal adjustments in the higher octaves can bring it closer to a reference point one can trust. Always listen at one recallable level and keep the correction in the monitoring path only. Keep crossfeed off for critical listening: because headphones exaggerate low-frequency material in the sides, it’s easy to identify oddities down there. Crossfeed surrenders that advantage.
Why I kept the LCD-X
Even uncorrected, my LCD-X and DCA’s ÆON CORE came closest to my room in the lowest octaves. With EQ, most of the others would likely get close as well. After all, I only listened to high-end headphones. But EQ can’t stabilize a seal that moves. For a portable rig I’m keeping my Audeze LCD-X as its seal barely reacts to my glasses and I already own a pair. For those who don’t wear glasses, the ÆON CORE ($899) deserves a direct comparison.
- THIRTY-FIVE MILES FOR A DB
- ON WHAT HEADPHONES CAN AND CANNOT REPLACE
- ON LISTENING TO THE BOTTOM OCTAVES
- ON HEADPHONE BASS
- TWO FACTORIES: A CASE STUDY
- ON SEAL, GLASSES AND PADS
- ON CORRECTION & LEVEL
- ON CROSSFEED
- AN EXAMPLE RIG
- COMING HOME
- WHAT I LISTENED TO
- REFERENCES & ACKNOWLEDGEMENTS
Thirty-Five Miles for a dB
The studio is just over seventeen miles from my front door. Most days, it’s a pleasant enough drive. On some days, it’s a thirty-five-mile round trip to make a 1 dB bass move on a revision that takes all of five minutes!
The room is worth it, of course. It’s acoustically treated, digitally corrected and designed around monitors that present the lowest octaves of music without compromise. Unfortunately, the room can’t follow me home to my shoebox apartment in Leucadia.
So this quest began with one question: how do I take a mastering studio on the road?
The question is not mine alone. Mix engineers and producers often ask me how to get more out of a compromised listening environment — typically a spare room in an apartment or house. The conversation invariably arrives at the same constraint: budget. Square footage is expensive. A room with enough space for meaningful low-frequency treatment costs more still. Throw in digital correction and suitable monitors and the undertaking lands firmly in five figures. Once serious room treatment is off the table, the next question arrives right on cue:
Which headphones?
This piece is my attempt to answer this question for critical listening.1
Along the way I visited Dan Clark Audio here in San Diego and Audeze in Santa Ana, expecting to come home with a new piece of kit.
I ended up keeping the headphones I already owned, the Audeze LCD-X.
On What Headphones Can (and Cannot) Replace

Let’s be precise about what a good room provides. Some of its gifts are spatial: a phantom center suspended between two speakers and an image that sits in front of the listener rather than inside the listener’s cranium. At a healthy level (~80 to 85 dBA), the bottom octaves are felt in the body as well as heard in the ears. A good room’s supreme offering is bass that behaves. In a room of ordinary size, low-frequency waves form modes that inflate some notes, minimize others or ring on after a note ends. And these qualities often shift whenever the listener moves their head. Controlling these standing waves with acoustic treatment takes physical depth, floor space and a real budget.
Headphones sidestep room modes. At low frequencies there is simply no room inside a headphone for modes to form. A good pair sounds essentially the same in a guest bedroom as it does in a world-class reference room. They do trade one problem for another deceptively costly problem, though. More on that subject in On Seal. Headphones are also microscopes of near-infinite utility for edits, fades, noise floors, reverb tails and the crimes of lossy encoding.
Stereo imaging is another strength, with one caveat. On loudspeakers each ear hears both speakers, so what one dials in on headphones is not exactly what a listener hears in a room. One has to learn the relationship between one’s headphones and speakers. That said, an image balanced entirely on headphones without a room to reference against is not necessarily a bad one. A massive number of listeners now hear music primarily on headphones and many others are sadly subjected to systems that barely preserve a stereo image at all. Many single-box Bluetooth speakers sum to near-mono and a car seats a driver far closer to one speaker than the other. For those systems the most informative test of translation is a mono fold-down.
The upper mids and treble can also be dialed in on headphones with confidence if three conditions are met:
- the listening level is at a fixed reference
- the headphones are corrected for one’s own taste and anatomy
- the result has been checked against speakers (or masters) one trusts.
All three conditions are covered below.
In my experience there are two decisions that typically require loudspeakers in a treated room:
- the size of the low end against everything else: how big should the bottom be?
