Digital Music Jargon Explained – Aliasing

The fourth installment of this series covers the topic of aliasing, and the Nyquist Theorem. This follows on from the previous posts which covered FFT, oversampling and dithering…

What Is Aliasing?

Aliasing occurs where there is not enough information present to accurately reproduce the source material. The most common example of this can be seen on TV when you look at the wheels of a car or other vehicle which has spoke-like hubcaps – at a certain point, the wheels seem to start rotating backwards. This is because the frame rate of the footage isn’t high enough to provide an accurate depiction of the wheel’s actual rotation.

In terms of audio, the same principle applies. For a digital system, if the sampling rate is not high enough to capture a particular waveform, then that waveform may be indistinguishable from another waveform at a lower frequency.

Back To The Nyquist Theorem

The Nyquist Theorem states that the sampling rate required to reproduce a particular signal must be at least twice the highest frequency in that signal. In theory, you can describe an oscillation with only two plot points. At a given sampling rate, lower frequencies will have a lot of plot points per cycle, because the waves take so much longer to develop.

However, as the pitch increases, the number of plot points per cycle decreases. In a 48kHz system, the highest frequency it can reproduce is 24kHz. If you go above this frequency, then a 26kHz tone will be identical to a 22kHz tone, a 30kHz tone will be identical to an 18kHz tone, and so on.

This sort of aliasing (caused by low sample rates) is a form of distortion that is undesirable in most situations. However, the chiptune genre thrives on pushing sounds through bitcrushers and lo-fi distortions to generate 8-bit bleeps…