** this is not a question that can be answered with short sentance or two, TL/DR version DNE **
Dither or Dithering is a rather misunderstood concept. If your making or mixing music you wouldn't ever actually have to use it. Dither is the very LAST thing that should be applied to audio...so, unless your finishing up something to be released out into the world (i.e Mastering), dither isn't something you need to do!
However, to answer your question in the shortest form possible...
--> What is it? -->
Dither is VERY low (quiet, low volume, low amplitude) noise added to a digital audio file.
It is used on BOTH full size audio files (.wav, .aiff)
and on compressed files:
--> lossy (.mp3, .ogg)
--> loseless ( .aac )
--- > Why is it used? ---
Why in the world would noise be added to digital audio files at the very last stage of finishing?
To understand the answer it important to understand that anything and everything that is digital (audio or what have you), at its core is represented using a system of "OFF" and "ON" switches... ie everything is represented as a '0' or '1'.
Audio in the digital world is no different! To facilitatate more resolution in the digital representation of analog sound, a group of switches can be used to represent more then just ON or OFF. When you use a LOT of these switches, you can approximate the analog wave form pretty well... (** see below for an example of how many switches are used in common digital media**)
However, no matter how many switches are used; it is still an inescapable fact that at the core of anything digital you only have ON or OFF.
So, if you have a piece of digital audio that is NOT dithered, and you listen to it fading out... at some very quiet point near the end of it fading to silence.... It will eventually drop off the cliff, or it will go from very very quiet to dead silence all at once. Why? because... deep down, digital anything is based on that ON or OFF switch.
When you apply dither to audio, you are adding an almost imperceptible noise floor, so that as audio fades away to nothing, it will fade away in to this noise, instead of falling off of that cliff.
How dither is actually applied at the sample level, or why its not a good idea to dither an already dithered digital audio file, or a discussion about why dither is less useful (or needed) when bit depth increases beyond CD quality ... are all very valid questions but they are outside the scope of this answer.
Interested to learn more about dither or any audio/sound/music production concepts or best practices? Feel free to contact me here. Always happy to help out where possible!
- Raven
******
CDs use a digital audio format of 44.1K/16 bit, meaning that every second of mono audio is approximated by 44,100 samples, with each sample using 16 switches (bits) to represent each sample -- && --- X 2 because CDs are stereo.
(sample rate & bit depth in depth questions, are outside of the scope here. Feel free to contact me here for further discussion)
*****
