Remove wind noise from video: a real-world test

Last updated: November 5, 2026

Choose a file

or drop audio or video here

MP3, WAV, M4A, FLAC, OGG, MP4, MOV, WebM · up to 60 minutes

Upload your MP4 or MOV above and play the free 30-second preview. You get the same video back with a cleaner soundtrack, and the picture is not re-encoded. In our tests, Standard removed more wind when gusts hit the microphone, and both modes handled wind in the background. Standard runs in your browser and is free for files up to 5 minutes, so start there.

Hear it on a real windy interview

This interview was filmed outdoors the day after a typhoon, with wind hitting the camera mic while the speaker talks. Press play and switch between the original and the two modes. Only the sound changes; the video keeps playing.

Standard removed most of the wind, including the gusts under the words. Enhanced removed less, and gusts were still audible while the speaker was talking. In the quietest moments of the clip, the wind fell from −21.8 dB to −69.3 dB with Standard and to −56.6 dB with Enhanced.

Sample: a U.S. Marine Corps interview recorded outdoors after Typhoon Khanun, with wind hitting the mic (DVIDS #893006, public domain).

Which kind of wind is in your video?

Wind makes two different sounds on a recording, and the two modes handled them differently in our test. Listen to a moment where nobody is talking and match it below.

Two kinds of wind noise and which mode to start with, based on our test
What you hear	What is happening	Start with
Low rumble, thumps and crackle that cover words	Wind is hitting the microphone itself	Standard
A steady whoosh or rustling behind clear speech	Wind is moving around you, but the mic is sheltered	Either; Enhanced was slightly quieter

Wind hitting the mic: Standard removed more

When wind blows straight onto a microphone, it makes loud, low bursts that change from one second to the next. That is the sound in the interview above, and Standard handled it better there.

We saw the same thing in a mixed sample: a narrator over a recording of wind blowing straight onto a microphone. The loudest gusts left in the pauses were −39.0 dB with Standard and −22.1 dB with Enhanced, down from −5.5 dB in the original. Enhanced left the strongest gusts about 17 dB louder.

Wind in the background: both modes work

Wind that moves around you rather than into the mic is softer and steadier. Both modes removed it. In a second real interview from the same day, with lighter wind, the quietest moments fell from −38.1 dB to −77.6 dB with Standard and −75.4 dB with Enhanced, and we could not hear a difference between them.

In two mixed samples, a light breeze and strong wind through palm trees, Enhanced left the pauses quieter: −81.5 dB against −68.1 dB for the breeze, and −74.2 dB against −68.9 dB for the strong wind. Standard took a little brightness out of the voice in the breeze sample (2.6 dB less energy above 4 kHz than the clean voice).

If your clip sounds muffled after Standard, preview Enhanced on the same file. In our earlier noise test, where the wind was almost as loud as the voice, Standard made the voice sound duller. Enhanced uses 2 credits a minute, so compare before you clean the whole video.

Results on the real interviews

Both clips are 30 seconds from interviews recorded outdoors after Typhoon Khanun, published by the U.S. Department of Defense on DVIDS. Their footage is in the public domain, and its use here does not imply any endorsement. A real recording has no clean copy of the voice to compare with, so we measured the quietest 20% of 10-millisecond slices, where wind is most of what you hear, and the typical level while someone is speaking. We also listened to every version.

The speech level fell by about 4 dB in the windier clip with both modes. Most likely, part of what was loud in the original was wind under the words.

Measurements on two 30-second clips, tested 5 November 2026 (dB; lower is better for the quietest moments)
Clip	Version	Quietest moments	Typical speech level
Strong wind on the mic	Original	−21.8	−17.1
Strong wind on the mic	Standard	−69.3	−20.9
Strong wind on the mic	Enhanced	−56.6	−20.5
Lighter wind	Original	−38.1	−15.4
Lighter wind	Standard	−77.6	−16.2
Lighter wind	Enhanced	−75.4	−15.9

Results on the mixed samples

Each sample is 18 seconds of a narrator from a public-domain LibriVox recording, mixed with a real wind recording released under CC0 on Freesound. The wind was set 10 LU quieter than the voice for the breeze, 2 LU quieter for the strong wind, and 2 LU louder for the gusts on the mic. Both modes got the same file, and the outputs are time-aligned with it. We measured the pauses between words (10-millisecond slices where the clean voice is at least 40 dB below its loudest point) and, in the louder 30% of the speech, how much energy sits above 4 kHz, where consonants live. See how we test for the rules we follow.

Measurements on each 18-second sample, tested 5 November 2026 (dB; lower is better for noise, closer to the clean voice is better for energy above 4 kHz)
Wind	Version	Noise in pauses	Loudest noise in pauses	Voice energy above 4 kHz
Light breeze	Clean voice	—	—	−32.2
Light breeze	Original	−33.1	−14.6	−17.7
Light breeze	Standard	−68.1	−45.6	−34.8
Light breeze	Enhanced	−81.5	−50.4	−31.0
Strong wind	Clean voice	—	—	−21.7
Strong wind	Original	−22.6	−3.1	−21.8
Strong wind	Standard	−68.9	−29.2	−21.6
Strong wind	Enhanced	−74.2	−31.4	−18.1
Gusts on the mic	Clean voice	—	—	−22.2
Gusts on the mic	Original	−17.7	−5.5	−23.8
Gusts on the mic	Standard	−67.3	−39.0	−20.8
Gusts on the mic	Enhanced	−64.7	−22.1	−18.9

Stop wind noise before you shoot

A windscreen does more than any software. A foam cover helps in a light breeze; for real wind, use a furry cover over the mic. A clip-on mic hidden under a jacket or shirt is also well protected.

Turn so your body or a wall blocks the wind from the mic, and stand on the sheltered side of a building or car. On a phone, find where the mic is (usually at the bottom edge) and keep it out of the wind, without covering it with a finger, which muffles the voice. Some phone cameras also have a wind noise setting worth turning on before you film outside.

What this test doesn’t show

This is a small test. The two real clips come from the same day and place, with one speaker each, and have no clean copy of the voice, so their numbers only show how much wind is left in quiet moments. The three mixed samples use one narrator, and the Freesound files are compressed previews. We used Standard at its strongest setting. We did not test wind on a clip-on mic, wind with music or several speakers, or gusts loud enough to distort the recording. That distortion is in the sound itself, and noise removal can’t undo it.

Hear the difference in 30 seconds