Abstract
The time-sequential sampling of signals that vary in time and have one or more spatial dimensions is analyzed for both deterministic and nondeterministic signals. It is shown that the severity of frequency-domain aliasing because of undersampling is strongly dependent on the temporal ordering of the spatial sampling points in each period of the sampling pattern. The signal-to-aliasing-noise power ratio provides a quantitative means of assessing this dependence. Conventional lexicographic sampling, a bit-reversed sampling pattern proposed by S. Deutsch [IEEE Trans. Broadcast. BC-11, 11–21 (1965)], and random sampling are compared for white band-limited signals. It is found that, in some cases, bit-reversed sampling is an improvement over conventional patterns, but other patterns probably exist that would be even better.