> ## Content Index > Fetch the complete content index at: https://soundforgepro.com/llms.txt > Use this file to discover other available public pages before exploring further. # How to Remove Wind Noise in iZotope RX 12: De-wind and Other Fixes - URL: https://soundforgepro.com/izotope-rx-remove-wind-noise/ - Published: 2026-10-25T12:43:00.000Z - Updated: 2026-10-25T12:42:59.000Z - Description: Sort the wind before you pick a module: De-wind for gusts in RX 12 Advanced, Spectral De-noise, Dialogue Isolate or Voice De-noise for a steady bed, Spectral Repair for leftover thumps. Published De-wind settings, edition limits, prices and when to re-record instead. - Author: Erick Finn - Tags: iZotope RX, Voice Recording **To remove wind noise in iZotope RX 12, use De-wind on gusts that thump the microphone, Spectral De-noise or Dialogue Isolate on wind that never lets up, and Spectral Repair on any thumps that survive.** De-wind is an RX 12 Advanced module that runs only inside the RX Audio Editor, so Standard and Elements owners work with the other tools. I choose the tool from the shape of the wind, not from the module list. A passing gust, a steady roar under the whole take and a gust that has already distorted the mic each need a different fix, and the wrong choice takes away more voice than wind. Below is the route for each case, the De-wind settings iZotope has published, and what reviewers and recordists report when the wind gets serious. The rest of the RX toolkit is mapped on the [iZotope RX hub](https://soundforgepro.com/izotope-rx/). ## Which RX tool removes wind noise De-wind is the only RX module built for wind, and it targets one kind of wind. The [RX 12 De-wind manual page](https://docs.izotope.com/rx12/en/de-wind.html?ref=soundforgepro.com) describes it as removing intermittent low-frequency rumble from light to moderate bursts hitting the microphone diaphragm. The same page sets two limits: the module is not designed for heavy bursts that blow out and distort the mic, and it keeps constant background wind on purpose, treating it as part of the noise floor. iZotope's own [RX 6 De-wind tutorial](https://www.youtube.com/watch?v=xxUNY6RVhIA&ref=soundforgepro.com) introduced it as a new RX 6 Advanced module; in RX 12 it is still Advanced-only. Everything else follows from those limits: | Wind problem | RX 12 tool | Editions | Where it runs | | --------------------------------------- | ------------------------------------- | ----------------------------------------------------------------- | --------------------------------------------------------------------- | | Gusts between or under words | De-wind | Advanced | RX Audio Editor only | | Steady wind bed | Spectral De-noise (Learn or Adaptive) | Standard, Advanced | Editor module and plug-in | | Steady or shifting wind behind dialogue | Dialogue Isolate | Standard, Advanced (Quality menu and band panel in Advanced only) | Editor module and plug-in | | Light steady wind, Elements owners | Voice De-noise | Elements, Standard, Advanced | Plug-in only in Elements; module and plug-in in Standard and Advanced | | Thumps left after a pass | Spectral Repair | Standard, Advanced | RX Audio Editor only | | Rumble below the voice | EQ high-pass | Standard, Advanced | RX Audio Editor only | | Gust that distorted the mic | Patch from clean audio or re-record | Any | Editing, not a single module | The edition and plug-in columns come from the [RX 12 overview tables](https://docs.izotope.com/rx12/en/rx-overview.html?ref=soundforgepro.com). De-wind appears only in the Advanced module column and has no row in the plug-in table, so it cannot sit on a DAW or NLE track as an insert or AudioSuite effect. Two order rules apply to every route. Run De-plosive before any high-pass: the [RX 12 De-plosive manual](https://docs.izotope.com/rx12/en/de-plosive.html?ref=soundforgepro.com) says it looks for plosives between 20 and 80 Hz and may miss them in audio that has already been high-passed. After that, a steep high-pass is a fair wind tool in its own right; the [RX 12 EQ manual](https://docs.izotope.com/rx12/en/eq.html?ref=soundforgepro.com) lists "steeply high passing out wind noise from a location recording" among the EQ module's jobs. Pops and wind thumps both live at the bottom of the spectrum, and the guide on [removing plosives without thinning the voice](https://soundforgepro.com/remove-plosives-from-audio/) covers the pop side. If the problem is hiss, hum or traffic rather than wind, the [RX background-noise workflow](https://soundforgepro.com/remove-background-noise-izotope-rx/) is the better starting point. This page stays on wind. ## Read the wind before you pick a module Open the clip in the RX Audio Editor and look at the bottom of the spectrogram. Wind rumble shows up low in the display, and its shape over time tells you which fix applies. - **Gusts.** Bursts of low-frequency energy that come and go, lined up with the thumps you hear between or under words. This is De-wind's job: the module tracks the noise floor over time and reacts to the sudden changes a gust causes on the diaphragm. - **A constant bed.** A band of low energy that runs through the whole take, under pauses and words alike. De-wind treats it as noise floor and leaves it, so this one goes to a de-noiser. - **Blown-out gusts.** A burst so strong that the voice inside it breaks up or crackles. The microphone itself distorted, which puts that moment outside De-wind's design. Where the wind sits depends on the mic, the windscreen and the gust. In one RX 11 example from [iZotope's 2024 guide to background noise in dialogue](https://www.izotope.com/community/blog/how-to-remove-background-noise-from-dialogue-recordings?ref=soundforgepro.com), the wind was heard mostly under 100 Hz, while crickets in the same clip sat around 3.2 kHz. That is one clip, not a rule. Read your own spectrogram before you set a crossover or a filter; the [audio frequency chart](https://soundforgepro.com/audio-frequency-chart/) shows which bands a voice needs. ![Decision chart with three lanes: intermittent gusts lead to De-wind in RX 12 Advanced, a constant wind bed leads to Spectral De-noise, Dialogue Isolate or Voice De-noise, and a blown-out gust leads to patching or re-recording, with dashed paths from De-wind and the steady-wind tools to Spectral Repair for leftover thumps](https://soundforgepro.com/content/images/2026/10/wind-type-decision.webp) Sort the wind first: gusts, a steady bed and a distorted burst each go to a different fix. Prevention is still cheaper than repair, and the De-plosive manual names wind filters on location as the standard protection for dialogue. If you edit in Sound Forge rather than RX, the guide to [reducing wind noise in a voice recording with Sound Forge](https://soundforgepro.com/reduce-wind-noise-voice-recording-sound-forge/) covers windscreens and the Sound Forge tools. ## De-wind in RX 12 Advanced, step by step Select only the gusts, start gently, set the crossover just above the rumble and render. If wind remains, run a second light pass rather than pushing every slider to its end. 1. **Bring the clip into the RX Audio Editor.** De-wind has no plug-in version, so a dialogue clip from a DAW or video editor has to be processed in the editor. If the editor is new to you, start with the [RX 12 beginner tutorial](https://soundforgepro.com/izotope-rx-12-tutorial-beginners/). 2. **Select the problem areas.** iZotope's [De-wind walkthrough from the RX 7 release](https://www.izotope.com/community/blog/remove-wind-noise-from-audio-with-de-wind-in-rx-7?ref=soundforgepro.com) applies the module across problem areas and credits it with keeping the dialogue intelligible. Leaving clean stretches unselected keeps them untouched. 3. **Open De-wind and click Preview.** Listen while you adjust; every control below is a trade. 4. **Set Reduction.** It balances the depth of wind reduction against preservation of the original signal. Start low and raise it until the thumps sit behind the voice. 5. **Set Crossover freq \[Hz\].** This is the upper frequency limit of the processing. Raise it only until the rumble is covered; every step higher lets the module work on more of the voice. 