What is a Bit?
A Bit of audio is the resolution of audio. Audio is turned into a binary word of 0’s and 1’s.
8 bit audio example: 01010010
16 bit audio example: 0100111010101101
24 bit audio example: 011100101001110000111010
Bit Depth and Levels
The more number combinations you can get in these binary words the more steps of level you can get in amplitude. Bits are with each sample of audio. You can choose any sample of audio and freeze it in time to see the Bit Depth.
8 bit audio example: 256 possible level
16 bit audio example: 65,536 possible levels
24 bit audio example: 16,777,216 levels
With every bit of greater resolution, the number of levels double.
What This Means For 16 Bit Vs 24 Bit Audio
It comes down to signal to noise ratio of the reconstructed signal to a maximum level determined by quantization error, in other words the dynamic range. You can imagine that more steps we can record for each sample the more distance there is in between the signal and the noise. There are theoretical maximums a signal can be reproduced before Quantization noise becomes an issue. Quantization noise is a model of quantization error introduced by the sampling process during analog-to-digital conversion (ADC)
16 Bit – 96.33 dB
24 Bit – 144.49 dB
This all means the more steps you have, the better the resolution from a signal to noise ratio you have. Dynamic range is the difference between the smallest to loudest signal.