Before we can understand digital audio, bit-digital we must first have a very general understanding of how sound occurs naturally. Sound is actually a mechanical wave that is propagating through the air (typically, but it can travel through other mediums as well). A basic visual illustration of sound is the sine wave.
A sine wave a represents a pure tone which is a single frequency. This sounds something like the emergency broadcast signal or the sound the TV makes when the colored bars appear, a long sustained beep. All sound is made up of these types of waves at different frequencies superimposed on each other ichimame (this forms the waveform you see on your DAW when you are editing).
Analog recording such as to a magnetic tape or a lathe (for vinyl records) used a microphone as a transducer (makes the same wave in the air, but with electricity), and stored the wave information either magnetically or with actual grooves as on a record.
Digital audio is recorded differently. One can think of digital audio being captured and recorded by taking very fast snapshots of the waveform (on the order of more than 44000 snapshots per second). This process is called sampling.
This is where digital audio really shines. Once the waveform or “data” is captured in this digital form, it becomes very malleable when using a computer and mathematical formulas. This is how we gain the power to do non linear editing and apply convolution algorithms.
Sample rate is measured in Hertz which is a unit per second (samples/seconds). There is a direct correlation between how accurately sound is recorded and the sample rate. That is, thewordcounter the higher the sampling rate, the more accurately the waveform is recorded. The downside of this is that often each discreet sample takes up at least sixteen bits of storage space. That means 44.1 kHz sample rate and 16 bit bit depth (CD quality) about 88.2 KB of space is need per channel.
Bit depth correlates to the number of different levels of dynamic range (volume) possible. So the higher the bit depth the higher the resolution of the dynamic range. Another way to look at is is this. The human ear can tolerate 0dB (silence) to 120dB (threshold of pain) of dynamic range. We gain roughly 6dB of dynamic range for every bit, therefore 16 bits will give us 96dB or dynamic range. (24 bits will yield 144dB of dynamic range)
In summary higher sample rates and higher bit depths yield higher fidelity recordings, only-and-one but they are also more taxing on processors, and take up more storage space on a hard drive. Remember that if you want to apply Digital Signal Processing to your mixes and you recorded your tracks at 88.2 kHz every plug-in you use will use roughly twice as much CPU time compared to tracks recorded at 44.1 kHz. 99% of people can no longer tell a difference beyond 48 kHz, 24 bit, and most are happy at CD quality 44.1 kHz, 16 bit (it will probably end up in this form anyway).