6. **Add Fundamental recovery if the voice thins out.** It re-synthesizes lower voice harmonics that the wind obscured or that were lost. 7. **Adjust Smoothing.** The manual calls it Artifact Smoothing, and the RX 12 panel labels it Smoothing. Raise it if the output sounds watery; lower it if speech turns muffled. 8. **Check against the original, then Render.** Bypass and Compare sit beside Preview and Render along the bottom of the module. ![RX 12 De-wind module panel with a vertical Reduction slider and horizontal Crossover freq, Fundamental recovery and Smoothing sliders, numbered markers 1 to 4 on those four handles, and Preview, Bypass, Compare and Render buttons along the bottom](https://soundforgepro.com/content/images/2026/10/de-wind-controls.webp) Set 1 and 2 first, then use 3 if the voice thins and 4 if the result turns watery. ### Published starting points iZotope has published three De-wind panels with readable values. None of them promises to fit your clip, but together they show a sensible range: | Published panel | Version | Reduction | Crossover | Fundamental recovery | Smoothing | | ----------------------------------------- | ------- | --------- | --------- | -------------------- | --------- | | RX 12 manual image | RX 12 | 3.0 | 1000 Hz | 5.0 | 5.0 | | "Remove Rumble" preset, RX 8 manual | RX 8 | 8.0 | 200 Hz | 5.0 | 5.0 | | Windy-field film scene, iZotope blog 2019 | RX 7 | 3.7 | 284 Hz | 9.3 | 5.0 | | Median of the three | \- | 3.7 | 284 Hz | 5.0 | 5.0 | The RX 12 manual does not call its image values defaults. The [RX 8 De-wind help page](https://s3.amazonaws.com/izotopedownloads/docs/rx8/en/de-wind/index.html?ref=soundforgepro.com) shows the Remove Rumble preset, which is not documented for RX 12, so treat it as a reference rather than a menu item. The 2019 values come from [Nick Messitte's dialogue-repair article on the iZotope blog](https://www.izotope.com/community/blog/6-dialogue-repair-problems?ref=soundforgepro.com), where a film scene shot in a windy field was treated in RX 7. I would start at the median row and move Crossover first, because it decides how much of the spectrum De-wind is allowed to touch. The 1000 Hz panel shows the module can reach well into the voice range, so I would only set it that high when the spectrogram shows the wind going that high too. When one pass is not enough, do not jump to extreme settings. The De-wind page warns that the more audio FFT-based processing removes, the more likely artifacts become, and it lists running De-wind more than once as an option for recordings with background wind. The 2024 iZotope dialogue guide makes the same point about denoising in general: a couple of gentle passes are more transparent than one extreme pass. ## When the wind never stops: Spectral De-noise, Dialogue Isolate, Voice De-noise Steady wind needs a de-noiser. De-wind keeps it as noise floor, and its own manual page points to Spectral De-noise instead. Dialogue Isolate and Voice De-noise are the other candidates; the choice comes down to your edition and how much the wind changes. ### Spectral De-noise: learn the wind (Standard and Advanced) The [RX 12 Spectral De-noise manual](https://docs.izotope.com/rx12/en/spectral-de-noise.html?ref=soundforgepro.com) names wind and outdoor environments among the noise it handles, and the tool runs as both an editor module and a plug-in. Learn captures a fixed noise profile, which suits wind that stays the same. In the editor, select the longest stretch that holds only wind, ideally a few seconds, and click Learn. In the plug-in, engage Learn and play the audio, or use Preview in AudioSuite on the current selection. Wind under speech often leaves no clean gap. The standalone editor can learn from several spectral selections at once: hold Shift and use the Lasso or Brush to grab wind only where the voice is absent in frequency, and RX offers to complete a partial profile. The plug-in cannot do this, which is one more reason to fix constant wind in the editor. For wind that keeps