Highly accurate frequency estimation of brief duration signals in noise

A highly accurate frequency estimation providing suppression of windowing effects, denoising performances and frequency resolutions in excess of Gabor–Heisenberg limit, is proposed for brief duration signals. It is shown that unbiased frequency estimation with vanishing frequency variances is achieved far below Cramer–Rao lower bound when signal-to-noise ratio reaches vicinity of threshold values. Observed performances provide novel and valuable perspectives for efficient and accurate frequency estimation for brief duration signals in noise.