NKT

NKT

Soundstage junkie

Philosophy

The best IEMs I have heard behave like a high-end display with a professional calibration. They get out of the way. They reproduce what is on the recording with the precision that every transient starts at the intensity it should, holds for as long as it should, and decays as the recording calls for. They let you forget there is a chain in between, the same way a good monitor lets you forget you are looking at a panel.

This is what mastery looks like in audio reproduction. Faithful rendering is the purpose. Mastery is the gentle, smart bending of the rules to make what is reproduced a bit nicer, without breaking it. IEMs that bend the rules well are the interesting ones. IEMs that bend the rules just to be different are not.

I write reviews to find them.

What I look for

Immersion, in the most literal sense. The orchestra around my head. The cello present and dark in the lower-midrange, not thin. The cymbal rendered with stick, shimmer, and decay rather than as a splash of high-pitched noise. Each instrument pannable in space, near or far, but never flat. The bass attacks land like clean punches, tactile, fast, gone. Detail surfaces without effort. The soundstage is 3D.

I am not picky about tonality. A well-judged warmer presentation is fine if it is done without sacrificing the precision of transients, soundstage, and resolution. What I reject is the idea that a particular tonal flavour is, on its own, a measure of quality, that an IEM is “musical” or “non-musical” because of how it tilts at 3kHz.

I treat tonality as preference. I grade on the things that are not preference: whether transients are clean, whether the soundstage has depth, whether separation holds up under a busy mix, whether detail emerges when the recording asks for it.

How the ratings work

The 1-to-5 scale is not about my taste. It came from a “championship” of A/B tests I ran a few years back, where I pitted reference IEMs against each other to identify which one sits at the entry of each rating band.

  • The AFUL Cantor sits at the entry gate of the 5/5 band. It replaced the 64 Audio U12T in that role.
  • The CFA Andromeda 2020 sits at the entry gate of the 4/5 band.
  • The Moondrop Blessing 2 sits at the entry gate of high 3.
  • The Moondrop Aria sits at the entry gate of low 3.

When a new IEM comes in for review, I A/B it against these gate IEMs to place it. A 4/5 here does not mean “as good as Andromeda 2020.” It means “this IEM is in the same performance band as the gate, on the dimensions I grade.” The gates have largely held their positions as the market has improved, with Cantor being the one explicit replacement I have made so far.

The full method, including the four dimensions I rate on (Timbre and Tonality, Bass and Dynamic, Resolution, Soundstage and Imaging), the comparisons, and the test tracks, is on the methodology page.

What I believe

A few commitments follow from the calibration view.

  • Sound is physical. Drivers vibrate, air compresses, the eardrum bends. Anything I hear has a physical cause. Vendor claims about how a crossover is “tuned” or how a nozzle “preserves phase” are interesting, not dismissible. When an engineer describes a design decision, I want to engage with it. The interesting ones are the ones I can trace, measure, or verify by listening carefully.

  • I run my own measurement rig, and that is what makes the difference. I use a 711-coupler microphone, the same widely-used standard in this hobby. Everything in my graph database is my own measurement, not somebody else’s. Frequency response encodes most of what I care about: tonal balance, technical capability, resolution floors, even some resonant pathologies. When the graph and my impression disagree, I get curious, because one of the two is wrong, or I am missing something. I do not let the measurement tell me I am wrong, and I do not let my impression tell me the measurement is broken.

  • Tuning shapes the tone, and engineering is what lets the tone come out cleanly. Tuning changes the overall tone. It can also influence perceived technical performance, since technical performance, in my view, is right there encoded in the frequency response, nowhere else. The engineering of the drivers, crossovers, shells, and nozzles is what allows a tuning to deliver without compromising on transient precision and the rest. A well-engineered IEM stays technically capable even when it is tuned warm and dark for a comfortable listen. A poorly-engineered IEM will sound mushy and lacking in detail even when it is tuned cold and thin and shrill. “Musical” lives in tuning. “Technical” lives in engineering. The interesting IEMs are the ones that have both.

  • Price is informational, not determinative. Think of it like a character-creation screen in an RPG. A higher price means more R&D budget, which translates to more “dice rolls” for engineering decisions. You might roll three sixes and end up with a “worth-it” TOTL IEM. You might also roll three ones and end up with a $5000 IEM that sounds worse than mediocre. On average, more money buys more performance, but the variance is real and the distribution is wide.

  • No IEM is universally the best. Every IEM is a set of trade-offs. Some make classical bloom and electronic lifeless. Some make orchestral magical and vocal dry. My job is to help you figure out which trade-offs sit well with the music you actually listen to.

Why this site exists

I call this blog In-ear Gems because, a few years back, I was watching chifi IEMs improve at an almost alarming rate. Many were tantalisingly close to the established flagships but consistently missed something on the technical side, transients, imaging, separation. I started writing here to track which ones were the actual gems, and to keep a record of when an average chifi IEM would catch up to the established benchmarks.

That is still the project. The reviews on this site are the record of that hunt.

What this means for the writing

  • I write reviews because, first and foremost, it is a productive way to justify owning enough IEMs to test them properly. If I have them anyway, I may as well write them up.

  • I write the review I would want to read. Heavy on technical performance. Heavy on explainability. A blend of objective measurements and subjective impressions, in that order.

  • Bias exists. I am not going to pretend it does not. That is why I disclose the source of every unit I review (loan, tour sample, self-purchase), so you can weigh what I say accordingly. Anyone who tells you their judgment is unaffected by source is selling you something.

  • I will not lie to flatter a manufacturer. I will also not lie to flatter the anti-manufacturer crowd by calling a good IEM bad. If something genuinely deserves praise, it gets praise.

  • I am not interested in moralising. There is enough of that elsewhere in this hobby. I want to listen to things and tell you how I hear them, in as much detail as I can. That is one more data point for your own assessment.

What I am trying to do

If you finish a review and you can picture what the IEM does to your favourite track, that is the goal. If you finish a review and you decide this one is not for you, that is also the goal. The point is not for you to agree with me. The point is for me to make the listening experience legible enough that you can decide.