N-sample level-crossing estimator methods and devices are provided that extract more information from given time samples than the current two-sample approach and that are more resistant to interference from noises. The two-mean level-crossing time-interval estimation method extracts more information from given time samples than existing methods, advantageously estimates a level-crossing time interval with a limited number of time samples and is quieter than current noisy estimation techniques. The two-mean level crossing time-interval estimation method for N-sample estimation uses all N time samples by calculating the mean value of the first N/2 time samples and subtracting it by the second N/2 time sample to average out the noises in time samples. The two-mean level crossing time-interval estimation method can be implemented by using a Finite Impulse Response filter to take level-crossing time samples as inputs, take the differential level-crossing time samples as inputs, or take the N/2-step differential level-crossing time-interval as an input. An addition only one-step differential level-crossing time-interval estimator device and a one-step differential level-crossing time-interval estimator device are also provided.