What impedance and sensitivity actually mean for your IEMs

Every product page in my shop lists two numbers that almost nobody explains: impedance, in ohms, and sensitivity, in decibels. People see them, decide they're for engineers, and scroll past to the driver count. I get it. But those two numbers are the whole answer to "do I need a DAC for this?", and if you understand them you stop buying sources you don't need and start buying the one you do.

Before anything else: I sell every earphone and every player named here. That's not a disclaimer I'm hiding at the bottom, it's the reason to read the rest carefully. Where the honest answer is "your phone is fine", I've said so.

What impedance really tells you

Impedance is how much the earphone resists the signal your player sends it. Low impedance means the set draws current easily and gets loud on very little. High impedance means it asks for more voltage to reach the same volume.

For full-size headphones this is the big one, because 300-ohm sets genuinely need an amplifier. For in-ear monitors it almost never is. Look at the range in my own catalogue: the Unique Melody Maven III Moonshard sits at 8.8 Ω and the Dunu DN142 at 37 Ω, and both will play plenty loud from a €10 dongle. Nothing I stock is hard to drive in the way people mean when they say hard to drive.

What impedance actually predicts for IEMs is something more interesting, and I'll come back to it under output impedance, because that's where the real money is.

Why the sensitivity number is a mess

Sensitivity is how loud the set plays for a given input. Higher number, louder for less. Simple — except manufacturers publish it in two different units and rarely tell you which one matters.

dB/mW is loudness per milliwatt of power. dB/Vrms is loudness per volt of signal. They are not interchangeable, and comparing them directly will send you to the wrong conclusion every single time.

Here's the clearest example I have, and it's two products from the same brand sitting on the same shelf in my shop:

• Dunu Kima 2 — 32 Ω, 108 dB/Vrms
• Dunu DN142 — 37 Ω, 107 dB/mW

Those look like the same earphone from a power point of view. They are not remotely the same. Convert the Kima 2 to the DN142's unit and it comes out around 93 dB/mW, which makes the DN142 roughly 14 dB more sensitive. In plain terms, the Kima 2 wants about twenty-five times more power than the DN142 to hit the same volume. Same brand, same price bracket, same spec table format, and the numbers are quietly telling you opposite things.

This is why I won't build a comparison tool on my site until I've normalised the whole catalogue to one unit. Half my spec sheets say dB/mW, some say dB/Vrms, and several just say "dB" with no unit at all, which is the manufacturer's fault rather than mine but still ends up on my page. I'd rather tell you that than pretend the table is cleaner than it is.

The two ways a source actually fails

Not enough is the obvious one. The volume sits at 90 out of 100 and the music still feels flat and small, with drums that push instead of hit. That's a source running out of headroom, and it's almost always voltage running out rather than watts.

Take the Kima 2 again. At 108 dB/Vrms it needs roughly 1.25 volts to reach 110 dB peaks, and most phone dongles top out somewhere around one volt. So the set that measures as easy on paper is the one that makes a cheap dongle sound tired, while the HiBy R4 delivers 2.3 Vrms from its 3.5 mm output and 4.1 Vrms balanced and never comes close to the limit. It isn't about power. It's about having voltage in reserve.

Too much is the failure people don't expect. A very sensitive set on a noisy source gives you hiss in the quiet parts, a volume control that goes from silent to loud between step one and step two, and every bit of interference your phone generates. The sets in my shop most likely to find a noisy source are the all-balanced-armature ones, the Letshuoer Mystic 8 and the 7th Acoustics Supernova, because BA drivers are efficient and reveal everything sitting behind them. FatFreq puts an impedance adapter in the box with the HBB Deuce for exactly this kind of pairing, and an adapter works by throwing away some sensitivity to buy back a clean background and usable volume travel.

Output impedance, and the number nobody prints

This is the one that actually changes how your earphones sound, and it's the number that's hardest to find.

