Dither subtractive

Dither can be subtracted out ( subtractive dither ) after quantization and thereby whiten the signal. But in most systems this may be either difficult or impossible to do because the dither signal is not available. Therefore non-subtractive dither is the most common use of dithering. Yamasaki [Yamasaki, 1983] discusses the use of large amplitude dither (as much as 32A) in subtractive dither systems. 

A typical example of subtractive dither use in an A/D converter can be found in a Teac patent [Nakahashi and Ono, 1990], Notable features include the use of limiting and overload detectors as well as the ability to control the dither amplitude. A clever example of non-subtractive dither [Frindle, 1995][Frindle, 1992] separates the input signal into two paths. One of the paths is now inverted and then both paths are added to a random noise source. After the DAC, the two analog signals are subtracted the result is the sum of the conversion errors and twice the input signal. 

---