Jittered

Jittered is the past participle of jitter and is used as an adjective to describe something that has been subjected to jitter—small, random variations in time, position, or value. In technical contexts, jittered items are perturbations added within bounded limits to reduce systematic artifacts or to reflect timing variability in a system.

In computer graphics and numerical analysis, jittered sampling is a technique for selecting sample points with a random offset inside each cell of a regular grid. The unit area is partitioned into subregions, and one sample per subregion is placed at a random position within that subregion. This approach reduces structured aliasing and correlation between samples, improving anti-aliasing, texture sampling, and the accuracy of Monte Carlo integration. Compared with purely random sampling, jittered sampling often yields more uniform coverage and faster convergence.

In signal processing and digital systems, jitter refers to timing variation of sampling events or signal transitions. When such timing deviations occur, the resulting samples are described as jittered. Clock jitter can distort digital and mixed-signal systems, motivating design techniques to minimize timing uncertainty and to compensate for jitter in receivers and converters.

In data augmentation and machine learning, jittering describes small random perturbations of inputs (for example, translations, noise, or other perturbations) applied during training. The goal is to produce a jittered dataset that improves model robustness to real-world variations.