Bit depth and sample rate each relate to two specific aspects of your audio. Let’s compare bit depth and sample rate and specifically identify the best ones to use in your audio.
Bit Depth and Sample Rate Compared
Let’s compare the two by explaining how each one relates to your audio.
What is Bit Depth
We’ll begin by defining what is bit depth.
Bit depth refers to the amount of dynamic range possible between the point of clipping and the noise floor.
Different settings allow for different volume range potentials with higher bit depths unsurprisingly accounting for a greater range.
Why does this matter?
The more of a difference you have between the point of clipping and the noise floor essentially means that background noise will be much quieter and virtually unnoticeable.
More than that, bit depth also determines how accurately the amplitude, or measurement of the loudness of a sound wave, is able to be calculated from sample to sample.
When audio is captured, the value of its amplitude from one sample to the next is recorded. If you’re recording at too low a bit depth, the correct amplitude can’t be properly recorded. Instead the sample’s amplitude will be rounded to the nearest possible value which results in a quantization error, manifesting in some harsh and unpleasant noise.
You can set the bit depth both in your audio recording interface when initially recording audio as well as when you render a mix into a WAV file in your DAW.
How MUCH of a difference is there between each bit depth, though?
32 vs 24 vs 16 Bit Depth
The exact number of possible values available is determined through a simple equation: “2 to the power of (bit depth)”. The number of decibels it provides between clipping and the noise floor is a simpler equation: simply the bit depth multiplied by 6dB (6dB per bit).
As such, 8 bit depth is 2 to the 8th power, which yields 256 values. In terms of difference between the noise floor and clipping, this would be 48dB (8 times 6).
16 bit depth features 96dB of dynamic range (16 times 6), and in terms of amplitude values it has 65536 values (2 to the 16th power).
Because we’re dealing in exponents, a higher bit depth provides an exponentially (see what I did there) larger range of values.
So 24 bit depth, or 2 to the 24th power, yields a staggering 16777216 possible values! Of course its dynamic range is still just another 48dB higher than 16, so 144dB (24 times 6).
Lastly, 32 bit depth (integer) yield a redundant 4294967296 values along with 192dB of dynamic range. That said, as I mentioned in my bit depth comparison, a concert is typically 100dB, so 24 bit depth works just fine for recording the loudest practical musical audio in terms of worrying about the noise floor.
It’s worth mentioning that there’s also something called 32 bit float which, as I mentioned in that same comparison, is what DAWs process audio in and isn’t limited to fixed values and instead utilizes scientific notation to substantially open up the array of values and dynamic range that it is able to capture to a literal limitless practical range.
What is Sample Rate
If bit depth relates to dynamic range and volume potential of your audio, sample rate relates to the frequency range potential of your audio.
Sample rate refers to the number of samples or cycles which are capable of being captured per second via your audio hardware or software and it’s measured in hertz (Hz).
When I covered the parts of the sound wave, I explained that the frequency, measured in Hz, is the number of cycles a sound wave completes per second.
So essentially the sample rate refers to up to what frequency can be captured or reproduced in your audio hardware/DAW.
A typical sample rate for recording and mixing is 44,100Hz. Knowing that the human ear can only perceive up to roughly 20,000Hz, you might wonder why it necessary to be able to capture more than double that frequency?
There’s something known as the Nyquist–Shannon sampling theorem which dictates that digital audio is only able to reproduce half of the sample rate frequency. In other words, setting a sample rate of 44,100Hz means that we can go up to 22,050Hz which is still above the frequency of human hearing.
48,000Hz, another popular sample rate typically used in video production, covers up to 24,000Hz by the Nyquist-Shannon sampling theorem. Still MORE than enough to capture all perceptible frequencies. A relative few number of mixing engineers even use 96k… meaning 48,000 in practical frequency range to work with in the digital realm.
I talked about this as well in my overview of sample rates, but the long and short of it is that 96k is essentially unnecessary today thanks to protections which modern plugins use to ensure that there’s no issues when adding harmonics which stretch well above the 22k range via overtone distortion which they add (one of the major arguments for using a 96k sampling rate).
What Bit Depth and Sample Rate to Use
As I answered in my article on what sample rate and bit depth should I use, generally speaking, going with 24 bit depth and a 44,100 (44.1k) sample rate give you more than enough dynamic and frequency range, respectively.
While a relative few in the audio mixing community swear by extremes like 32 bit depth and a 96kHz sample rate, 99% of the time it’s unnecessary and simply creates a bloated file size needlessly.
There’s a reason 24 bit depth and 44.1k sample rate has been the standard for decades now, and even today if you’re submitting your music to a distribution service like Distrokid or Ditto Music, they’ll likely recommend 24 bit if not 16 bit depth and 44.1k sample rate.
Speaking of which, check out my overview on dithering audio whenever lowering the bit depth to avoid the noise associated with the aforementioned quantization errors.
Hi and thanks for this tutorial. Can one downsample music from, say, 48 k to 44 k? I started a project with a 48 k sample rate, but for final export, it’ll probably have to be 44 k.