Let me start with something it took me far too long to accept: wind noise is not a "just pull it down with a noise reducer in post" problem. That's exactly how I used to think, and the result was a whole exhaust note smeared under a low, droning whoosh. I'd throw it into a noise-reduction plugin and crank it, and the thickness of the exhaust got shaved off right along with the wind — muddy, fake, worse than if I'd left it alone. Eventually I gave up on that idea: the wind has to be blocked the instant you press record, and post is only there to tidy the edges. So everything in this post is about pushing wind noise down to the minimum at the moment of recording.
What wind noise actually is, and why it's so hard to fix
What we casually call "wind noise" is really two different things, and telling them apart matters. One is the ambient sound of wind blowing through distant trees or buildings — that's an actual "sound," and a mic captures it just fine. The other, and my biggest enemy when riding fast, is airflow hitting the mic diaphragm directly — that isn't sound at all, it's a low-frequency overload, the diaphragm getting physically shoved by air pressure.
That distinction is exactly why it's so hard to fix. Ordinary ambient noise is spread across the spectrum, so noise-reduction software can still pick out its signature and separate it. But airflow slamming the diaphragm produces one giant blob of low-frequency energy, almost all of it jammed below a couple hundred hertz — thick, flat, and able to clip the signal in an instant. The most infuriating part is that this low band overlaps perfectly with the exhaust low end I want to keep, so the moment you cut it in post, the thickness of the note goes with it.
There's another variable: speed. The pressure of airflow on the diaphragm climbs steeply with road speed — not linearly, it explodes as the speed goes up. So puttering around town is fine, but the second I hit the highway and the speedometer reaches 80 or 90, the wind pressure blows past what the diaphragm can take. I've had this happen too many times — "the test recording in the alley sounded fine, then the whole highway section was trash" — and this is exactly why.
Once that clicked, the logic of my four-layer method got simple: the first three layers all try to keep wind from touching the diaphragm, and if it does touch, to make the impact as soft as possible; the last layer makes sure that even the bit that slips through doesn't get amplified into clipping. Let me break it down layer by layer.
My four-layer method for reducing wind noise
My workflow is always the same four layers — outside in, hardware to settings — each one blocking a slice, stacked together until it's enough. No single layer does the job alone: mount a windscreen but aim it into the wind, and it still gets eaten; nail the placement but crank the gain to clipping, and the throttle blip still distorts. These four layers complement each other, and skipping one leaves a gap.
The order goes like this: (1) a deadcat windscreen, breaking up the airflow at the outermost edge; (2) downwind placement, tucking the recorder somewhere the wind barely reaches, with the mic facing the tail and going with the wind; (3) narrowing the pickup on a Zoom H2n, framing the pickup angle on the exhaust behind me; (4) low mic gain, leaving enough headroom for the exhaust's high volume so the peaks don't clip. Let me walk through what I'm thinking at each layer.
Layer 1: A deadcat windscreen, breaking up airflow outside the diaphragm
My first line of defense is the deadcat windscreen — also called a deadcat or windjammer, that furry sleeve that slips over the outside of the mic. A lot of people assume the foam plug that came with the mic is enough; I thought so too at first, but the moment I ran it at highway speed I realized the two aren't even in the same league. A foam windscreen is fine for a faint indoor breeze, but it's hopeless against the wind pressure a motorcycle throws at it.
A deadcat works because of its structure. Those long strands of fur, before the airflow ever reaches the mic, break a concentrated gust apart into countless tiny eddies — turning one heavy punch into a scatter of small forces, so by the time the air touches the diaphragm it's far gentler, all without blocking the actual sound much. That's exactly why location sound people almost always run a furry cover and never rely on foam alone.
The numbers: a foam windscreen knocks down roughly 10 to 15 dB of wind noise, while a deadcat can hit 20 to 30 dB or more, depending on the model and wind speed. After I switched from foam to fur, that low-frequency roar in the high-speed sections dropped a full notch. It won't make wind noise vanish completely, but it pulls the situation from "the whole section is ruined" back to "post can still save this" — and it's the cheapest step in my entire method with the most direct payoff, which is why I always start here.
Deadcat Windscreen (Deadcat / Windscreen)
A long-fur windscreen that slips over the mic, breaking airflow into tiny eddies and drastically softening the impact on the diaphragm. For high-speed exhaust recording you need a deadcat — foam alone won't cut it.
