kvantiseeritust

Kvantiseeritus, or quantization, is the process of mapping a continuous quantity to a finite set of discrete values. It applies in physics, where many observables such as energy levels or angular momentum are inherently discrete, and in engineering, where continuous signals are represented with a limited number of levels in digital systems.

In physics, quantization means that certain physical properties can only take specific values. This idea emerged from Planck’s work on black-body radiation and the development of quantum mechanics, leading to discrete energy spectra in atoms and quantized light as photons. Quantization underpins much of modern physics, including atomic structure, spectroscopy, and condensed matter phenomena.

In signal processing and measurement, quantization occurs during analog-to-digital conversion. A continuous signal is divided into steps of size Δ, and each sample is assigned to the nearest discrete level. Uniform quantization uses equal steps; non-uniform quantization uses variable steps to better match signal statistics (for example, μ-law or A-law companding in audio). The difference between the original sample and its quantized value is the quantization error or quantization noise, which degrades fidelity.

Key properties include bit depth (number of quantization levels), dynamic range, and the resulting signal-to-noise ratio. Quantization error is bounded by roughly half a step for mid-rise or mid-tread quantizers. Techniques such as dithering can reduce perceptual artifacts and alter the statistical nature of the error.

Applications cover digital audio, imaging, communications, and instrumentation. Quantization is typically paired with sampling, as dictated by the sampling theorem; together they transform a continuous-time, continuous-amplitude signal into a fully digital representation.