Is 44.1 kHz the right sample rate for my music? What about 48 kHz? Why doesn’t everyone use 192 kHz for everything if it houses the most data? These are questions that gentle folk embarking on an audio production journey invariably come across, and this article is designed to answer these questions and more!
Firstly, it should be known that 44.1 kHz and 48 kHz are widely recognized as the industry standard sample rates for recording and mixing. But why did we settle for, what are usually, the lowest options available on most DAWs? Surely technology and computation power has rendered these choices obsolete? Well, the answer to this lies in two places; the Sampling Theory, formerly known as the Nyquist Theory, and the fact that human hearing operates roughly over a range of 20 Hz to 20 kHz.
The Nyquist Theory states that for an audio frequency to be accurately recorded and reproduced over speakers or headphones the sample rate needs to be twice (2x) the audio frequency. So, a 4 kHz sine wave only needs to be sampled at 8 kHz, a 500 Hz wave only needs to be sample at 1 kHz.
Now considering that the highest frequencies that a person can hear are around 20 kHz, this hints at why 44.1 kHz and 48 kHz became and stayed the industry standard. Both of these sample rates accommodate a frequency response in excess of 20 kHz; 44.1 kHz can sample up to 22050 Hz, and 48 kHz can sample up to 24 kHz.
For the most part, that’s all there is to it! We make music and other media for human audiences, so we choose a digital sampling rate that is sufficient for all the frequencies that humans can respond to. Done!
You might be asking, but would 192 kHz sample a frequency ‘better’ than 48 kHz would? And the answer is no, not for frequencies below 24 kHz.
A few other things I’ve heard over the years and a few quick responses to them:
What about transients, don’t transients need higher frequency response than 20 kHz?
No, they don’t. Over the audible range, your ability to capture transients is not improved with higher sampling rates.
What about distortion and harmonics? Isn’t it better to record and mix at higher sample rates to accommodate for higher order harmonics?
On the recording side, not really. On the mixing side, there is an element of truth to this, which is some distortion plugins such as FabFilter Saturn offer ‘oversampling’ as a feature allowing you to work in 44.1 kHz but benefit from the bandwidth that a much higher sample rate would offer. That could be its own article in itself!
So, where do higher sample rates actually fit into the industry if they’re not used? Great question!
The main application to higher sample rates, in my opinion, is when utilizing extreme pitch shifting and time stretching. Slowing down a 192 kHz sound all the way down to a crawl will yield far less artifacts than doing the same thing with a 48 kHz sound. Similar results can be found for pitch shifting. This is why sound designers might record samples at a very high sample rate, so that they can be manipulated in postproduction.
In summary:
If you are recording music and you don’t expect to be doing extreme pitch shifting or time stretching, 44.1 kHz or 48 kHz is perfectly fine.
If you know that the project will go to CD production, probably choose 44.1 kHz (as that is the sample rate that CD audio must be in the end). If it’s for streaming, you can choose either, in my opinion.
The industry standard for all things video-related is 48 kHz, so if you’re composing for film, choose 48 kHz.
An interesting dilemma is the case of ‘this is going to CD, but we’re also making a music video’ … I would choose 44.1 kHz, even though there is a video thing involved; err on the side of the CD release on that one!
If you think you are going to do some wild pitch bending, world warping madness, choose a sample rate that is a multiple of either 44.1 or 48, based on what I’ve written above. For instance, if it’s destined for CD, choose 88.2 kHz or 176.4 kHz.
I haven’t mentioned bit depth yet, which is the other side of the digital coin. The choice with bit depth is simple. Record and mix at 24bit. If you need to deliver in a 16bit format, do the conversion from 24 to 16bit as the final, final step in the process.
And that’s it! Hopefully that answers all or most of the questions you may have had regarding sample rates in digital audio production!
