In digital samples amplitude is represented as “bit depth.” Bit depth determines both the number of steps, or levels, that are possible in a sample and how loud a signal the system can tolerate. CD-quality sound uses 16 bits. We can calculate the number of steps by raising the number two to the 16th power. 2 ^ 16 = 65,536 steps. The number of steps is divided into 32,767 positive (plus 0) and 32,768 negative steps, representing the crests and troughs in a sample of music. Each time we take a sample of a piece of sound the actual amplitude must be rounded to the nearest available level, introducing something called a “quantization error.” Simply put, this means that a small amount of noise is produced each time a digital recording is sampled.
The next section will cover the signal-to-noise ratio, which is the amount of inherent noise versus the system’s capacity for the desired signal. In theory the overall capacity of a digital system is approximately six decibels per bit. For a 16-bit CD-quality a system can tolerate 96 dB. (Source: Cornell University Music Department. http://digital.music.cornell.edu/cemc/book/export/html/1594).