Do the Math
You can think of digital recording as plotting a waveform on a graph whose x and y axis are time and level. The "x" time axis is the sampling rate, which is either 44.1, 48, 88.2, 96 or 192 kHz. A higher sample rate increases the fidelity of the high frequencies, but at a cost. A file recorded at a 96kHz sample rate uses twice as much disk space as a file recorded at 48kHz, and taxes the processors twice as hard. Most systems reduce their track counts by half with a 96kHz sample rate.
The y axis is the level of the signal. This is expressed in bits. The standard CD bit depth is 16 bits, which is 216, or 65,536. A bit-depth of 24 is 224, or 16,777,216. Yet, because of the way a computer stores data, 24-bit recording only uses 50-percent more disk space. Every bit that''s added to the bit-depth gives us 6dB more headroom. That means that 24-bit recording has 48dB more headroom than 16-bit. In other words, you can record at 48 dB and still get full CD quality. A 24-bit recording does not tax the CPU much more than 16-bit recording. So for a minimal increase in disk space and CPU load, 24-bit recording gives you a huge improvement in sound quality and headroom.
Signal processors have to round out to the nearest digit when they process a wave. If the bit-depth is low to start with, this can result in noticeable degradation of the sound. With 24-bit audio, the rounding-off is much smoother, hence more accurate and better-sounding.