Pros
- Long fur scatters concentrated airflow — the most noticeable drop in wind noise
- A full level above a foam windscreen for wind blocking
- Cheap, easy to fit, instant results
- Barely affects mids and highs, so exhaust detail survives
Cons
- Size has to match your mic's diameter — measure before buying
- Long fur traps dust, so it needs the occasional clean
When buying a deadcat I only watch one thing: the size has to be right. Too loose, and it flaps around at speed, generating its own rattling noise; too tight, and it simply won't go on. Measure the diameter of your mic or recorder's mic head before buying, and pick a model that hugs it snugly. I use a one-piece recorder like the Zoom H2n, and there are deadcats shaped specifically for it on the market — they wrap tighter than a universal cover and stay put better at speed, so that's what I look for first.
Layer 2: Downwind placement, mic facing the tail and going with the wind
The second layer costs nothing yet gets overlooked the most: placement. My thinking is simple — rather than fighting to block the wind, put the recorder somewhere the wind barely reaches in the first place. While you're riding, the bodywork, the fairing, the tank, even my whole body trails a relatively calm "wake" behind it, where the airflow is turbulent and slow rather than blasting in head-on. Tuck the recorder into that downwind pocket, and wind noise drops by a large chunk right at the source.
The second key is which way the mic points. I aim it toward the tail, in the same direction as the airflow, never into the wind. Think of sticking your hand out a car window: palm flat into the wind and it gets shoved hard backward; turn it edge-on, knife-style with the wind, and the resistance drops way down. Same with a mic — pointing it back lets the air "graze past" instead of "pouring in," and the force hitting the diaphragm head-on instantly drops a level. And since what I'm recording is the exhaust outlet behind me anyway, facing the tail handles wind resistance and pickup at the same time.
The exact spot is different on every bike, and I've found each one by trial and error. On my R6 the position I use most is back near the exhaust outlet, fixed with a zip tie or quick-release, with the front kept off the direct line of the wind as much as possible. The mount has to be rock solid — at speed, the slightest bit of play adds a pile of rattling noise inside the deadcat. In my experience, spending ten extra minutes on placement often beats an hour of noise reduction in post.
Layer 3: Narrowing the pickup with a Zoom H2n
The third layer relies on a tool that can switch pickup patterns, and mine is the Zoom H2n. It has five mics, top and bottom, and switches between several modes: the front X/Y pair crosses at 90 degrees for a narrower, more focused pickup; MS (mid-side) mode lets you adjust the side width, anywhere from very narrow up to 180 degrees; above that there are 2-channel and 4-channel surround modes. For wind noise, I switch to one of the narrower, more directional patterns so it concentrates on the exhaust behind me, leaving sound from other directions — including the turbulent wind noise pouring in from the sides and head-on — outside the pickup range.
The principle behind it is "directionality." Surround or wide-angle modes are sensitive in every direction, so wind gets picked up no matter where it comes from; a narrow, directional mode focuses on one direction and is naturally less sensitive to "wind that isn't in that direction." By narrowing the pickup and aiming it at the exhaust, I'm using the pickup angle to filter out one more layer of side wind noise — and as a bonus, the exhaust takes up a bigger share of the whole recording.
Being able to switch the pickup angle on the spot is exactly why I picked the H2n over just any audio recorder. I can choose, on the fly, whichever mode best dodges the wind and best aims at the exhaust for the current wind direction and placement — instead of getting home after recording and discovering I picked up too wide and let all the noise in.
One thing to watch, though: the narrower the directionality, the more sensitive it is to "did you actually aim it right." Miss the angle, fail to hit the exhaust outlet, and a narrow pattern records something hollow instead. So Layer 2's placement and Layer 3's pickup angle are really tied together — I put the recorder in the wake first, facing the tail, then use the pickup mode to "frame" that direction. Only when the two layers work together is it dialed in.
While I'm on pickup angles, here's a prerequisite I always insist on: for wind noise, I'd rather carry an extra Zoom H2n than use the sound straight off the DJI Osmo Action 4's built-in mic. The camera's onboard wind handling is tuned for "make speech clear," so it actively cuts the low end and scrubs the wind out — and for the exhaust note I want, that's exactly the wrong move. The low band it cuts is precisely where the thickness of the note lives; the wind gets scrubbed, but so does the thing I'm after. The camera's wind processing is in direct conflict with the logic of my first three layers: I want to block the wind at the source while keeping the exhaust low end fully intact, whereas it just chops out the low end after the fact, wind and meat together. That's why I use an H2n — one that lets me set the pickup angle myself and hands all of the low end back to me to process — so the work of the first three layers doesn't get flattened by the camera body for nothing. The gain setting in the next layer is also about how I set it on the H2n.
