As a college professor that teaches audio production (recording & editing), mainly for video but also for audio CD and digital distribution, the subject of “Sample Rate” inevitably raises its head. It’s an important subject, one that many find either boring or unnecessary.
Since you’re not in my class and we can’t devote the needed time to it, I came across this great article by Sweetwater Editorial Director, Mitch Gallagher. 7 Questions About Sample Rate. For those interested in recording and/or editing audio, this is a worthwhile read.
It’s easy to talk about the sample rates for sessions, but how much do you know about it? In this article, I’ll answer a few questions about sample rates.
What is “Sample Rate”?
Sample rate is literally how fast samples are taken. Picture an analog audio track. A “sample” is a measurement — a snapshot, if you will — at one specific time in that audio track, described in the binary language of 1s and 0s. Repeat that measurement tens of thousands of times each second; how often that snapshot is taken represents the sample rate or sampling frequency. It’s measured in “samples per second” and is usually expressed in kiloHertz (kHz), a unit meaning 1,000 times per second. Audio CDs, for example, have a sample rate of 44.1kHz, which means that the analog signal is sampled 44,100 times per second.
The science behind sample rates goes back to the 1940s, with the development of the Nyquist–Shannon theorem. The theorem states that when the sampling frequency is greater than twice the maximum frequency of the signal being sampled, the original signal can be faithfully reconstructed. As long as the Nyquist limit (the Nyquist limit is half the sample rate) exceeds the highest frequency of the signal being sampled, the original analog signal can be reconstructed without loss. If lower sampling rates are used, the original signal’s information may not be completely recoverable from the sampled signal or it may result in an audible artifact known as “aliasing.” (We’ll talk more about aliasing later in this article.)
If the sample rate is 44.1kHz, the highest frequency that can be captured and stored is a bit less than half of the sampling frequency, or around 22kHz. Remember that the accepted standard for the human hearing range is from 20 Hertz to 20,000 Hertz (or 20kHz), though in practice, most of us don’t hear frequencies that high. Age, exposure to loud sounds, and environmental issues lower sensitivity to high frequencies. Raising the sample rate to 48kHz raises the Nyquist frequency to just under 24kHz, and recording at 96kHz moves the Nyquist frequency to just under 48kHz — well over an octave beyond the range of audibility.
What Does Sample Rate Do?
You’ll hear people suggest that sample rate measures or captures a lot of things, but it really only does one thing: measure frequency. That’s it.
So how did we end up with 44.1kHz as the “standard” sample rate for so many digital formats? According to what I’ve been told, in the early days of digital audio, 48kHz was the “pro” standard, and manufacturers wanted to use a different rate for “consumer” devices to prevent direct digital copying. It’s not easy mathematically to convert digital audio from a sample rate of 48,000 to 44,100; so this sample rate was chosen for consumer gear for it’s incompatibility with pro gear.
If 44.1kHz Captures More than We can Hear, Why Use Higher Sample Rates?
There are a couple of reasons that higher sampling rates can be advantageous; the first is that while 44.1kHz is the standard for audio CDs, 48kHz is the standard for audio for video. Studios who regularly work in film and television may use 48kHz as their in-house standard. But higher sample rates such as 88.2kHz, 96kHz, 192kHz, and even higher may have a purpose — and maybe not the one that you think it is.
Remember what the Nyquist theorem states: frequencies below half of the sampling rate can be reconstructed. So what happens to frequencies that are more than half of the sampling rate? The theorem states that any frequencies above the Nyquist Limit will not be rendered properly, and this proved to be true; frequencies above the limit can appear as spurious signals in the audible audio spectrum. This is referred to as “aliasing,” and must be prevented by band-limiting (filtering) the analog audio before it’s converted to a digital format. Effectively, this means that analog-to-digital converters (ADCs) have a low-pass filter at the Nyquist Limit that stops those out-of-bandwidth-frequencies from getting to the converters. The implementation of that filter in the ADC is key; if done well, it should be completely transparent; done poorly, the filter will degrade the quality of the audio. By sampling at 88.2kHz, 96kHz, or even 192 kHz, the implementation of the anti-aliasing filter is moved above the audible frequency range (which means that even less than optimal filter design will be inaudible). This was a much bigger issue with early ADCs, where the filters could audibly degrade the signal. With modern technology, it’s much less of an issue regardless of sample rate. [read more]