Four Bone Conduction Headphones, 136 Miles, and the 11 Minutes I Didn't See Coming

Thirty-eight percent. That's the cut on the Shokz OpenRun as I file this — $129.95 down to $79.95 — and by my own tracking it's the lowest that model has gone in fourteen months, Black Friday included. I've spent eleven weeks and roughly 136 miles in bone conduction headphones, four pairs of them, plus one open-ear set that isn't bone conduction at all and got included so I'd have something to argue with. So I can tell you whether the discount is real.

The verdict, since that's what you came for: buy the OpenRun at $79.95 and skip the Pro 2. The Pro 2 costs $65 more and spends that money on low-end response, a better call mic, and two extra hours of battery — three things that matter in an office and don't matter at mile nine.

What took me longer to work out was where the sound actually goes once it's in your head. That turned out to matter more than the fifty dollars, and it's the reason this piece is running in a magazine about sleep.

The field, and what I paid

Prices move. These are what my tracker showed the morning I filed, alongside the lowest I've personally logged for each.

Every unit was bought, not sent. I ran in all of them on the same 6.2-mile loop — asphalt path, two road crossings, one exposed half-mile where the wind comes off the reservoir and ruins any audio test you were hoping to run.

First contact: a transducer on your cheekbone

Nothing goes in your ear. The pads sit on the zygomatic arch, the ridge of bone running forward from just in front of the ear opening, and the whole product is a bet that a small motor pressed against that ridge can shake your skull into hearing something.

The first thing you notice is that the bet requires pressure. The band has to clamp, and clamp is where these succeed or fail as running gear. The OpenRun's titanium band held its tension unchanged across eleven weeks. The Haylou's did not — by week six it had relaxed enough that the left pad drifted off the bone on downhills, and once the pad slides onto soft tissue the volume drops by what felt like a third. Not a subtle degradation. It's the difference between hearing a podcast and hearing that a podcast is happening somewhere nearby.

Weight matters less than clamp. All four sit in a 26–30 gram range and I could not reliably tell them apart blind on weight alone. What I could tell apart, immediately, was band geometry against glasses. With flat-temple frames the band tucks underneath and you forget it's there. With wraparound sport sunglasses — the kind with a thick curved arm — the band fights the arm for the same real estate and one of them wins. On the Pro 2 the loser was my sunglasses, which crept down my nose for six miles. If you run in wraparounds, try the combination before the return window closes.

The other physical fact nobody puts on the box: above roughly 70% volume, low frequencies stop being sound and start being sensation. You feel a fine buzz against the cheekbone. The Haylou does this earliest and hardest; the OpenRun tolerates about 80% before it starts tickling; the Pro 2 is the most composed, which is the first hint of what its extra $65 actually bought. If you're the kind of listener who runs things loud, budget for that ceiling being lower than you expect.

Meanwhile, your ear canal is wide open

Here's the part that's genuinely different, and it's not the bone conduction — it's the hole.

Your ear canal stays completely unblocked. So while the transducer is shaking your skull, the ordinary world is still arriving at your eardrum by the ordinary route. Two signals, sharing one set of ears, summing before your brain gets a chance to sort them out. That's the whole product. Everything good and everything bad about open-ear audio comes from that arrangement.

The bad first. Wind. On the exposed reservoir stretch, into a headwind I'd guess at 12 mph, turbulence in the outer ear generates a broadband roar that lands directly on top of the frequencies a podcast lives in. There's no isolation to fall back on, so you do the only thing available: turn it up. I ended up at 75–85% on that half-mile and back down to 50% the moment I turned.

That volume is where leakage becomes a real objection, so I measured it. Uncalibrated phone SPL meter, one meter away, my partner holding the phone in a quiet room reading 39 dBA at rest. At my 50% cruising volume: 44 dBA, and they described it as "a faint tsst." At the 80% I use into wind: 52 dBA, and the description became "cymbals and consonants — I could tell it was a man talking." Treat those numbers as directional, not laboratory. The finding that survives the imprecision is simple: at running volumes these are private; at windy-day volumes they are not. Don't take any of them onto a quiet gym floor and turn them up.

The good is the same mechanism read the other way. Nothing is compressing your ear canal, so there's no occlusion effect — no boom of your own footfalls conducted through your jaw, none of that underwater pressure that makes sealed earbuds feel like a helmet by mile eight. After a two-hour run in the OpenRun I took them off and nothing changed. My ears didn't pop, didn't feel warm, didn't feel used. That absence is hard to sell in a product listing and it's the single strongest argument for the category.

