This is is similar to MATLAB's pulsewidth function, with higher capability, better detection and it is less susceptible to noise. It has 4 methods, best of which is 'Dynamic' which extracts the signal using two thresholds, one of which is determined automatically according to the data and make the algorithm reliable in case of a noisy signal. Other methods are including 'Middle' and 'MiddleSmooth', later uses a low-pass filer (embedded in the same file) to filter the data drastically and determines the pulses from the filter data however extracts them from the actual data! Problem is that the filtering frequency is fixed and need to be adjusted according to your measurement setup.