changing, Adaptive mode follows the noise instead of holding a fixed profile. The manual warns that it uses a significant amount of memory and CPU, and it points to Voice De-noise's Adaptive mode as the efficient, zero-latency alternative. ### Dialogue Isolate: separate the voice (Standard and Advanced) Dialogue Isolate separates speech from constant or non-stationary background noise, and the [RX 12 Dialogue Isolate manual](https://docs.izotope.com/rx12/en/dialogue-isolate.html?ref=soundforgepro.com) lists weather in that group. You set Voice, Reverb and Noise gain, and Sensitivity decides how much of the signal counts as noise: more Sensitivity brings more reduction, but also more risk of artifacts and lost clarity. Advanced adds a Quality menu (Good / Real-time or Best / Offline) and a four-band panel (Low, Low Mid, High Mid, High). That band panel is how the 2024 iZotope example aimed its processing at the low wind band. The [Dialogue Isolate guide](https://soundforgepro.com/izotope-rx-dialogue-isolate/) covers every control in depth. ### Voice De-noise: the Elements route Voice De-noise is in every RX 12 edition, and in Elements it is the wind tool you have, as a plug-in. The [RX 12 Voice De-noise manual](https://docs.izotope.com/rx12/en/voice-de-noise.html?ref=soundforgepro.com) describes a zero-latency design built from 64 bandpass filters that act like a multiband gate. Learn needs at least one second of pure noise. Adaptive mode follows changes in the noise floor and keeps its threshold lower to avoid artifacts. My pick for a light, steady wind bed on a DAW or NLE track is Voice De-noise in Adaptive mode, followed by the host's own high-pass if low rumble remains. ## Paint out leftover thumps with Spectral Repair Thumps that survive a De-wind or de-noise pass can be removed by hand. Select each one in the spectrogram and open [Spectral Repair](https://soundforgepro.com/izotope-rx-spectral-repair/), an editor-only module in Standard and Advanced. The right mode depends on what sits under the thump, according to the [RX 12 Spectral Repair manual](https://docs.izotope.com/rx12/en/spectral-repair.html?ref=soundforgepro.com): | Mode | What it does | Use it on wind when | Longest selection | | --------- | --------------------------------------------------------------------------------------- | -------------------------------------------------- | ---------------------------------------------------- | | Attenuate | Lowers magnitudes in the selection to match the surrounding area; resynthesizes nothing | The wind overlaps words and the voice must survive | Unlimited in Vertical mode; 10 s in Horizontal or 2D | | Replace | Replaces the selection with audio interpolated from the surrounding data | The thump falls in a pause or room tone | 10 s | | Pattern | Fills the selection from the most similar nearby audio | A short thump in a steady background | 4 s | The rule in the 2024 iZotope dialogue guide is simple: attenuate where the noise overlaps dialogue, replace where it falls in sections that hold only room tone. Recordists use the same move outside dialogue. In a [taperssection thread on wind in live recordings](https://taperssection.com/index.php?topic=202629.0&ref=soundforgepro.com), one user painted over the rumble at the bottom of the RX spectrogram with the brush and applied Spectral Repair Attenuate, reducing the wind substantially though not always completely. Nick Messitte's 2019 article also describes his earlier manual approach to a windy scene: Spectral Repair, De-noise and EQ, then library wind laid back in and blended to match. That is still a workable plan for a Standard owner. If you patch with real background from the same location, the guide to [recording and using room tone](https://soundforgepro.com/room-tone-sound-forge/) explains how to collect it. ## When a gust has distorted the microphone No RX module promises to undo a gust that overloaded the microphone, and the De-wind manual excludes that case outright. The fix is replacement, not reduction. 1. **Swap in a clean take.