- the low-frequency microdynamic profile: how high should short bass impulses rise and how quickly should they decay?
Together these decisions determine much of whether a master translates as intended across listening systems. Make an ill-informed decision and the master sounds muddy or thin. More often, poorly defined bass dynamics lead to a song that doesn’t sound as loud or strong as intended. Peaks set the ceiling, but the ear judges a sound’s loudness largely by how much of the signal sits near it and for how long. A low end with insufficient energy near its peaks and too much in the middle of its dynamic range reaches the ceiling without actually sounding loud.
That’s what this quest is about: finding a headphone that can be trusted with those bass decisions. For me, the LCD-X comes closest. Can I now work on headphones exclusively? Not quite yet. Headphones deliver pressure to the ear canal and nothing else while a room delivers bass to the body as well. Without the cues the body supplies in a loud room, learning to judge bass weight and impact from pressure alone takes time.
At this stage I primarily use headphones for revisions and quality control. I’m confident in making broad, multi-octave adjustments below ~200 Hz on headphones. But I still prefer to make the drive for single-octave adjustments down there as I find the additional bodily cues from the room to be of great utility when dialing in one octave relative to its neighbors.
On Listening to the Bottom Octaves
Before this search began, I spent time with two well-regarded (and rather expensive) dynamic-driver headphones. Both were otherwise pleasant to listen to but shared two unpleasant habits: limited perceptible extension below 40 Hz and a tendency in dense material for resonant or legato bass to swallow a kick drum whole.
Picture a kick over a sustained bass line. Both live mainly in octaves 1 and 2.2 In most mixes the kick is meant to be perceptibly louder than the bass for the kick’s duration, say 100 milliseconds. After that, the kick shouldn’t loiter. It should clear out and hand the region back to the legato bass.
There are two ways playback can fail. First, the joint kick-bass envelope can be overstated when the kick lands: the resonant bass blooms under the kick and the bloom outlasts the kick itself. Second, the kick envelope can be understated when the kick lands: the kick just never rises above the bass line. In either kind of failure, the contrast between kick and bass is no longer what the production team intended. If that contrast is not presented faithfully, how can one determine what kind of low-frequency processing (if any) belongs on a master?

The human auditory system can’t distinguish low-frequency sounds by pitch with as much resolution as it does higher-frequency sounds. This is primarily because down low, auditory filters can span multiple notes (near 50 Hz a single auditory filter is around 20 to 30 Hz wide, more than half an octave). In other words, down low, kick and bass are perceived as different instruments largely because their envelopes differ.
Dynamics, then, are not a secondary quality of bass. The envelope is what allows a kick to be perceived as a kick in the first place!
Further, the auditory system adds up sound over roughly the first 100 to 200 milliseconds, about the duration of a kick. For the auditory system to report a short sound as loud as a longer, sustained sound, the short event needs more level. So part of the kick’s level advantage in the mix is spent on simply keeping up with the sustained bass. The margin one actually hears is small, carried by just a few cycles (about five at 50 Hz).
What would seem to be small errors in monitoring can eat into that margin, though not every error does. A monitoring system that presents the bottom two or three octaves evenly, but 2 dB too loud, does not alter the kick-bass margin. The error still affects one’s judgment of the size of the low end, but it leaves the kick and bass contrast intact. Now suppose the error is uneven: a bass line’s fundamentals sit around 50 Hz and the kick’s punch comes from around 90 Hz. A system that plays 50 Hz 2 dB louder than 90 Hz eats roughly 2 dB of the kick’s margin over the bass. The kick is swallowed whole, or the bass simply sounds “slower.” A system with the opposite error presents the kick (and the song) punchier than the mix intended.
Every monitoring system is uneven by a few dB in the bass, very good rooms included. What makes a monitoring system trustworthy is that its unevenness doesn’t change from one session to the next; an engineer who works in the same room long enough can learn it and their judgments account for its bumps and dips, provided a listening position is maintained.
Whether a pair of headphones can stay consistent from session to session is the subject of On Seal.
The test is therefore simple to state:
Does a kick stand apart from the bass and clear out the same way it does in a well-treated room?
On Headphone Bass
Why would one headphone pass that test and another fail it? My first assumption was driver type. The dynamic headphones had failed, so a planar-magnetic design with its large radiating area would surely pass! I knew better than to believe driver type alone determines bass behavior, but I still would have bet on the planars. A small bet, of course. As the factory visits demonstrated, I would have lost it.
