Dan Clark Audio Stealth

Five rig readings across five labs agree on the shape: ASR's GRAS 45C, Resolve's GRAS 43AG with anthropometric pinna, TechPowerUp's IEC711-class rig on a randomly selected retail unit, SBAF's MiniDSP EARS averaged over five pad positions, and Head-Fi HQ's Brüel & Kjær 5128 with an APx555. ASR reports the only meaningful deviation as a small bump between 100 and 160 Hz that the company describes as deliberate. Channel matching is specified at 0.25 dB weighted from 20 Hz to 10 kHz, and TechPowerUp's retail sample measured better matched than anything else he had tested.

Both features are measured. SBAF's EARS traces read a low bass shelf of about +5 dB at 20 Hz and a slight recession between 200 and 300 Hz; Resolve identifies the same dip on the GRAS 43AG as the target's own upper-bass contour and reports it as the source of the midrange clarity; TechPowerUp reads the same region as the headphone's one weak spot, calling the upper bass and lower mids a touch lean on harder-hitting tracks.

⚠ vs. listeners — How much of that shelf reaches you is a seal question, and it is the one thing about this headphone the rigs genuinely disagree about in practice. Resolve measured the leak cases directly and got full bass through thin-armed glasses; SBAF's own listening conclusion was that the lows and mids are very seal-dependent and that glasses wearers should avoid it; an owner in the same thread reports the opposite, that the pads warm up and conform, and that only thick frames are a problem. Resolve also notes the driver's resonance sits high, around 300–400 Hz, so a broken seal costs everything below that.

Resolve gives the mechanism rather than just the impression: the mid-treble is deliberately relaxed, which makes the upper-midrange tones more noticeable by contrast, and the feature moves with headphone position — Dan Clark Audio's own guidance is to seat the headphone so the back top of the ear, around ten o'clock, meets the rear of the pad. TechPowerUp arrived at the same positioning advice independently.

The rigs disagree about the one feature the argument turns on. Headfonics reads a nuanced 8 kHz bump as real and welcome; TechPowerUp, measuring on an IEC711-class coupler, thinks the AMTS Helmholtz resonator is interacting with his rig and that the peak is largely an artefact — he expects most listeners to hear a slight dip with peaks around 6–8 kHz, smoother than any published graph shows. Nobody measures it as bright: the disagreement is about what is absent.

⚠ vs. listeners — There is also a wear-in variable no graph of a new unit can show. An owner who swapped in fresh pads after two years reports that compressed pads had been darkening the top end, and that new ones brought clarity back — so part of the 'too dark' camp may be hearing older pads rather than the tuning.

SBAF measured the trade and then stated it as a trade: the slight recession between 200 and 300 Hz 'greatly improved soundstaging and imaging clarity at the cost of small loss in perceived notion of bass slam'. That single sentence accounts for most of this argument — the clarity everyone praises and the physicality a minority misses are the same design decision seen from two sides.

⚠ vs. listeners — This is the clearest case here of graphs failing to settle a listener. The critic who A/B'd it against an EQ'd LCD-5 and returned it drew exactly that conclusion — 'This proves to me beyond a doubt that you cannot compare headphones with Frequency response graphs alone'. Two practical findings cut the other way, though: TechPowerUp reports clamp pressure and ear position measurably change how dynamic it sounds, and an owner who pushed a large bass boost through it found the drivers took it without complaint, which the near-zero measured distortion predicts.

No rig measures stage, but there is a documented physical reason to expect the effect: the AMTS structure is designed partly to break up the standing waves and cup resonances that give closed-backs their boxed-in character, and the measured decay behaviour shows no obvious resonant signature for them to sit on.

The objective substrate is not in doubt. ASR measured distortion at 94 dBSPL and 2 kHz low enough to convert to a SINAD of 81 dB — a figure that includes the measurement microphone and the noise floor, not just the headphone — and describes the gap below the reference as frequency-independent, which points away from resonances. Head-Fi HQ's 5128 plots read harmonic distortion as extremely low for a non-electrostatic headphone.

⚠ vs. listeners — The dissent is not about the distortion figures; it is about which half of 'detail' you count. Resolve splits the quality in two — clarity for the trailing ends of tones, and separation during busy passages — and rates the Stealth competitive with anything on the second while a step behind its open-back rivals on the first.

415 g on the scale, well under the 600–700 g typical of the flagships it competes with, and the reason several reviewers report full working days in it. The jaw-pressure note is replicated: Resolve records mild pressure where the bottom of the cups meets his jaw and suggests rotating the headband forward, and separately notes the back of his ears touching the AMTS waveguide inside the cup.

SBAF ran an explicit noise-reduction test — a loudspeaker sweep measured with and without the headphone on the fixture — and benchmarked the result against studio headphones: 'Compared to other models, for audiophile headphones, the Stealth exhibited good noise reduction ability contestable to mixing headphones.' It beat an Ollo S4R by 3–5 dB across most of the range and lost to a Sony by 5–10 dB through the upper mids.

The two recurring serviceability complaints are structural rather than a matter of taste: the pads are adhered rather than clipped, which Headfonia notes is inelegant but makes the best airtight seal, and the Hirose connectors are uncommon enough that aftermarket cables are harder to source than for a mini-XLR or 3.5 mm headphone.