One of the resultant problems of digital recording is a phenomenon called quantization distortion. It’s a grainy sound that often happens when audio is processed from a higher bit depth, like 24bit, to a lower one, like the CD-quality Red Book Standard of 16bit.
The solution to this problem is a processing effect called dithering. Dithering essentially is adding an inaudible layer of low-level noise to cover up the quantization distortion.
Really bad quantization distortion can cause pops, crackling sounds and a reduced stereo image. That’s why it’s best to apply dithering almost always when creating a master that has a different bit-depth than your source.
Dithering is often the final step in the mastering process, when the track is bounced into the final digital format it’s being distributed in. Most often, this looks like creating a 16bit Redbook track from a mix that’s been recorded at 24bit or 32bit floating point.
There are different types of dithering–Apogee, Waves, POW-R…–but their differences are hard to appreciate without technical engineering expertise. You only need to use dithering when you’re truncating the bit depth of the audio you’re working with. The one exception to this is that many producers don’t bother to dither when they’re exporting straight to MP3.
While it’s generally best to leave mastering to a professional studio, where an engineer with the proper equipment, room acoustics and expertise can fine-tune your track most precisely, it’s helpful to know about dithering and its role in the mastering process.