Dither
Figure 8-13 Dither component.
Dithering (see Figure
If, after reducing the number of bits in a signal, the original high- resolution (more bits) signal is subtracted from the final
(Final Signal) - (Original Signal) = (Error Signal)
Some very simple algebra shows that the final signal can be thought of as the sum of the error signal and the original signal:
(Final Signal) = (Error Signal) + (Original Signal)
When audio data words are shortened by simple truncation, it turns out that the error signal is correlated with itself and with the original signal. This simply means that the error signal is not pure white noise, and that the frequency and amplitude of the error signal is in some way related to the frequency and amplitude of the original signal.
If noise with some very specific amplitude and spectral characteristics is added to the original signal before truncation, then the correlation is destroyed and the error signal sounds like low level noise instead of tones. Even though the added noise actually raises the noise floor in the final
This addition of noise before truncation is called dithering. The noise that is added is called the dither signal.
In summary, dithering is a way of making digital audio better by making it worse.
Bit Level
In the
Processing Components - Dither