Fifteen Meters: Measuring What Shokz Headphones Actually Let You Hear

Fifteen meters.

That is the gap I measured on a flat greenway at 5:50 on a Tuesday morning — the difference between how far away I could hear an e-bike coming up behind me while wearing Shokz headphones versus sealed in-ear buds playing the same podcast at the same measured loudness. Twenty-one meters against six. The bike was holding 25 km/h. Fifteen meters at that speed is about 2.2 seconds: roughly the time it takes to register a sound, look over your left shoulder, and move a meter to the right.

The verdict: bone conduction buys a real, measurable margin of awareness, and you pay for it in bass, volume headroom, and about $180.

Everything below is me working out what that number is worth — and, more usefully, what it doesn't cover.

Where the number came from

I have run in bone conduction for four years and I had never actually tested the claim I repeat to everyone who asks. So I tested it, badly but honestly.

A friend on an e-bike made repeated approaches from behind along a straight, empty greenway. I walked a marked lane with traffic cones set at three-meter intervals over 45 meters. When I first heard the bike, I raised a hand; the rider noted which cone they were level with. Twelve passes per configuration, four configurations, forty-eight passes total, between 5:40 and 6:35 a.m. Wind was 4 km/h on a handheld meter. Surface was dry asphalt. Same podcast episode, same segment, restarted each pass.

Loudness was matched by holding a cheap SPL meter at the ear canal entrance and adjusting each device until it read 68 ± 2 dBA. More on why that is the weakest part of this whole exercise in a minute.

The sealed-bud number is the one that rearranged my thinking. Six meters. On three of the twelve passes I did not hear the bike until it was inside three meters, which is to say I heard it as it went past. Those are the passes I think about.

What fifteen meters actually measured

It measured detection of one specific sound, by one specific person, in near-silence. That is a narrower claim than it sounds like.

An e-bike at 25 km/h on dry asphalt is roughly the quietest thing that can hurt you. It has no engine note. What you hear is tire hiss and a faint motor whine, and it arrives from behind, where your ears are least directional. It is close to a best case for demonstrating a difference between headphone types, because there is almost nothing to hear and any masking at all wipes it out.

So the honest framing of the result is: in the hardest listening condition, the open designs preserved roughly two-thirds to three-quarters of my unaided detection distance, and the sealed buds preserved about a fifth. Bone conduction did not make me hear better than my own ears — it cost me six meters against bare ears, which is a 22% haircut nobody advertises.

Against a car, the gap would compress. A sedan at 50 km/h announces itself far enough out that all four configurations would probably have caught it. The number matters most for the quiet threats: bikes, scooters, joggers overtaking, a dog off-leash, someone reversing out of a driveway.

What the number does not measure

This is the section I would want to read, so it is the longest one.

It does not measure reaction. Hearing a bike at 21 meters and doing something correct about it are different skills. I raised a hand in a controlled test where I knew a bike was coming and my entire job was to listen for it. On a real run I am thinking about a hill. Expectancy inflates detection substantially, and my test is saturated with it.

It does not measure wind. Four km/h is a still morning. On a bike at 30 km/h, the wind noise across the ear is the dominant sound in the entire environment and it flattens the differences between all four configurations toward zero. I tried a truncated version of this test on a bike and abandoned it because I could not detect the following rider reliably in any configuration, including bare ears. If you cycle, the awareness argument for open-ear is much weaker than the running argument. Nobody selling these will tell you that.

It does not measure transparency mode. I tested sealed buds with ANC off, not with a good transparency mode running. Modern transparency passthrough is genuinely competent and would have closed a lot of that fifteen meters. I did not have a fair second pair to test it with, and I am not going to guess at a number. If you already own something with strong passthrough, my result does not tell you what you would gain by switching. It tells you what you would gain over plugging your ears.

It does not measure a population. One listener, 43 years old, with unremarkable hearing that I have not had tested in a decade. One rider, one bike, one surface, one morning. The ranges in that table overlap between the two open configurations — 21 m and 18 m are not meaningfully different at n=12 with a spread that wide. Treat the open-versus-sealed gap as real and the open-versus-open gap as noise.