** If the line was recorded more than once, cut the damaged word from another take. 2. **Rebuild short gaps.** In a pause, Spectral Repair Replace (up to 10 seconds) or Pattern (up to 4 seconds) can fill the damaged stretch from the surrounding audio. 3. **Match the background.** Lay room tone under the patch so the edit does not drop out. Advanced also has Ambience Match, which the 2019 iZotope article paired with Spectral Repair on windy dialogue. 4. **Re-record.** When the damaged words carry meaning and no clean source exists, a retake is the honest fix. Field reports agree that this kind of wind seldom comes out clean. [Larry Jordan's 2020 review of CrumplePop plug-ins](https://larryjordan.com/articles/review-five-powerful-audio-plug-ins-from-crumplepop-for-mac-nles/?ref=soundforgepro.com) calls wind "a full-frequency mess" that is, most of the time, impossible to clean up. An answer on Reddit's field-recording forum gives the order I would follow: cut the distorted wind, replace it with an undistorted segment, then smooth the join with Spectral Repair. ## Wind on music and field recordings On music, De-wind and steep high-pass filters can take the bass with the wind. Fundamental recovery is described for voice harmonics, and the manual offers no equivalent for instruments. The taperssection thread is the clearest record. A live-concert recordist reported that De-wind removed the rumble but stripped the low end with it, behaving like a "smart high pass filter". Another contributor noted that a high-pass may fall short when the wind reaches higher frequencies, and that spectral editing is the most likely way to preserve low bass, at the cost of doing it by hand. A third relayed a Reddit tip: skip De-wind on music, highlight the wind in the spectrogram, Learn it in Spectral De-noise, and expect it to take a time or two. For music I would follow that order: Spectral De-noise learned on a wind-only selection at gentle settings, a second pass if needed, then Spectral Repair Attenuate on the bursts that remain. I would keep De-wind for spoken word. ## No RX Advanced? Standard and Elements routes Standard handles most wind without De-wind. Elements handles light, steady wind and little else, because it has no editor: the [RX 12 overview](https://docs.izotope.com/rx12/en/rx-overview.html?ref=soundforgepro.com) describes it as six plug-ins for your DAW (De-click, De-clip, De-hum, De-reverb, Repair Assistant and Voice De-noise), while the RX Audio Editor comes with Standard and Advanced. | Edition | US store price, October 21, 2026 | Wind tools | RX Audio Editor | | -------- | -------------------------------- | ------------------------------------------------------------------------------------------------------------------------------ | --------------- | | Elements | $99 | Voice De-noise plug-in, plus your host's own high-pass | No | | Standard | $399 | Spectral De-noise, Dialogue Isolate (without the Quality menu and band panel), Voice De-noise, Spectral Repair, EQ, De-plosive | Yes | | Advanced | $1,399 | Everything in Standard, plus De-wind, Dialogue Isolate's Best / Offline quality and four bands, Ambience Match and De-rustle | Yes | The prices are from iZotope's US store on October 21, 2026: [RX 12 Elements](https://www.izotope.com/products/rx-elements?ref=soundforgepro.com) at $99, [RX 12 Standard](https://www.izotope.com/products/rx-standard?ref=soundforgepro.com) at $399 and [RX 12 Advanced](https://www.izotope.com/products/rx-advanced?ref=soundforgepro.com) at $1,399\. The Advanced page lists the paths for current owners: $999 from RX 12 Standard, $799 from RX 11 Standard, $1,199 from RX Elements, and $269 to update from RX 11 Advanced. The same page lists De-rustle and Ambience Match as Advanced-only, the 64-bit plug-in formats (AU, AAX, AAX AudioSuite, VST3, AU ARA and VST3 ARA), and support for macOS Sonoma 14.7, Sequoia 15.8 and Tahoe 26.x on Intel and Apple silicon, plus Windows 10 22H2 and Windows 11 24H2\. None of those formats carries De-wind. Repair Assistant has no wind stage in any edition. The [RX 12 Repair Assistant