What are sample rates and why should I pay attention to them?
Simply put, it is the rate at which an electrical current is sampled, per second. You should pay attention to them if you want to learn the language of studio engineers and give your clients the best possible end product. If you have always wondered about this mysterious world and what benefits some rates have compared to others, then this blog post is for you.
Lets Talk Basics
First, you must understand that the world around us is analog. What do I mean by that? Every sound that we hear is created by vibrating air waves at certain frequencies that our ears pick up and interpret as birds chirping, car horns honking or people talking. And at the basis of those sounds are sine waves that infinitely vary with regards to frequency (how often or wide) and amplitude (how high/low).
As humans, we have a limited range of hearing from about 20 hertz to 20,000 hertz. More commonly, this is abbreviated as 20k-20khz. Hertz is a nomenclature for number of cycles per second. So therefore, 1hz would be 1 pulse or cycle per second.
To give an example, have you ever seen a live orchestra play? Been to a musical performance or opera? Before the show starts, you will hear everyone tuning up and creating that sense of anticipation before the first note of the show. It is not random what they tune to, but is standard practice that everyone tunes to the note of A, which is 440hz (or cycles per second).
In the sense of a microphone, it physically picks up pressure differences in the air between two conductive layers of metal and changes that pressure differential into an electrical voltage. This tiny electrical charge gets passed down a signal chain through a preamp to magnify the voltage to line level for conversion to digital.
If the analog-to-digital converter (ADC) measures the incoming voltage at a rate of 44.1khz, it assigns 44,100 values to the relative height of the sine wave over a period of one second. If the rate is 48khz, then it assigns 48,000 values to the position of the sine wave over one second. Think of it as increasing the pixel resolution on your TV. The higher the sample rate, the more accurate your digital audio file will be.
According to the Nyquist-Shannon sampling theory, in order to accurately sample a piece of sound, the sound must be sampled more than twice the limit of hearing, so that gets us to 40khz and above.
But why 44.1khz?
For those who aren’t afraid of getting a little nerdy, follow this link to an excerpt of John Watkinson from his book “The Art of Digital Audio”.
In a nutshell, 44.1khz became the standard for high-definition audio due to the video industry having the best storage mediums and equipment to work with at the time. Therefore audio specifications conformed to those storage requirements.
What I Use
Most all sessions I record at 48khz now. I prefer it because it balances quality versus storage size and it is an even integer of bit rates/easier to convert. Why not go all out and record at 192khz? Bigger is better right? Not all the time.
First, it takes an enormous amount of CPU to edit sessions with 40+ tracks at 192khz.
Second, think about this. When you record audio for the first time, you have the highest resolution file and the most control over how you want things to sound. However, when a consumer listens to it, most devices only playback at rates of 44.1khz or 48khz.
If you give consumers a recording that is natively 96khz or 192khz, why leave it up to the consumer’s devices to down convert the sample rate with possible errors? I would rather err on the side of caution and remove any possible barriers to experiencing the audio the way it was meant to be.
Our job as engineers is to be the best quality control department possible, so that the fans have the same experience no matter where or how they listen to a song. Sample rates may be a small part of the equation, but nevertheless, a critical one.
Let me know your thoughts!
What do you use for converters or what rates do you sample at?