The cochlea does the mixing, and that's where the bass dies

Skull vibration reaches the cochlea directly, skipping the eardrum and the three small bones behind it. Those bones are a lever system — a mechanical amplifier that exists precisely because moving fluid takes more force than moving air. Bypass the lever and you lose the amplification. That loss is not evenly distributed. It falls hardest on low frequencies, which need the most displacement.

Which is why the bass isn't weak on these. It's largely gone.

The test track that made this unarguable was "bad guy" — the sub-bass line that carries the entire song. On the OpenRun it isn't quiet, it's absent; the track becomes a vocal, a snap, and a lot of implied space where the hook used to be. On the Haylou, same, with a little cheek buzz standing in for it. On the OpenSwim Pro, same again.

On the Pro 2 it's actually there — thin, but present, with a real sense of pitch. And that's because the Pro 2 quietly stops being a pure bone conduction device. It adds a small air-conduction driver aimed at your open ear canal to fill in the lows. It's a hybrid, and it works, and it comes with a tax: the Pro 2 is the leakiest unit here at matched volume, because bass radiating into the air is bass radiating into everyone's air.

Now, the useful reframe. Speech energy lives roughly between 200 Hz and 4 kHz, and that band is exactly where these are competent — clear, articulate, entirely sufficient. Every one of the four handled an audiobook at conversational volume with no strain. These are voice-content headphones that will also play music, not music headphones that also do podcasts. Once I stopped auditioning them against my sealed earbuds on albums and started using them the way I actually use them — talk radio, audiobooks, race-day playlists I know well enough to hear the missing parts from memory — the complaint dissolved.

The OpenSwim Pro deserves a line of its own. It has 32 GB of onboard storage and a genuine swim rating, and underwater it is not a compromise, it's the only thing that works — water in the canal makes air conduction hopeless while making bone conduction, if anything, better. I did nine pool sessions with it. If you swim, it's a different purchase with a different justification and it isn't competing with the OpenRun at all.

Further up the line: the reflex that hears the car

Signal leaves the cochlea, runs the auditory nerve, and hits the brainstem before your cortex has any opinion about it. That's the level where startle and localization live, and it's the level where open-ear earns its safety reputation.

Locating a sound depends on comparing what arrives at each ear — timing and intensity differences, plus the way your outer ear colors sound arriving from different angles. Put anything in the canal and you distort that coloring. Even good transparency mode is a reconstruction: microphones outside, drivers inside, a few milliseconds of processing in between, and your pinna cues rebuilt by software rather than delivered by anatomy.

I tried to test this honestly and only half succeeded. On a shared park path over three weeks I logged every cyclist who called "on your left," and whether I registered it on the first call. Open-ear: 28 first-call responses out of 30 passes. The AeroFit 2 in its vented, no-ANC configuration: 19 of 30. I knew which device I was wearing every single time, so treat that gap as an impression with arithmetic attached rather than a result. But the qualitative difference wasn't marginal. With open-ear I knew where the voice came from. With the earbuds I knew a voice had happened and had to turn my head to find out where.

Road crossings were the same story in a form that matters more. Car direction was immediately obvious with all four bone-conduction units, at every volume I tested including the windy-day 80%. That's the purchase argument for anyone who runs on roads, in the dark, or anywhere a bike might come from behind — and it holds regardless of which of these you buy, which is worth saying plainly, because it means the safety case is not a reason to spend more.

Last stop, four hours later

Here's what I didn't expect to be writing about.

Eleven weeks of evening runs, finishing between 7:15 and 7:50 p.m., lights out between 10:30 and 11:15. Thirty-eight of those nights I logged properly: run type, audio, a written lights-out time, and my wearable's sleep-onset estimate the next morning. Caffeine held constant throughout — nothing after 2 p.m., no exceptions, because I wasn't going to write this up and then have to blame the espresso.

Runs split into two audio conditions that emerged on their own rather than by design. Talk audio — news podcasts, one long argumentative interview series, two true-crime seasons. And everything else — instrumental music, familiar albums, or silence.

Median sleep onset after talk-audio runs: 24 minutes. After music or silence: 13 minutes. An eleven-minute gap, consistent enough across the log that I noticed it before I went looking for it.