A headphone divides the air around its diaphragm into two volumes: the front volume, between the diaphragm and the eardrum, and the back volume behind it. “Open-back” and “closed-back” describe the back. At low frequencies the front volume is very small relative to a bass wavelength. The diaphragm isn’t launching a room-filling wave as much as it’s compressing and releasing a small pocket of air. The front volume is a pressure chamber rather than a miniature room.
To play one octave lower at the same level, a speaker radiating into a room needs ~four times the excursion. In the sealed front volume that excursion requirement doesn’t exist, so the same excursion (less than a micron for a large planar at normal listening levels) grants the same pressure at 200 Hz as at 20 Hz, as long as the seal holds.

Deep bass is therefore well within reach of a sealed headphone, but not every headphone will deliver it. The pressure that actually reaches the eardrum depends on a proper seal (no leaks) and tuning. The dynamic headphones I had auditioned stopped short below 40 Hz, but extension alone doesn’t tell me anything about the kick-bass relationship. Two headphones can both reach 20 Hz and present the bass dynamics very differently.
My next explanation for that difference was timing. Every mechanical and acoustic system has resonances: frequencies it favors much as a plucked string favors its fundamental. A resonance builds up more slowly than the event that excites it. Moreover, a resonance keeps ringing after the event itself ends. A sustained bass note lasts long enough to collect the full gain of a resonance while a kick has begun and ended before it fully develops. The bass bed swells, the kick blurs and a tail persists after the kick should have exited the soundstage. On the surface level, it’s a convincing story that matches what I heard.
But upon investigation, it’s probably not the true explanation. How long a resonance rings depends on how sharp it is. The modes of an untreated room are sharp: at 50 Hz they can ring for several hundred milliseconds, longer than the kick itself. The majority of headphone bass resonances are broad and damped (shortened). A well-damped resonance at 50 Hz, say with a Q around 1, falls 20 dB in about 15 milliseconds. That’s a small fraction of a kick. For a headphone resonance to smear a 100-millisecond kick it would have to be sharp enough to show up as a narrow peak in measurements (and very likely as one-note bass upon audition).
There was a second reason I doubted the resonance explanation. When I pitched it to a mentor, he informed me that with a good seal, a headphone’s bass behaves close to what we engineers call minimum-phase: its timing is tethered to its frequency response. Change one and the other changes with it. Correct the response with EQ and one largely corrects the timing along with it. Minimum-phase EQ filters operate the same way resonances do: narrow boosts or cuts ring longer than broad ones. A narrow cut employed to correct a ringy peak cancels the peak’s ringing, but only if the peak stays put and is actually there. If the peak is only reported by a fixture-derived measurement and is not present on one’s head, or if it moves with reseating, a narrow cut will introduce its own ringing without removing any.

So what does “slow” mean when I describe a headphone? I believe it is the inter-octave bass balance. Headphones voiced with excess energy in the octave where a bass line’s fundamentals reside will lift the resonant bass without lifting kick punch, which may live in another octave. “Fast” would sound like the opposite: headphones voiced with more energy where kick punch and note articulation live and less energy where resonant bass is.
It’s tempting to look at headphone measurements to find a bass response that is even, but the balance that reaches one’s eardrum isn’t necessarily the balance a measurement fixture reports. Below 200 Hz, pads and seal determine much of bass behavior. A fixture has a head and ears, just not one’s own, and it almost never wears glasses.
Some qualities of a headphone are not fixed. For example, a driver can compress or distort as level rises. That kind of quality is not reported on a frequency response chart. Driver type didn’t determine the kick test, but it matters here: a big, light planar diaphragm barely has to move to pressurize the front volume, so it can stay clean and, most importantly, consistent across a huge range of playback levels. EQ is not a solution for a headphone whose behavior drifts considerably based on dynamic qualities like level, seal or program material. The only solution there is to select a different pair of headphones.
As for open versus closed, I generally prefer open-back headphones for critical listening if I’m working in a quiet room (below ~40 dBA ambient). That is just my preference. Well-executed closed-back designs are quite capable.
Two Factories: A Case Study
I visited both factories in September 2026. Every audition and comparison was made with the DAC/amplifier set flat (no EQ engaged). I matched levels to roughly 85 dBA using an SPL app on my phone, so very close calls could be due to small level differences. Each factory had a different DAC/amp, so I’m more confident in comparing headphones within each visit than across factories.
