digital-audio.org
Information centre for all about digital audioResolution
Related Pages
Try these next ...
External Links
Sampling Resolution
As explained on the A to D page the resolution at which audio waves are sampled refers to the accuracy of the measurement. This accuracy can be thought of as how fine a measurement we take.
In the picture below, we have shown the samples having values to one decimal place. Well they could have been recorded more accurately, for example the first sample could have been recorded as 1.3276 volts. Since the data is ultimately used to re-create the wave at a later stage, then the accuracy of the sample is very important.
So - why would we use less accurate data? Well the numbers that are measured have to be stored somehow, and in the world of digital storage the number 1.3276 takes up more storage space than the number 1.3 . We therefore have to strike a balance between accuracy and the size of the data we have to store.
When CDs were invented, the resolution was decided on as 16-bit. Since we are talking about digital data, ultimately the numbers are stored as binary digits, and 16 is a round number in binary terms being 2x2x2x2 = 16. 16-bit data allows the storage of a number between zero and 65535. In our example in the picture, if 5 volts is the maximum level then 1 bit represents 0.000076 volts ( 76 micro volts ), so that's pretty accurate!
Since the invention of CDs and DAT ( which also uses 16-bit sampling ), it has become apparent that more up to date storage and electronics can cope with a higher resolution. this means we can increase the quality of digital audio, to make the overall performance better. Most recording studios, and computerised audio recorders now use 24-bit instead of 16-bit. The sound quality is markedly better at this higher resolution. The same waves are now sampled with data between zero and 16777215. In the example above 1 bit now represents 0.000000298 volts ( 0.298 micro volts )