What is Aliasing? How to Avoid Aliasing in Your Mix

As I covered in my overview of sample rate, the sample rate you set with your recording hardware and in your DAW determines up to what frequency is able to be recorded and reproduced. With the Nyquist frequency in mind, in the digital realm this is half of what we set our sample rate to. So a sample rate of 48kHz records and reproduces frequencies up to 24kHz. But what about frequencies in our audio which exist above that 24kHz? We can’t hear them, so no harm, no foul, right? Unfortunately that’s not the case, which brings us to today’s topic: what is aliasing? Understanding this concept and how it relates to and affects your mix will help you achieve better sounding mixes, so let’s get into it.

What is Aliasing

I discussed quantization noise in my overview of what is bit depth. This manifests as white noise which results from recording or mixing at a lower bit depth than can accommodate the amplitudes of the sound.

In other words, this white noise is an error which is the result of a limitation.

Aliasing is essentially the same thing, but with frequencies instead of amplitudes.

If we have our sample rate set to 44,100Hz, according to the aforementioned Nyquist frequency, this means we can reproduce up to 22,050Hz in digital audio (half the sample rate).

If we have audio at frequencies above that limit of 22,050Hz in our mix, this produces an error which manifests as distortion in which those frequencies get “folded” back below 22,050Hz.

These unwanted artifacts are called aliasing.

What Causes Aliasing

You might be wondering why we’d even have such high frequency information in our mix in the first place considering 20k is the top end of what we can even hear.

High frequency information is often created without our realizing it with a lot of the plugins we use.

Harmonic related plugins like saturation or exciters as well as compression create overtones which can exceed the Nyquist frequency in our mix, thus causing the unwanted distortion associated with aliasing.

How to Prevent Aliasing

There are several ways to prevent aliasing, so as long as you’re conscious of it in your mix, it shouldn’t be a problem.

Oversampling

Oversampling is a setting (default or otherwise) on many modern plugins today with aliasing in mind.

When oversampling is enabled on a plugin in your mix (when applicable), it ignores the native sample rate you have set in your DAW and instead applies a much larger sample rate like 88.2k or 96k to the output to that particular plugin.

So let’s say for example you’re using an exciter in your signal chain (a common practice in the mastering process).

This plugin is designed to create overtones at higher frequencies to make the mix or track sound brighter and more pleasing to the ear. These overtones can easily exceed half of your sample rate, normally causing aliasing.

With oversampling enabled, that plugin is effectively raising the frequency ceiling so that those high frequencies which are created via the plugin are still able to be properly reproduced.

Oversampling uses more CPU, but it’s well worth it to prevent aliasing on the types of plugins which are prone to causing aliasing issues.

Low Pass Filtering

Analog to digital audio convertors have something called an anti-aliasing filter.

This is specifically designed to filter out frequencies above the Nyquist frequency so that the system doesn’t even attempt to reproduce them, hence preventing these errors.

You can use a linear low pass filter like FabFilter’s Pro-Q and set it to your Nyquist frequency after every problematic track in your mix.

So if your sample rate is 44,100Hz, you’d set your filter point at 22,050Hz. If your sample rate was 48,000Hz, you’d set your filter point at 24,000Hz.

This is useful for keeping older plugins in check which introduce aliasing issues but don’t have oversampling options.

Raising Your DAW’s Sample Rate

If you’ve got the CPU for it, there’s no substitute for raising your DAW’s sample rate either to 88.2k or 96k.

Like oversampling but for your entire mix, you’re ensuring that every frequency can be represented by drastically raising the frequency ceiling. You’re effectively doubling the ceiling so that aliasing is virtually impossible.

When it comes time to render down, an anti-aliasing filter can be used to remove everything above whatever sample rate you’re bouncing to.

Again, the trade off here is it’s taxing on your CPU and creates much larger files if you leave them at that frequency.

If nothing else, mixing at 88.2k or 96k gives you peace of mind that aliasing is a non-issue.