Dan Clark Audio. Through a Luxsin amplifier I auditioned DCA’s flagship Stealth and Expanse along with the E3 and the ÆON CORE. All four were lightweight, compact and comfortable. In the midrange and treble the flagships were excellent. Detail and depth were all there and complex material was easy to inspect without fatigue. From the upper bass up they also sounded surprisingly close to my corrected reference, even in stock form.
The low end, however, was another matter. At about 0:55 in Apparat’s “Hailin From the Edge,” two kick drums of distinct weight enter. Listening on the closed-back and open-back flagship headphones (Stealth and Expanse, respectively), neither kick rose high enough above the tonal bass occupying the same octaves. I took my glasses off and listened again. The tonal balance changed, but the kick still didn’t rise as high as I felt it ought to. That made it difficult to contrast the two distinct kick envelopes or to judge how quickly either kick decayed. The ÆON CORE was the exception. I listened to the same passage and the kick-bass relationship was much closer to what I expect in a good room, albeit without the excellent frequency response the flagships had above the bass.
I initially heard the flagships’ bass limitation as a timing or resonance problem, but now I suspect it’s just their bass tuning and the way it lands on my head. I didn’t audition them corrected, so I can’t confidently claim that EQ would bring them closer to what I was listening for. Mostly, I was surprised that some of the finest planars available had a problem similar to the dynamic headphones’.
Audeze. For the visit to Audeze I brought my own LCD-X. (Yes, I brought headphones to a headphone factory.) Through an RME ADI-2 DAC FS, I compared it to the factory’s brand-new LCD-X and two of Audeze’s newest models: the flagship LCD-5s and Manny Marroquin’s MM-520 collaboration. I also heard the LCD-5s and factory LCD-X through a dCS Bartók. The electrostatic CRBN2 played through the Bartók and a Mjölnir energizer.
The MM-520 is quite a bit lighter and easier to wear than the LCD-X. Its bass resembled the LCD-X’s but sounded lighter in the lowest octaves and a little slower with less punch in octaves 1 and 2. Its top end sits further back than I need for judging transients at higher SPL. Broad EQ would likely address both. Besides, a signature collaboration should be allowed to remain specific to the engineer who informed the design. The LCD-5s was easily the most enjoyable listening of the day. It sounded open and refined, if a bit midrange-forward. That said, its bass still sounded a little slow in the lowest octaves for critical work.
The CRBN2’s low end was another story. It sounded like a pair of loudspeakers strapped to my head and I imagine it could serve as a primary monitor. The bass was lifelike with none of the limitations above, though one still doesn’t feel sub bass in the body. At $5,995 before a sizable dedicated energizer and a converter worthy of both, the system that came closest to replacing loudspeakers in these auditions costs about what the loudspeakers do. And it is certainly not something one travels with.

Of the headphones that fit in a bag, DCA’s ÆON CORE, my own LCD-X and the factory’s LCD-X were the only uncorrected pairs doing what I was listening for from octaves 0 to 3. The clearest moment came while I switched rapidly between the MM-520, LCD-X and LCD-5s on Strong Maurice’s “Rice Bowl,” which I mastered at RQ in 2021 before I owned planar headphones. On the LCD-X the bottom octaves came closest to what I hear in a well-treated room. The best description I have is that it sounds like speaker excursion: I hear and nearly feel the low end move forward and back. Of course it is not literal excursion given how little a headphone diaphragm travels. I suspect it’s what deep extension sounds like when each of the lowest octaves sits in proportion to its neighbors. The other Audeze headphones invited me to think, “What gorgeous music.” The LCD-X had me inspecting the music instead.
On Seal, Glasses and Pads
Everything discussed so far depends on the seal. A leak in the seal compromises the front volume of the pressure chamber, allowing air displaced by the diaphragm to escape instead of compressing in front of the eardrum. Worse, a leak generally interacts with the cavity in ways that vary with jaw movement, head and pad position, creating a position-dependent response reminiscent of an improperly treated room.
A leak is expensive down low because it doesn’t turn down the bass evenly: it turns down one octave or part of an octave relative to its neighbors. Worse still, a leak isn’t typically static because removing and reseating the headphones changes the amount and frequency of bass attenuation. Sean Olive and Dan Clark, whose factory I visited, presented a study in 2025 measuring the same headphones on groups of real listeners. Some models varied from listener to listener by a few dB as one would expect, but others changed by more than 10 dB in the bass.