Layer 4: Low mic gain, leaving headroom for the note
The last layer is on the settings side: set the mic gain (input gain) low. Intuition says crank the gain so the sound is clear, but for the exhaust note it's the exact opposite. Exhaust SPL is already high, and the peak of a throttle blip is faster and harder still. Crank the gain too high and that peak slams straight into the recording ceiling (0 dB) and clips — and once it clips, that section is dead, no amount of post can bring it back.
There's a second downside to high gain: it amplifies the leftover wind noise along with everything else. The wind noise I worked so hard to push down in the first three layers gets shoved right back up because the gain is too high. So setting the gain low handles two things at once — it neither clips the note nor amplifies the wind.
My setting rule: keep the loudest throttle-blip peaks landing somewhere between -12 and -6 dB, leaving plenty of headroom, never touching the ceiling. I'd rather record the whole thing a touch quiet and bring the level up in post than crank the gain and gamble on clipping just so it "sounds satisfying live." In digital recording, a small signal that didn't clip can still be saved later; a clipped one is gone for good.
Stacked together, the four layers work like this: the deadcat breaks up the airflow at the outermost edge, downwind placement keeps the wind off the mic, the narrow pickup angle keeps side wind out, and low gain makes sure that even the leftover doesn't get amplified or clip. Each layer blocks a slice, and what's left for post is just a little, easily handled bit of residue.
How much post can fix, and what it can't bring back
Even with all four layers fully in place, the high-speed sections inevitably leave a bit of low-frequency wind, and that's when post finally steps in — but its role is to finish, not to do the heavy lifting. I use just two moves, both low-risk. The first is a high-pass filter to cut the low band that the ear can barely hear yet is almost entirely wind energy. Wind noise energy mostly piles up below 100 Hz, and my compromise for the exhaust is to start cutting downward from somewhere around 80 to 100 Hz — that range barely affects how the exhaust sounds, but it removes a lot of that rumbling floor noise. The second is a very light touch of noise reduction, just a hair.
The key word: restraint. The moment you get heavy-handed with noise reduction, the thickness and texture of the exhaust get shaved off with it, turning into that muffled, behind-a-blanket fake sound — the exact opposite of what I want, since my whole thing is a real, clean exhaust note with barely any audible post-production. Given the choice, I'd rather leave a touch of natural wind than squeeze out a dead, noise-reduction-flattened sound.
The conclusion fits in one line: all post can ever bring back is the part that wasn't recorded badly in the first place. A clipped peak can't be recovered, and a low end smeared by wind across a whole section is very hard to clean out. So I put all my effort into these four layers at the moment of recording, which lets post stay both easy and clean. If you want to hear what these methods actually sound like recorded, start with the latest videos on the home page.
FAQ
Why is wind noise so hard to kill when recording motorcycle exhaust at speed?
Because it isn't normal ambient sound — it's a low-frequency overload caused by airflow slamming straight into the mic diaphragm. Almost all of the energy is jammed into the low end, thick and flat below a couple hundred hertz, and it can clip the signal instantly. Worse, it overlaps with the exhaust low end I'm trying to keep. The lesson: you have to block the wind the moment you hit record. Post can only clean up the edges — anything that already clipped is gone for good.
Does a deadcat windscreen actually work? How is it different from a foam windscreen?
Hugely different. A foam windscreen only knocks down about 10 to 15 dB of wind noise, while a deadcat uses long fur to break the airflow into tiny turbulence, blocking 20 to 30 dB or more, so far less force hits the diaphragm. The moment I switched, that low-frequency roar in the high-speed sections dropped a whole notch. It's the cheapest, most noticeable step in the whole method.
How should I position the recorder so the wind doesn't eat it?
My rule is to tuck it into the wake behind the bodywork or my own body, with the mic facing the tail, going with the airflow rather than into it. Then I switch the Zoom H2n to a narrower pickup mode and frame the pickup on the exhaust behind me. That cuts both the head-on impact and the side wind noise.
How low should I set the mic gain?
I like to keep the loudest throttle-blip peaks landing around -12 to -6 dB without touching the ceiling. Exhaust SPL is already high, so you don't need much gain — crank it and you just amplify the wind noise and clip more easily. Better to record a little quiet and bring it up in post than to record it blown out.