And the loudness matching is the soft joint in the whole structure. A $40 SPL meter held near the ear canal is not measuring what a bone conduction transducer delivers to your cochlea, because the transducer bypasses the ear canal entirely. Matching a device that vibrates your cheekbone against one that pressurizes a sealed cavity is, at the level of physics, matching two things that are not the same quantity. I did the best I could and then checked subjectively that all four sounded about equally loud to me. If you want to discount the fifteen meters by some amount for that, I would not argue.

What survives all of that: sealed earbuds with nothing running to counteract them cost me most of my hearing range for quiet approaching vehicles, and bone conduction cost me some of it. That is a directional finding, not a spec.

What bone conduction actually is, and what it isn't

The transducer sits on the zygomatic arch — the ridge of bone in front of your ear — and vibrates. That vibration travels through skull bone to the cochlea directly, skipping eardrum and middle ear. Your ear canal stays completely open, which is why the awareness advantage exists at all: there is no physical obstruction and no occlusion effect, so you never get the thumping of your own footfalls that sealed buds produce.

Here is the distinction that most coverage blurs, and it matters if you are shopping: open-ear is not a synonym for bone conduction. Shokz sells both. The OpenRun and OpenSwim lines are bone conduction — a band around the back of your head, pads on your cheekbones. The OpenFit and OpenDots lines are open-ear air conduction: tiny speakers hooked over your ear that fire sound into an open canal from a short distance. Different physics, different sound, different fit problems. The air-conduction ones sound better — actual bass exists — and leak more. The bone conduction ones survive being rained on and stay put when you shake your head hard.

If someone recommends "Shokz headphones" to you without specifying which family, they have not told you much.

What you give up

A review where everything is good is a brochure. Here is the ledger.

Bass is not attenuated; it is largely absent. Below about 100 Hz, the cheekbone stops delivering pitch and starts delivering sensation. A kick drum arrives as a small pressure pulse against your face rather than as a note. On the sub-heavy end of electronic music the fundamental simply isn't there — you get the harmonics above it and your brain fills in a ghost. The Pro 2's dual-driver arrangement, which adds a small air-conduction driver to the bone transducer, is a genuine improvement over the original Pro: on the same track, the Pro 2 gave me an audible low end where the Pro gave me a buzz. It is still not earbud bass. It is closer to good laptop speakers with a better midrange.

Speech, by contrast, is excellent. Podcasts and audiobooks are what these are actually for. Mid-heavy content lands cleanly and the open canal means you hear a voice in the room rather than a voice in your head.

Above about 80% volume, you feel it. The transducers tickle. It is not painful, but it is a distinct buzzing on the cheekbones that I find distracting, and it is the ceiling on how loud these get usefully. Running beside four lanes of morning traffic, I could not get the audio comfortably above the road. That is a real functional limit: in loud environments, bone conduction loses to sealed buds outright, and no amount of enthusiasm changes it.

Leakage is condition-dependent. At 60% in a quiet kitchen, my partner sitting a meter away heard nothing. At 80%, they could identify the podcast from across the table. On an open-plan office floor with ambient chatter, nobody has ever noticed. On a quiet train, I would not.

Glasses are a genuine compatibility issue. My frames have 4 mm temple arms and there is no conflict. A friend with chunky acetate frames borrowed mine and the arm lifted the transducer a few millimeters off the bone — the volume dropped noticeably and the bass, such as it is, vanished entirely. Bone conduction depends on contact pressure against bone. Anything between pad and cheek is a problem. Try before you commit if you wear thick frames.

Durability, from my own pair: the band on my first set developed a hairline crack at the left transducer joint at around sixteen months, running roughly five times a week. Shokz replaced it under the two-year warranty without argument. I mention it because it is the failure point I would expect on any of these — the flex hinge, not the electronics.

Three models, three jobs

I have owned or borrowed four. Weights are from my own kitchen scale; battery figures are the manufacturer's claim followed by what I got at ~60% volume with a mix of podcasts and music.