A good room can be a few dB off at the listening position, too, but that error is consistent from day to day, session to session. A headphone’s error can change every time it’s used if the seal isn’t strong. That’s the trade in question: a room’s bass response depends on where one sits; a headphone’s depends on how it sits. Since the kick’s margin over a sustained bass is only a few dB, an error that moves between removal and reseating headphones does more damage than a larger static error that stays put.
I wear glasses — it’s part of my sophisticated Bohemian look. At both Dan Clark and Audeze I took them off and put them back on repeatedly while listening to every pair. I found that the DCA headphones lost some deep bass extension when I had my glasses on. Moreover, each time I put my glasses back on, the bass response sounded slightly different. Audeze’s MM-520 and LCD-5s were even more sensitive while the LCD-X was rather unaffected. For those who wear glasses, leather pads, thin temple arms or a dedicated pair of frames for headphone work are worth trying. Some pads are even cut with channels for temple arms, which Rtings notes can make a headphone more consistent for bespectacled folks like me.
A headphone sporting all the right qualities can be rendered unusable for low-end decisions due to an unstable seal. Because the bass response was slightly different with some headphones each time I put my glasses back on, those headphones could not be corrected with static EQ — the bass response would “wander” underneath any correction curve.
Earpads wear out. I learned what that costs at the factory when I compared my LCD-X to the factory’s demo pair of the same headphones. Compared to the demo pair, mine gave up a touch of punch in octave 2. The drivers had not changed; the factory confirmed so when I noted the difference. A static change in the cushions registered to me as lost punch, once again the frequency response passing for a speed difference. My physical reference had drifted (Le Grand K, anyone?), its betrayal revealed only by comparison. The kilogram’s drift was identified by weighing it against its official copies. If one is serious about maintaining a reference, pads can be policed the same way: keep a fresh spare set of pads and compare every so often, replacing the worn set when the difference becomes consequential.

On Correction and Level
I performed every audition flat, but for critical listening I employ correction. In my view, correction is not optional for critical work as both the headphones and one’s ears deviate from a neutral reference point. Everyone’s ears and heads are unique.
Correcting headphones is rather easy these days as one can employ fixture-based corrections to hit the Harman target response. How are these corrections derived? Basically, superheroes like oratory1990 place the headphones on an artificial head and ear (with a facsimile ear canal and everything) and measure the response. The response is then compared to a target curve, typically the Harman Over-Ear, and a difference filter is derived.
The catch, of course, is that the fixture from which the corrections are derived is an average ear, which is to say, it is nobody’s. A fixture-perfect correction curve isn’t going to be perfect on a real listener’s head.
Taste is also a real variable here. Harman’s own research found listeners don’t agree on how much bass and top end they prefer relative to the target. Around 3 kHz, preference and anatomy are difficult to differentiate. The ear canal and outer ear lift that region by over 10 dB, but everyone’s ear is different: the exact amount of boost and the exact frequency at which the boost peaks vary. A bit of what we call taste around 3 kHz is likely the result of one’s anatomy.
There are utilitarian preferences as well: one engineer asks that a monitoring system overstate 2 to 5 kHz so harshness is made obvious while another asks for the same frequency band to be recessed for less ear fatigue over a long session.
I used the LCD-X uncorrected for years. During this quest, I settled on an EQ that has turned the headphones into a reliable reference.
My correction comprises two layers:
- oratory1990’s Harman target ADI-2 preset for the 2021 revision of the LCD-X
- a personal (or anatomical) adjustment: the preset’s 4 kHz filter set 1 dB lower than oratory1990 prescribes, with the same frequency and Q

In oratory1990’s preset for the LCD-X, the bass correction is just one broad shelf. The inter-octave relationship stays intact. That’s why I auditioned all the headphones uncorrected. If a headphone gets close without EQ, it is unlikely that narrow boosts or cuts down low are required to bring the bass closer to the target.
I found my personal layer of correction by comparing the preset-corrected LCD-X to my studio in an A/B test. To my ears the preset was too hot in the high midrange, so I pulled back the 4 kHz boost by 1 dB. With the personal layer of adjustment, the corrected LCD-X sounds as I “expect” music to sound for critical listening.
If narrow boosts or cuts are required to suit one’s taste or anatomy, I would advise that one audition different headphones instead — remember, narrow EQ affects timing more than broad EQ.
