The 711 vs. 5128 Coupler War: Why Your IEM's Graph Might Be Lying About the Bass

Here's something that breaks a lot of people's brains the first time they see it: you can take a single IEM, measure it on two different professional rigs, and get two noticeably different frequency response graphs — different bass levels, a different treble shape, sometimes a dramatic dip in one that barely exists in the other. Neither machine is broken. Neither reviewer is lying. They're just using different couplers — and the quiet shift from one standard to another is currently rewriting how the entire hobby reads sound. If you've ever been confused why two trustworthy graphs of the same earphone disagree, this is your answer.

What a coupler actually is

To measure an IEM, you need something shaped enough like a human ear to capture what an eardrum would receive. That something is a coupler — a precision cavity with a calibrated microphone where the eardrum would be. For decades, the industry standard has been a design built to the IEC 60318-4 spec, universally nicknamed the "711." It's essentially a carefully dimensioned metal tube that approximates the acoustic behavior of an average occluded ear canal. Crucially for our hobby, cheap clone versions of the 711 can be bought for around a hundred dollars, which is the single reason IEM measurement exploded into a hobbyist sport in the first place. Nearly every graph database you've ever browsed was built on a 711 or a 711 clone.

Enter the 5128 — and the problem it exposed

The newer standard, the Brüel & Kjær Type 5128, is a far more sophisticated and expensive machine — tens of thousands of dollars — built to a different spec (ITU-T P.57 Type 4.3). Instead of a simple tube, it models the acoustic impedance of a real human ear much more accurately, and it does so much further up into the high frequencies. Where the old 711 is only really rated as accurate to roughly the upper treble, the 5128 is rated across the full audible range.

And when reviewers started putting the same IEMs on both, the differences were striking. On the 5128, the same earphone typically shows noticeably less bass relative to the rest of its response, and a different — often more believable — treble shape above 5 kHz. The 711, it turns out, has its own built-in resonances and quirks that the 5128 doesn't share. The machine many of us trusted as ground truth was quietly imposing its own fingerprint on every graph it produced.

The "10 kHz hole" was largely a lie

Here's the most useful single takeaway from all this. For years, people stared at a sharp dip around 8–10 kHz on IEM graphs and concluded the earphone "lacked air" or had "rolled-off treble." It turns out that notch is, in large part, a coupler artifact — an effect of the 711's internal structure, not the earphone. It even moves around depending on how deep the IEM is inserted and which tips are used. The 5128, modeling the ear more accurately, often doesn't show the same dramatic hole. So a great deal of hand-wringing over "treble extension" was, all along, people reading the machine's quirk as the earphone's flaw.

The rule that protects you: never compare across rigs

This is the practical heart of the matter, and it's simple: you cannot directly compare a 711 graph to a 5128 graph. The bass-to-treble balance, the absolute levels, the shape of the highs — all of it shifts between the two systems, so laying one over the other and judging "this set has more bass" is meaningless if they came from different rigs. The same caution applies, to a lesser degree, even between two different 711 clones, which vary in quality. When you're comparing IEMs by their graphs — a skill we walk through in How to Read a Frequency Response Graph — make sure every graph came from the same source. A reviewer's database compared against itself is reliable; a screenshot from one site laid against a screenshot from another is a trap.

None of this means measurements are useless — quite the opposite. It means measurements are a tool with a known fingerprint, and using them well requires knowing which machine made them. The hobby is in the middle of a slow, messy migration from the old standard to the new one, with preference targets being painstakingly re-derived for the 5128 as we speak. For now, the takeaways are refreshingly practical: compare like with like, distrust the 10 kHz hole, and remember that the most accurate graph in the world still isn't your ear. If this is the kind of thing you find fascinating rather than exhausting, you'll enjoy our companion piece on what genuinely changes your sound versus what's audiophile folklore.