manual](https://docs.izotope.com/rx12/en/repair-assistant.html?ref=soundforgepro.com) lists De-noise, De-reverb, De-hum and De-click as its Voice-mode clean-up stages. My advice: buy Standard if wind is an occasional problem, and pay for Advanced when gusty location dialogue is a regular job. The full split between [RX Elements, Standard and Advanced](https://soundforgepro.com/izotope-rx-elements-vs-standard-vs-advanced/) covers the non-wind differences, and the [RX price, trial and license guide](https://soundforgepro.com/izotope-rx-price-trial-license/) covers the ways to try and license RX. ## Test results: what published RX wind fixes show Published reviews offer no independent measurements of RX wind removal, so the composite result has two parts: the three De-wind panels iZotope itself published, and the verdicts of nine reviews and field reports from 2018 to 2025. **Published settings (three panels, one publisher).** Crossover runs from 200 to 1000 Hz, with a median of 284 Hz. Reduction runs from 3.0 to 8.0, median 3.7\. Fundamental recovery sits at 5.0 in two panels and 9.3 in the third. Smoothing is 5.0 in all three. Every panel comes from iZotope, in the RX 12 manual, the RX 8 manual and a 2019 blog article, so they are the vendor's starting points rather than independent measurements. **Verdicts (nine sources).** - De-wind cleared intermittent gusts from location dialogue in both vendor worked examples, 2 of 2: iZotope's 2018 De-wind post and the 2019 dialogue-repair article. - Wind, especially heavy or distorted wind, often does not come out completely, according to 4 of 9 sources: the taperssection thread, a Reddit field-recording answer, Larry Jordan's review and [Oliver Peters' 2023 CrumplePop write-up](https://digitalfilms.wordpress.com/2023/06/30/borisfx-adds-crumplepop-audio-suite/?ref=soundforgepro.com). - Manual spectral work is the fallback in 3 of 9, either painting and attenuating in Spectral Repair or cutting and patching. - Several gentle passes beat one heavy pass in 2 of 9, including iZotope's own 2024 guidance. - The one music-focused thread, 1 of 1, found that De-wind and high-pass filtering took bass along with the wind. - CrumplePop WindRemover handled wind behind speech well in 3 of 3 reports, and 2 of those 3 add that voice-isolation processing can leave speech flat or need a manual touch-up. My verdict: De-wind is worth the Advanced price when gusty location dialogue is regular work, because it handles the common case in one pass. For everyone else, Standard's Spectral De-noise plus Spectral Repair covers most wind, and a word the wind distorted still needs a patch or a retake, whichever edition you own. ## CrumplePop WindRemover vs RX for wind WindRemover is the quicker route when the clip lives on a video timeline; RX gives more control, plus a manual fallback when one-pass processing fails. [Boris FX describes WindRemover](https://borisfx.com/products/windremover/?ref=soundforgepro.com) as an AI plug-in that removes low- and high-frequency wind with real-time playback inside Premiere Pro, Media Composer, Pro Tools, Final Cut Pro, Audacity, Logic, GarageBand, Audition and DaVinci Resolve. De-wind, by contrast, means a trip to the RX Audio Editor. The reviews line up. [ProVideo Coalition's 2025 CrumplePop review](https://www.provideocoalition.com/review-crumplepop-2025/?ref=soundforgepro.com) found the default settings did "a pretty good job" on a windy clip in Resolve. Oliver Peters found that it largely eliminated wind behind someone speaking but did not fully clear a clip with wind and no speech. [RedShark's CrumplePop review](https://www.redsharknews.com/boris-fx-crumplepop-reviewed-does-it-make-amazing-audio-easy?ref=soundforgepro.com) noted how little flattening it caused, while Resolve's built-in voice isolation left the voice somewhat flat. The full [CrumplePop vs iZotope RX comparison](https://soundforgepro.com/crumplepop-vs-izotope-rx/) covers prices and the rest of both lineups, and the [guide to RX alternatives](https://soundforgepro.com/best-izotope-rx-alternatives/) lists other options.