On level: listen at one recallable level. 80 to 85 dBA on a handheld meter when placed at the driver is a good place to start. Of course, the level at the driver or cup is not the level delivered to the eardrum. When level-matching headphones, measure with the meter in the same position on each driver. Consistency is more important than the number — pick a level that is both comfortable and revealing.
Wherever headphone correction occurs, ensure it is only in the monitoring path and never in the mastering chain.
The proper order of operations follows:
- Before buying, check the published correction for that model, if it exists. If narrow filters (especially below 200 Hz) are required to meet the reference target, choose a different pair of headphones. Consider Rtings’ Frequency Response Consistency for an idea of how much a model’s bass moves with fit.
- Choose the headphones for what EQ can’t fix. Audition them three ways:
- Play a 15-second passage with sustained bass with a kick drum (perhaps “Afghamistam” by Botch) at a ladder of playback levels: quiet, a casual level, an attentive level, critical (80 to 85 dBA) and too loud (around 95 dBA, briefly). Listen for the kick envelope, particularly at the casual, critical and too loud levels. Compare against the speakers or a second known pair at the same levels: does this headphone change more than they do? Does the kick stop gaining impact as the level rises? Are there harmonics imparted at higher levels?
- Play a song known intimately and repeat the test above, listening for midrange and high-frequency dynamics and harmonics.
- Remove the headphones and reseat them repeatedly, listening to the same 15-second passage of a song each time at the same level. Remove and replace eyeglasses as well, if applicable. Does it sound different each time?
- EQ the headphones to a reference target (ideally the published preset)
- Optionally, add a personal layer by comparison, keeping EQ moves broad
- Set one listening level and store it along with the correction
On Crossfeed
Crossfeed is an attempt to replicate the in-room imaging presented by stereo loudspeakers.
When one listens in a room with loudspeakers, each ear hears both speakers, left and right channels simultaneously. If one were to mute the left speaker, both ears would still register input from the right speaker. When the same listener uses headphones, each ear can only hear its own channel. Mute the left driver and the left ear hears silence.
Generally crossfeed works by feeding a low-passed and slightly delayed portion of each channel into the opposite ear.
It’s a fun feature offered by DACs and headphone amps. In my experience it often sounds better for casual listening.
I keep it off for critical listening because headphones exaggerate low-frequency side content, making excessively wide or out-of-phase bass easy to identify quickly. The same stereo bass imbalances are difficult to perceive if listening in a mediocre room. Crossfeed surrenders this advantage, as the filter involved blends low frequencies more than higher frequencies. Because I can hear exactly what sits in the sides, I only narrow the bass where it needs it rather than “mono-ing” the low end by default, a process that I advise against as it can cause unintended changes in the sides.
If I were still mixing, I might toggle a simple crossfeed on my DAC every now and then. Newer approaches to crossfeed are particularly interesting. Topping offers a convolution-based crossfeed that models head-related transfer functions and room acoustics, but I haven’t auditioned it yet.

An Example Rig
If budget is the constraint, the best place to start is with the headphones one already owns. One can use the method in this piece with almost any decent pair. The following example is just my own configuration and is by no means a minimum.
Cans, conversion, current, cables and correction: here is my rig at published prices as of November 2026.
- Cans: Audeze LCD-X, $1,199 new plus $90 for a spare set of pads. Used pairs of the current revision pop up online for $700–$900.
- Conversion and current: RME ADI-2 DAC FS, $1,299. Loads of power for the LCD-X and many other demanding headphones. Bear in mind that headroom requirements depend on a pair’s sensitivity, impedance, EQ boost and intended listening level. Its recallable states also hold the reference level and the correction curve. The alternative I’d happily recommend is Topping’s DX9 Discrete.
- Cables: The cable in the box. A planar’s comparatively flat impedance means a few tenths of an ohm of cable resistance creates a tiny level change that’s essentially uniform across the audible band, roughly a tenth of a decibel on the LCD-X. There may be esoteric reasons that different cables may sound better (or just perceptibly different). Because I answered my question before getting to cables, that debate will have to wait.
- Correction: Free with any digital parametric EQ or built into the RME.3
That totals $2,588 new before tax, which is comfortably four figures rather than the five (or six) figures of the room or the CRBN2 system I heard.
Coming Home
For casual and attentive listening, the LCD-5s through the Bartók was easily the best system I heard at either factory. Of the other Audeze headphones I could carry, the MM-520 came closest to the LCD-X in the bass and is far easier to wear. Both may well be more rewarding to live with than the LCD-X, and I could see either being used for serious work by someone who doesn’t wear glasses.