Your source has its own impedance on the way out. When that output impedance isn't much lower than the earphone's, the earphone's own impedance curve starts shaping the frequency response. A single dynamic driver has a fairly flat load, so a sloppy source mostly just makes it quieter. A multi-driver set with a crossover has a load that swings around across the frequency range, and a sloppy source turns those swings into audible bass bloom or thin, recessed mids.

The rule of thumb is a ratio of one to eight: your source's output impedance should be under an eighth of the earphone's. Run that against my own lowest-impedance sets and it gets strict fast. The Maven III Moonshard at 8.8 Ω wants a source under about 1.1 Ω. The MEST Fortune at 10 Ω and the MEST Jet Black at 12.6 Ω aren't far behind, and all three are multi-driver flagships with crossovers, bone conduction and electrostatic tweeters in the mix.

So the €2,295 earphone is the one that's fussy about the €50 dongle, not the €120 one. That's backwards from what most people expect, and it's the single most useful thing on this page.

HiBy doesn't publish output impedance for the R4 or the R6 III, which annoys me more than it probably should. I'm not going to put a number on my product pages that I haven't measured myself, so for now I'll say what I can stand behind: both are quiet enough with the flagships that I demo them that way in Utrecht, and you're welcome to come and hear whether you agree.

What scaling with power actually means

More power does not add bass. It does not fix a lean set, it does not make a bright set smooth, and it will not turn a €150 earphone into a €600 one. Anybody selling you a source on that promise is selling you a story.

What better amplification does is hold control when the music gets busy. Bass notes stop and start where they should instead of smearing into each other, the stage keeps its shape at volume, and dense passages stay readable. You notice it most on planars, which draw real current and are the sets in my shop that most obviously wake up on a proper output — the Letshuoer S12 Ultra at 16 Ω and 101 dB, and the Astralis at 20 Ω and 104 dB. Planars aren't my personal favourite tuning and I'll say that openly, but they're the clearest demonstration of the effect I can hand you.

My two players, and the thing my own shop gets wrong

I stock the HiBy R4 and the HiBy R6 III (2025), and the natural assumption is that the more expensive one drives more. It doesn't, and I'd rather tell you than let you find out at home.

The R4 is the more powerful player by about 30%, and it's the cheaper one. What the R6 III buys you is composure rather than muscle: a quieter background, the Class A/AB switch to shift the tonal character, and line out so it can sit on your desk feeding an amplifier.

So I pair them by character, not by price. The R4 is my energy source, and it's the one I reach for with warm, smooth or rolled-off sets that want waking up. The R6 III is my composure source for bright, technical and sensitive sets where the last thing you want is more heat. One pairing I actively avoid: the Hisenior Mega5EST Bass+ on the R6 III in Class A. Warm amplifier into a bass-lifted set is too much of the same thing, and it turns a set I like into a set I don't.

If you don't want a player at all, the HiBy W4 does 125 mW single-ended and 475 mW balanced, takes Bluetooth from your phone or USB from your laptop, and clips to your shirt. It's the answer for most people who asked this question in the first place.

The whole catalogue, as published

These are the manufacturer figures exactly as they appear on my product pages, units and gaps included. Sorted by impedance, because that's the axis that matters for source matching.

Four sets publish nothing useful at all. I've left those rows honest rather than filling them with figures I found on a forum.

So what do you actually need?

If you're under €300 and using a phone, a decent dongle is enough and the money is better spent on tips. The Mega7, the Explorer and the DN142 all run happily off a basic USB-C dongle, and the DN142 in particular is seven drivers that ask for nothing.

If you own something multi-driver above €700 — the MEST Jet Black, the Fortune, the Maven III, the Dawn-X — the source stops being optional. Not because it needs power, but because at 8.8 to 15 ohms a sloppy output impedance will change the tuning you paid for.

And if you're on a planar, more current is the one upgrade that does something you can hear on the first track.

The thing I'd rather you do than read another spec table is bring the earphone to Utrecht and plug it into both players yourself. Fifteen minutes settles what an afternoon of forum posts won't.

What's the source pairing that surprised you — the one where the cheap thing beat the expensive thing?