The most common headphone amplifier in the world was built for 250Ω

Anyone who records anything already owns a headphone amplifier: the socket on the front of their audio interface. There are 65 interfaces in this database out of 280 sources, which makes them the second largest category here, and by installed base they are almost certainly the largest one in the world.

They are also the category where one specification goes wrong so consistently that it stops looking like a fault and starts looking like a design brief. The median published output impedance of an interface here is 10Ω. For every other kind of source in this database it is 0.5Ω. That is a factor of 20, and it is not a rounding difference.

What 10Ω does

The one-eighth rule says a source should sit below an eighth of the headphone’s impedance, or it starts shaping the frequency response rather than just making it louder. The median in-ear in this database is 18Ω, so the ceiling is 2.25Ω.

An interface sitting at the category median of 10Ω satisfies that rule for 0 of the 222 in-ears here. Of the 43 interfaces that publish an output impedance at all, 15 clear it for an 18Ω in-ear. 29 clear it for a 250Ω studio headphone.

That second number is the explanation. These outputs are not badly designed; they are designed for the headphones a studio actually owns. A 50Ω output into a 250Ω monitor is inside the rule with room to spare. The same output into a 16Ω multi-driver in-ear is four times over it, and the in-ear was not on the drawing board.

The eight highest here

The last column is not a verdict on the product. It counts how many of the 567 models in this database that output impedance leaves alone under the one-eighth rule — and for the top of that table the answer is close to none of them. The simulator shows what the response actually does at those numbers.

It is not a price problem

The cheapest interface here that clears the one-eighth rule for an 18Ω in-ear is the IK Multimedia AXE I/O ONE at $130, which publishes under 0.1Ω — less than a hundredth of what the most expensive offenders publish. Output impedance is a design decision about what the socket is for, and some makers decided it should be for headphones. Reading down the two tables, the thing that predicts the figure is not price and not brand prestige: it is whether the maker treated the headphone output as a monitoring convenience or as an amplifier.

Power is not the problem

The median interface here delivers 112 mW into 32Ω against 380 mW for everything else, so on paper it is the weaker half of the database. In practice that rarely matters, because the median headphone needs 1.0 mW for 100 dB SPL. The median interface reaches that on 561 of the 567 non-electrostatic models here, and the weakest one in the category — the Steinberg UR22C at 12 mW into 32Ω — still reaches it on 512 of them. Every interface here clears nine tenths of the catalogue.

Driving everything is a different bar, and only 6 of the 65 clear it, because the last few percent of this database is HiFiMAN planars and low-impedance in-ears that want more current than a monitoring socket was ever asked for. But if your interface sounds wrong with in-ears, power is the least likely explanation on the list.

The specification none of them publish

Output noise is the other half of in-ear matching, and this is where the category is at its worst. 4 of 65 interfaces publish an output noise figure. Across every other kind of source in this database, 79 of 215 do. The exception is the RME Babyface Pro FS and the RME ADI-2 DAC FS and the RME ADI-2 Pro FS R and the RME ADI-2 Pro EX.

So for 61 of the 65 interfaces here, the question “will this hiss with my in-ears?” cannot be answered from the manufacturer’s own documentation at all. Not answered badly — not answerable. This site leaves those blank rather than filling them with someone else’s measurement, which is why the hiss estimate is missing on most interface pages.

What to do about it

If you use an interface with full-size headphones of 80Ω or more, none of this is your problem and the numbers above are trivia. If you use it with in-ears, three things are worth knowing: check your own model on the output impedance checker, remember that an in-line attenuator lowers hiss but raises the impedance the headphone sees — the attenuation calculator shows both effects — and that the full list of interfaces here is sorted by price with both figures beside it.

And if a maker does not publish the output impedance of a socket they sell as a headphone output, that is itself information. 22 of these 65 do not.

Every headphone output impedance these makers publish

The one-eighth rule says a source suits a headphone when its output impedance is no more than an eighth of the headphone’s. Run it over the 65 interfaces here and the category separates into two halves that have nothing to do with price. The median published output impedance of an audio interface is 10 Ω; for every other kind of source in this database it is 0.5 Ω, about 20 times lower. 43 of the 65 publish a figure at all.

The second column is the consequence and no maker prints it: where the rule stops being satisfied, and what share of the 567 non-electrostatic models in this database is left on the right side of it. At the bottom of the table the answer is a handful of models, all of them 400 Ω and above.

The 9 entries at fifty ohms are the interesting ones, because that group contains the cheapest and most widely owned interfaces on the list — the Audient EVO 4 at $129 among them. Fifty ohms satisfies the one-eighth rule on six of the 567 models here: Beyerdynamic DT 880 Edition (600Ω), Beyerdynamic DT 990 Edition (600Ω), Beyerdynamic T1 2nd Gen (600Ω), Audio-Technica ATH-R70X, Audio-Technica ATH-R70xa and Audio-Technica ATH-ADX5000. Nothing below 420 Ω qualifies, which means the 300 Ω studio classics do not, and an in-ear is not in the conversation at all. The worst case is the Tascam US-2x2HR at 66 Ω, which wants 528 Ω and therefore satisfies the rule on three. At the other end the IK Multimedia AXE I/O ONE publishes 0 Ω for $130 and is inside the rule on every model in the database.

None of this makes a high figure a design fault, and the engineers who chose it were not being careless. A monitoring socket was specified to feed studio headphones of two hundred and fifty ohms down a long lead, and a resistor in series there costs a fraction of a decibel while buying a kinder load and some protection against a short. What changed is not the socket but who plugs into it: tracking vocals on a four-driver in-ear puts eight or ten ohms across an output designed for thirty times that, and the fraction of a decibel turns into an audible change in tone. The column exists to say which of those two situations you are in.

The 22 that publish no output impedance at all

This is the half of the question that sends people searching, and the honest answer is that the number does not exist in the maker’s own documents. It is not scattered at random either: Universal, Steinberg and PreSonus publish no headphone output impedance for any interface in this database. MOTU publishes one for MOTU UltraLite-mk5 — 1 Ω — and for MOTU M2, MOTU M4 and MOTU M6, the three that most people actually own, it publishes nothing.

There is nothing to deduce from a blank. An unpublished output impedance can be half an ohm or it can be fifty, and the only way to find out is to measure it, which this site does not do. What it does mean is that for those 22 interfaces the one-eighth rule cannot be run at all, so the three-axis verdict elsewhere on this site leaves that axis empty rather than guessing.

Built for 250 ohms, in the makers’ own numbers

The claim at the top of this page is usually made as an opinion. It can be made as arithmetic instead, and only from figures the makers published themselves. 27 of the 65 interfaces here print a power rating at two or more loads. Turn each of those into the voltage it implies — V = √(P × Z) — and a true voltage source gives the same answer at both ends.

It does not. The median interface implies 2.08 times more voltage at the highest load it publishes than at the lowest. For the 114 non-interface sources that publish two or more loads, the same median is 1.29. 25 of the 27 interfaces are at one and a half times or more, against 39 of 114 everywhere else.

Read it the other way round and it is the useful statement: these sockets run out of current long before they run out of voltage. That is why an interface that looks weak beside a dongle on a 32 Ω in-ear can be the better of the two on a 300 Ω headphone, and why the figure a maker quotes at 32 Ω tells you so little about the socket.

Interfaces are not alone in this — several valve and high-output desktop amplifiers spread further still, for the same reason. What is unusual about the category is how consistent it is: almost every interface that publishes two loads behaves this way, and almost nothing else does.