OpenRun Pro 2 is the one I reach for and the one I'd buy again. The dual-driver low end is the meaningful upgrade over the original Pro, and USB-C ended the proprietary-magnetic-cable era, which on its own is worth something to anyone who has been stranded at a hotel with the wrong charger. Ten and a bit hours real-world is enough that I charge it weekly, not daily.

OpenRun 2 is the sensible purchase for most people and I have talked more than one runner down to it. You lose the bass driver and a couple of hours. Check the ingress rating for your specific unit before you buy: the original OpenRun was IP67, and the Pro tier is IP55. That is not a typo, and it matters if your plan was to rinse them under a tap after a hot run. Mine live with a damp cloth instead.

OpenSwim Pro solves one problem completely. It has 32 GB of onboard storage and works as an MP3 player with no phone present, which also makes it the best of these for a run where you want to leave the phone at home. In the pool it works exactly as advertised, which still surprises me every time — water conducts, bone conducts, no earbud stays in anyway.

OpenComm2 is the outlier and the one I underestimated.

The calls, which I did not expect to care about

The headline mic finding: on a Teams call from a café with moderate background noise, wearing the OpenRun Pro 2, two people asked me to repeat myself in a thirty-minute meeting. Same café, same time of day, same meeting cadence, wearing the OpenComm2 with its boom mic: zero.

That is not a rigorous test, but it matches what recording voice memos on both showed. The OpenRun's mics are small and body-mounted at the jaw and they are competing with everything else in the room. The boom puts a noise-cancelling capsule two centimeters from the corner of your mouth. On a bike at any real speed, the OpenRun mic is unusable — the recording is wind and a suggestion of a person. Do not take calls while riding, for several reasons.

For an office worker whose actual daily requirement is six hours of headset with reliable outbound audio and the ability to hear a colleague walk up, the OpenComm2 is the correct product and the running models are a compromise. It is also the ugliest thing I own.

The 5 a.m. argument

My runs start at 5:40, which means a double espresso — about 130 mg — at 5:20 and out the door before it has fully landed. In the dark, on roads with no footpath, the awareness case stops being theoretical. Cars reverse out of driveways at that hour. Delivery riders run without lights.

There is a smaller thing that I did not anticipate and now would not give up: with an open canal you hear your own cadence. Footstrike, breathing, the change in both when a gradient starts. Sealed buds mask all of it, and the occlusion effect replaces your real footfall with a dull internal thud that tells you nothing. I pace better in bone conduction, and I noticed it in split consistency before I noticed it consciously.

I tried to sleep in them. Don't.

Given where this magazine lives, I owe you the test: I spent three nights trying to use bone conduction for sleep audio. It does not work. The band wraps the occipital bone, which is precisely the part of your skull a pillow is for. On my back it was tolerable for about twenty minutes. On my side it pressed the transducer into my cheek and the band into the pillow, and I gave up on night three. Battery also runs out before you do.

Bone conduction sleep headbands exist and are a different product category. These are not that, and no marketing claims they are — I just wanted to know.

What to actually pay

The Pro 2 has held a $179.95 sticker since launch and lands in the $130–$150 range during the usual sale windows a few times a year. The standard OpenRun goes below $100 in the same windows. The discounts are real but they are not rare, and there is no version of this where missing one is a problem.

My rule: I would pay sticker for the Pro 2 only if I ran daily or took calls on it. Otherwise the OpenRun at $99 does about 85% of the job, and the 15% is bass you will half-notice.

Who this is for

Buy bone conduction if: you run on roads or shared paths, especially before dawn or after dark; you mostly listen to speech; you want to hear a colleague approach your desk; you have never found an earbud that stays in when you sweat; you swim and want audio (OpenSwim Pro specifically).

Don't, if: you want these as your only headphones; you commute on a loud train or fly often, where you need to hear less, not more; you listen mostly to bass-forward music and care about it; you wear thick-armed glasses and can't try a pair first; you cycle fast and are buying on the safety argument alone — wind noise takes most of that back.

And if you're an office worker buying for calls: get the boom mic. The running models are a downgrade for the thing you'd actually use them for.

Fifteen meters is not an audio spec. At 5 a.m. on a dark road, it's the only one I care about.