Sampling converts a continuous-time signal into discrete time, and quantization maps amplitude to a finite set of levels. For an ideal uniform quantizer with step size $\Delta$, the quantization error is often modeled as uniformly distributed over $[-\Delta/2, \Delta/2]$, giving mean-square error $\Delta^2/12$ under standard assumptions.

Modern data converters push into energy-efficient sparse sampling, time-interleaved architectures, machine-learning-assisted calibration, and in-sensor/near-sensor analog compute. The hardest issues are mismatch, interleaving spurs, finite op-amp gain, and robust calibration under process and temperature drift.