Pro Audio Files

Increase Your Income Improve Your Mixes

How to Use De-wind in iZotope RX 6

In this video, I’m going to cover how to use the brand new De-Wind module in RX 6 in a post-production context that features dialogue.

Once you’ve downloaded the test file called, “Post Production_De-Wind” for this tutorial from the RX web page and opened it in RX 6 by clicking and dragging it into the RX application window, or by opening RX and pressing Command+O or Control+O on a PC to locate and open the file.

Your RX window should look like mine with the file, Post Production De-Wind in a tab on the top-left hand side of the screen. Let’s first have a listen to this sample, which has a very noticeable wind noise coming into contact with the microphone’s diaphragm, distracting us from the dialog.

I’m going to press return to bring the playback head to the beginning of the sample, and space bar to play the sample.

[voice sample]

So as you can hear, that’s pretty distracting. Let’s use the De-Wind module to reduce these intermittent wind noises. I’ll locate it on the right. A little wind sock next to it. And you’ll notice a few parameters here. Let’s go through them together, and I’ll make some choices about how I want to process this wind noise.

Reduction determines the balance between the depth of wind reduction and the preservation of the original signal. I wasn’t hearing a lot of rumble here, so I’m going to leave it at its default value of 3. The crossover frequency sets the upper frequency limit for the De-Wind processing algorithm, so anything below the number I choose here will be processed.

To choose a frequency to dial in, you have to think about when the wind is most prevalent in the sample. We can tell by looking in the spectrogram and listening to the sample that the wind is most prevalent in these lower frequencies. To get a better look at the low end and increase the visual real estate, if you will, we’re going to right-click on the frequency scale and change the frequency scale from its current setting of Mel to extended log. Watch what happens when I do this.

Now, we get a much better look at the low end frequencies, where the wind is causing so much trouble.

So before, if I switch back from extended to Mel, it looks like this. Now, watch this 100Hz value here as I switch back from Mel to extended log. Watch what happens. So after we switch, we get a much better visual readout of our low frequency values like 5, 10, 20, 30, 50, 70Hz. Those are all visible now.

The next thing I’ll do is press F on my keyboard and switch to my frequency selection tool to find a good place to identify, in the spectrogram, where the wind is most pronounced.

So it looks like in between zero and 369 is where the wind is causing the most trouble. I got that value by making a selection and looking at this little box here, which gives us a readout of the frequency selection that we’ve made. The highest frequency value is 369.68, and the lowest is zero, and that makes sense, because I dragged my selection all the way to the bottom of the spectrogram.

Remember, I’m doing this to tell De-Wind’s crossover frequency parameter where to focus its processing to attenuate the wind.

Now, just to make sure I’m processing as much of the wind as possible, and not the important tonal noise from the dialogue, I can listen back to just the frequency selection that I’ve made by pressing this little play button here, which will only play what’s inside of our selection frequency wise.

So have a listen as I play this back.

[selection playback, filtered to 369Hz]

I’m doing this because if I notice that there’s a lot of dialogue in there in the selection I’ve made, I might want to reconsider where I want to move my crossover frequency cut-off, because I don’t want to affect the dialogue too much, just the wind.


So for example, if I went a bit higher with my selection and played it back…

[selection playback, filtered to 610Hz]

We can definitely hear some dialogue in there, so I have to be careful and make sure that I plug in a value of about 360, which is where we were just a moment ago.

I’m going to right-click and switch back from extended to Mel. Perfect.

Next, we have fundamental recovery, which recovers lower voice harmonics that might be lost or obscured by the wind. So essentially, if after we’ve attenuated wind noises, some voice harmonics were also attenuated in that process, this parameter can help to recreate the harmonics that might have been lost.

Lucky for us, this is a female voice over, so the intermittent wind noises weren’t treading on any low harmonics belonging to her voice. So I’ll leave this where it is at its default setting of 5.

Finally, we have artifact smoothing, and this eliminates the musical noise that is often characteristic of the processing that powers this module. Musical noise can be described as how something may sound underwater. We can increase the slider if the output sounds watery, or decrease it when too much smoothing makes things sound muffled.

Again, I’m going to leave it at 5, and maybe increase it if after I process these settings, I’m noticing either some wateriness or some muffled sort of voice over sounds.

So I’m going to press process. And as we can see in the spectrogram, those intermittent bursts of wind have definitely been attenuated. In fact, I can’t even see them anymore, so let’s listen to the De-Wind module processing on that voiceover.

Here’s after.

[voiceover after processing]

Now, just for context, let’s do the before. So my undo history window, I’ll go back to the initial state before we did processing. So here’s before.

[voiceover before processing]

And here’s after once more.

[voiceover after processing]

So those effects are pretty dramatic with De-Wind.

For more information and to download your own samples to use with RX 6, visit




Our friends at iZotope develop award-winning audio software and plugins for mixing, mastering, restoration and more. These are some of their great videos that we're currently featuring on The Pro Audio Files.

Free Video on Mixing Low End

Download a FREE 40-minute tutorial from Matthew Weiss on mixing low end.

Powered by ConvertKit
/> /> /> /> /> /> /> /> /> />