I want to be careful about what that is and isn't. It is n=1. It is a consumer wearable's estimate of sleep onset, not a lab measurement, and those estimates are known to be soft at exactly the boundary I'm measuring. The nights weren't randomized — I chose the podcast when I felt like a podcast, and "felt like a podcast" is not independent of what kind of day I'd had. And the effect I'm describing is almost certainly about the content, not the hardware. An hour of a well-made argument at 7 p.m. leaves you holding the argument.

But the hardware is why the content was there. That's the honest chain, and it's worth following. Sealed earbuds made evening talk audio uncomfortable enough on a road route that I mostly didn't — the isolation felt unsafe and I'd take them out at crossings. Open-ear removed that friction completely. Something that used to happen once a week started happening four times a week. The device didn't wake me up. It removed the reason I wasn't doing the thing that did.

I've kept running in the evenings. I switched to instrumental after 7 p.m. and my median went back to 14 minutes over the following three weeks. That's not a controlled result either, and I'm not going to dress it up as one.

What I couldn't test: anything at all about long-term hearing. Bone conduction is often assumed to be gentler because it skips the eardrum, but it's still delivering energy to the same cochlea, and eleven weeks tells you nothing about years. Nobody reviewing headphones can tell you that, including the ones who imply it.

Is $79.95 actually the floor?

The question underneath the buy. Fourteen months of my own price notes on the OpenRun: it sat at $129.95 most of the year, dipped to $99.95 for a spring event, went to $89.95 over Black Friday week, and came back to $99.95 in July. Today it's $79.95.

So this is not a repeat of a known low — it's twelve and a half percent below the best I have ever recorded for it, and the record it beat was Black Friday's. The generational logic supports that: this is the second-gen OpenRun with the newer Pro 2 sitting above it in the line, and this is the phase where the older model gets cleared rather than protected. Waiting three months for November to maybe match $79.95 is not a plan, it's a hope with a deadline.

The Pro 2 at $144.95 is a different situation. It's $5 above its own record low and it's the current flagship, so it will be discounted again, and harder, when the calendar turns. If the hybrid bass driver is genuinely what you want, wait. It'll be cheaper.

The Haylou at $47.99 is above its March low and I wouldn't chase it at any price given what its band did by week six.

Who this is for, and who it isn't

Buy these if you run on roads or shared paths, if you listen to podcasts and audiobooks more than albums, if sealed earbuds leave your ears aching after long efforts, or if you've ever taken a bud out at an intersection. That last one is the tell. If you're already removing earbuds for safety, you've been running with the wrong hardware and this fixes it for eighty dollars.

Don't buy these if you train mostly indoors on a treadmill in a quiet gym, where you'll either annoy people or run too quietly to hear anything. Don't buy them if bass is the reason you listen to music while running — you will hate them and no amount of adjustment will change the physics. Don't buy them if you want isolation on a flight; there is none, by design. And don't buy the bone-conduction models for swimming — buy the OpenSwim Pro, which is a different product wearing the same shape.

If you wear wraparound sunglasses, this is a try-before-you-commit purchase rather than a click-buy.

The winner

The Shokz OpenRun at $79.95 is the buy, and the $65 you'd spend upgrading to the Pro 2 buys bass you can't hear over your own footsteps.

Second place goes to the OpenSwim Pro, but only for swimmers, and only because nothing else in the category can do what it does. The Haylou is a real product at a real price that came apart mechanically inside two months. The AeroFit 2 sounds better than all of them and localizes worse than all of them, which tells you exactly which problem each design set out to solve.

Back to the thirty-eight percent

It's a real discount. I've checked it against fourteen months of my own notes and it's the lowest the OpenRun has been, by a margin, in a window that includes a Black Friday. Fifty dollars saved, and the case for buying now rather than in November is as solid as this sort of case ever gets.

It's also the least interesting number in this piece.

The number I'd actually make the decision on is eleven — the minutes that appeared between lights-out and sleep on the nights I ran with a good argument in my ears, in a log too small and too messy to prove anything, and clear enough that I changed what I listen to after seven. Open-ear headphones are unusual in that they don't so much add something to your run as remove an obstacle, and what rushes into the space is whatever you were already inclined toward. For me that was more talking, later in the day, all the way to the pillow.

Buy them at $79.95. Then decide what you're going to put through them, because that's the part the discount doesn't cover.