Despite using synthesizers, I tend to make realistic sounds. That was one of the original uses of synths, to recreate sounds of other instruments, with varying degrees of success. Very few synthesizers set out to be, or are capable of being, used to synthesize sounds of nature, like birds, frogs, or… wind. The simplest and earliest commercial synthesizers were capable of rudimentary wind synthesis, of course: simply sweeping a resonant lowpass filter across white or pink noise was often enough to sell the concept. But, we can go much deeper than that, with some fantastic results. The techniques shown here are useful for much more than wind, including water, whooshes, and much more.

All samples are 320k mp3 at -6dB. Despite this, please turn down your headphones or monitors.

Filters

The lowpassed noise variety is nice, but lacks character. Using the other standard filter types alone tends to lack the required spectral presence, such as bandpass, highpass, and notch. However, we can combine any two of the four filter types to create a much more reasonable wind sound, such as using a highpass and lowpass in series:

However, this is still lackluster, and we’re well outside the realm of what most synthesizers can do already, which is why i tend more towards modular synths.

We can augment the above by using a formant filter (like the alt mode of MI Rings or the Kilohearts formant filter). This, when mixed with the previous filtered noise, can create very interesting results.

We can go a step further by replacing the above formant filter+HP/LP setup with a modal filterbank (like normal Rings with 0 Damping or Moodal), which allows for much more precise tuning of the noise – at the cost of potential phase issues. This works well if you use a dry/wet mix, possibly even along with any of the above filtering methods, but this is just raw modal filtered noise:

Noise Type

The type of noise you use matters as well. White noise is functional. It equally covers the entire spectrum, but… that’s not how humans hear things. We can use a tilt filter to create pink noise, which is much better suited to how we hear and is better to use for this. However, violet, blue, and brown noise can work just as well, and are attained similarly with a tilt filter. We can use a wide asymmetric notch to create grey noise, which also works. Below is a sample of white, pink, brown, violet, blue, and grey noise:

You can also use digital noise. The easiest form of this is to run any of the above noise into a comparator. Another way is to use an audio rate clocked sample and hold with noise as the sample input. You could also just use a dedicated digital noise generator like Befaco’s Noise Plethora, but those tend to be too detailed for wind. Here’s a couple examples of digital noise:

A less obvious choice is a detuneable wavetable oscillator. This is easily achieved in software like Serum, Vital, and Phaseplant, or you can use something like E352 in cloud mode or Quadnic in detune chord mode. These allow for more tonal and modulating noise with a ton of options:

Finally, you can also just use a reverb as a synthesizer. This works best with dedicated reverbs like Erbeverb, Supermassive, or Raum, but can work with multifx like RT-60 and Vybz; just crank up the feedback and play with the settings (this has WAY more uses than wind):

There are many other ways to create noise, of course; these are just the easiest and most accessible. I encourage you to explore and experiment – often the most seemingly unusable garbage is rather useful in the right context.

Modulation

Modulation, as with all sound design, is crucial. Modulation is how we add nuance to any given sound, and that’s no different for wind. So far all the examples have been modulated a tiny amount with smoothed random generators or manually by turning knobs, which works for ambient wind, but if you’re designing something more specific you will need to employ envelopes, LFOs, sequencers, and possibly other modulation.

For example, if you’re making a whoosh (which is effectively localised wind), you’ll likely use an AD envelope controlling overall amplitude as well as tonal characteristics, in this case the cutoff frequency of a lowpass filter:

Using LFOs to modulate wind is great if you need to design a cyclic wind pattern. I’ll be honest, I don’t do this much. Last time I did was when making a tornado, where I used 3 LFOs summed together to create an undulating control voltage to modulate the filters:

Sequencing wind is not very common, but still has uses. The one that comes to mind first is if you’re designing wind throughout a scene that switches viewpoints. Where you are in a room when there’s a storm outside will change your perception of that storm. Normally I would do this by using separate wind samples and stringing them together per shot, but you can sequence it as well. I don’t have a great example of this as it kinda requires video and I don’t feel like editing one together XD

Another kind of modulation, as noted above, is random modulation. It’s useful for almost all things in sound design, but for wind specifically, using noise as a modulator is extremely powerful. In the wavetable noise example, I modulated detune amount with bandpassed white noise to create that very noisy clouding sound. You can also modulate filters with the same noise they’re processing, with interesting results.

Conclusion

I hope this article has given you some fun ideas to try, and made you realise just how much can go into a good wind sound. If you have any questions, feel free to ask, and if you like this article, why not buy me a tea?