For critical listening in the bottom octaves, the LCD-X remains my selection. To my ears the uncorrected LCD-X and ÆON CORE both presented a kick-bass relationship that most closely resembled a well-treated room. The LCD-X wins for two reasons: it isn’t sensitive to my glasses and I already own a pair.
Admittedly, EQ would likely bring the other headphones I heard much closer to the room-like bass target. But for many of the headphones I auditioned, their seal was inconsistent. The LCD-X seals wonderfully on my head, unbothered by glasses. The downside is that the LCD-X is, by far, the heaviest pair of headphones in this entire discussion (612 g, against 480 g for the CRBN2 and 420 g for the LCD-5s). I’d tell a friend who doesn’t wear glasses to listen to DCA’s ÆON CORE against the LCD-X.
Total transparency: I’ve used the LCD-X for years now. Part of why it sounds right to my ears is simply familiarity. But familiarity doesn’t explain everything here. I mastered “Rice Bowl” in the previous RQ studio before I owned planar headphones. So my reference for that record’s bass response doesn’t come from the LCD-X. But when I heard the song on the LCD-X, it sounded close to the way that song sounds in my current room. Still, I can’t rule out the bias, so I’m stating it here.
My advice to engineers: method over model. Listen for the kick and bass relationship, change one thing at a time (glasses, pads, correction) and most importantly reference against speakers in a room one knows well. If a known room is unavailable, use records one knows intimately.
These days, I do most in-the-box revisions from my apartment with the portable rig. Rarely, I’ll master a song entirely on headphones. But I still opt to start every bass-focused master in the studio. The bodily cues one receives from a good room make for quick judgments about the size of the lowest octave of music.
And the truth is that listening to records on big ol’ speakers is just too fun.
What I Listened To
All auditions were carried out via Tidal and Qobuz at the highest available resolution.
Chains
- Dan Clark Audio: Luxsin headphone amplifier, set flat
- Audeze, primary: RME ADI-2 DAC FS with DC protection engaged (ON, not FILTER) for the LCD-5s, MM-520, factory LCD-X and my LCD-X
- Audeze, secondary: dCS Bartók for the LCD-5s and factory LCD-X; dCS Bartók with a Mjölnir electrostatic energizer for the CRBN2
Tracks
- Apparat — “Arcadia (Album Version)”
- Apparat — “Hailin From the Edge”
- Shudder to Think — “X-French Tee Shirt”
- Radiohead — “Everything in Its Right Place”
- Botch — “Afghamistam”
- Murkury — “Zenith”
- Polyenso — “Falling in Rain”
- Maps & Atlases — “4/25”
- Radiohead — “Weird Fishes / Arpeggi”
- Carla Bley — “Lawns”
- Tinariwen — “Amoss Idjraw”
- Toy Matinee — “Queen of Misery”
- The The — “August & September”
- The The — “August & September” (Remastered)
- Nils Frahm — “Forever Changeless”
- TTNG — “Goose Steppin’”
- Strong Maurice — “Rice Bowl”
References & Acknowledgements
- Equivalent rectangular bandwidth
- Behavioral estimates of human frequency selectivity at low frequencies (Jurado, PhD thesis)
- Equal-loudness contour
- What’s the definition of an “open” or “closed” headphone? (oratory1990)
- REW help: Minimum Phase
- Why acoustic impedance matters (Jerome O’Flaherty, Pragmatic Audio, November 2025)
- Frequency Response Consistency (Rtings)
- Listener Preferences for In-Room Loudspeaker and Headphone Target Responses
- Segmentation of Listeners Based on Their Preferred Headphone Sound Quality Profiles
- Factors That Influence Listeners’ Preferred Bass and Treble Levels in Headphones
- How can cables affect the sound of a headphone/loudspeaker? Is this important? (oratory1990)
I’d like to thank Brandon and the staff at Dan Clark Audio as well as Mark and the team at Audeze for hosting these auditions.
A massive thank you to oratory1990. I have learned a great deal about headphones, measurement and correction from his writing.
I’d also like to thank Isaac Asimov, whose writing instilled in me both rigor and the need to follow a line of reasoning through to a practical solution. To Dr. Asimov, I apologize for dropping the Oxford comma.
To the lovely Nat Calkins, thank you for your company, your photography and your editorial eye.