Ever since the release of compact disc in 1982, we’ve had 44.1kHz. Most of the music which has ever been released in a digital format has used this sample rate; millions of albums, both on CD and more recently as MP3 and AAC downloads. 44.1kHz has been with us for a long time but the question is, do we still need it in the modern world and why was it chosen as a sample rate in the first place?
During the development of CD, one of the requirements was that the format must be able to reproduce the entire audible frequency spectrum. This is generally quoted as being roughly 20Hz to 20kHz for human hearing. It was known from Nyquist’s Theorem that in order to reproduce any given audio frequency, the sample rate had to be at least double the highest frequency you wanted to reproduce. This meant that CD had to sample at a rate of 40kHz or higher in order to cover the audible frequency spectrum. In the early days of digital audio, storing the equivalent of a CD album worth of digital data on a hard drive wasn’t possible because the drives of the time simply didn’t have sufficient capacity. Video recorders were therefore re-purposed to store audio samples as black and white video signals. In the US, these video recorders ran at 30 frames per second and had 490 useable lines per frame (excluding blanking lines). It was decided that 3 audio samples should be stored per line of video.
3 samples per line X 490 lines X 30 frames per second. This gives us a total of 44,100.
In the UK, video recorders operated at a slightly different resolution and frame rate. Once again though, if you store 3 audio samples per line of video, the maths still works out:
3 samples per line X 588 lines X 25 frames per second. Again, we get 44,100.
As you can only store a whole number of samples per line (1,2,3,4,5 etc), 44.1kHz was the minimum sample rate possible in order to fulfil the nyquist requirement and also to allow for CD masters to be stored on video tapes.
The Two Families Of Sample Rates
By contrast, TV production and modern digital video workflows have always used 48kHz as their standard sample rate. Multiples of both 44.1kHz and 48kHz are also now fairly widely used. In certain workflows, higher sample rates can be of use. The two common sets of sample rates today are:
CD derived rates:
Video derived rates:
Why Do We Need So Many Sample Rates?
It’s evident that 44.1kHz was borne purely out of the technical constraints at the time of the development of CD. Its derivatives, 88.2 and 176.4kHz only exist because they’re mathematical multiples of it. As we move away from physical formats and into an era where most content is delivered electronically, is there really any reason to keep 44.1kHz? Obviously, we’ll need to retain the ability to work with legacy content at that sample rate and to be able to produce CD masters for as long as the format still exists, but shouldn’t we now just move to 48, 96 and (where required) 192kHz? I’m very interested to hear your viewpoints on this.