This document discusses quantization in digital signal processing. It defines quantization as the process of converting continuous amplitude signals to discrete amplitude levels. Quantization introduces quantization noise, which can be modeled as having a uniform probability distribution between +/- Δ/2, where Δ is the quantization step size. The signal-to-noise ratio increases exponentially with more quantization bits. Quantization can be uniform, where the step size is constant, or non-uniform, where lower amplitudes have smaller steps to reduce error. Compression techniques like μ-law are used prior to non-uniform quantization.