KDFX Reference
KDFX Algorithm Specifications
10-131
744 Quantize+Alias
Digital quantization followed by simulated aliasing.
PAUs: 1
The Quantize+Alias algorithm offers some of the worst artifacts that digital has to offer! Digital audio
engineers will go to great lengths to remove, or at least hide the effects of digital quantization distortion
and sampling aliasing. In Quantize+Alias we do quite the opposite, making both quantization and
aliasing in-your-face effects. The quantizer will give your sound a dirty, grundgy, perhaps industrial
sound. The aliasing component simulates the effect of having sampled a sound without adequately band
limiting the signal (anti-alias filtering).
Quantization distortion is a digital phenomenon caused by having only a limited number of bits with
which to represent signal amplitudes (finite precision). You are probably aware that a bit is a number
which can have only one of two values: 0 or 1. When we construct a data or signal word out of more than
one bit, each additional bit will double the number of possible values. For example a two bit number can
have one of four different values: 00, 01, 10 or 11. A three bit number can take one of eight different values,
a four bit number can take one of sixteen values, etc. An 18-bit digital-to-analog converter (DAC) like the
one in the K2661 can interpret 262,144 different amplitude levels (2
18
).
Let’s take a look at how finite precision of digital words affects audio signals. The figures below are plots
of a decaying sine wave with varying word lengths.
Figure 57 A decaying sine wave represented with different word lengths: (i) 1-bit, (ii) 2-bit, (iii)
3-bit, (iv) 4-bit.
(i) (ii)
(iii) (iv)
To Next Page
Loading...
