March 2009

So what are bit depth and sample rate?

I decided to write this after seeing a great number of postings on audio recording and DAW forums from people either asking that very question or people who thought they understood, but didn't. It is assumed that you have a basic understanding of sound and computing.

Most people understand that there is a correlation between these two settings and sound quality. And naturally most people assume that the higher the numbers, the higher the quality. Which is true, but there are consequences when using higher settings as well as benefits, which we will discuss later.

Lets begin by establishing what these terms actually are:

Sample Rate

The sample rate, or recording frequency is the number of times per second that the computer measures the amplitude and frequency of the sound coming in.

Sound is a wave. In order to understand how a computer records that wave, we need to understand a few basic principles about it's characteristics.

The amplitude is the height of the wave, and the frequency is the measurement of how many complete waves pass per second. The way that this relates to our perception of the sound is that the higher the amplitude, the louder the sound seems. And the higher the frequency the higher in pitch the sound seems.

In terms of recording that sound, the more frequently the computer takes a measurement of the sound wave the more accurate the description of the audio will be.

Because the human ear registers sound with a frequency roughly in the range of about 20-22,000Hz, there is a theoretical minimum sample rate that we need to use to record most sounds accurately. At the very minimum we need to take a sample of the audio wave more often than once per cycle. If we were to only measure the wave once per cycle, then for example, we might end up measuring the wave at the same point in the cycle each time. Which wouldn't give us a very accurate description of the wave's amplitude and frequency.

Therefore, the commonly accepted theoretical minimum sample frequency is 44,000Hz. It is no coincidence that the sample rate for audio CD's is 44,100Hz or 44.1KHz. Modern DAW's and sampling software often offer a range of sample rates such as 44.1KHz, 48KHZ and 96KHz.

Bit Depth

The recording bit depth determines how many discreet levels there are to describe that waveform at each measurement.

One way of thinking of this is to imagine a ruler or tape measure. If it is only marked in centimetres, then the most accurate measurement you could make of something would be to the nearest centimetre (or possibly every half centimetre at a guess!). But if it is marked in millimetres, you can make measurements that are ten times more accurate, down to the nearest millimetre.

So an 8 bit recording will have fewer discreet levels to describe the wave for each sample than a 16 bit recording. And a 24 bit recording has even more than 16 bit.

In practice

So naturally most people set their software to record 96KHZ at 24 bits, knowing that they are going to end up with the best possible sound quality once the signal is inside the computer. Of course the chain is only as strong as the weakest link, if the sound source hisses and the cables are cheap, and one of the jack plugs is loose, then all that gets recorded too.

A consequence of recording with such a high sample rate and bit depth it that there is much more data to store per second of audio being recorded (as there are more samples being taken of the audio). Likewise, the higher the bit depth, the more discreet levels of data there are for each sample taken. Which produces more data that has to be stored.

There is therefore a decision to be made. Is it worth recording at a sample rate of 96KHz? The resulting volume of data for the computer to process and store would have been prohibitive just a few years ago. But now, even very basic machines are more than capable of handling the data produced when recording several tracks of audio at this level of accuracy.

As I suggested before, if you are recording less than perfect sound sources or using cheap cables, mics, or connectors or are subject to the multitude of mains noise issues (which I'll go into at a later date) then is it really worth recording at the maximum possible sample rate and bit depth?

The most useful and practical way of assessing which sample rate and bit depth to use is to test the material being recorded as combinations of different settings whenever possible. If you can hear no discernible improvement in sound quality, then you need to decide if it is really necessary.