Comparison of some algorithms to estimate the low and high state level of pulses

The Subcommittee on Pulse Techniques (SCOPT) of the Instrumentation and Measurement Society is developing a new standard for pulse analysis. Currently we are working on the estimation of the low and high state levels of pulses. Since the waveforms considered contain two states, the terms “low state” and “high state” may be used interchangeably with the terms “state 1” and “state 2,” respectively. The estimation of other pulse parameters, such as pulse amplitude, transition duration (rise and fall time), overshoot and undershoot pulse duration or width, and settling parameters, depend on the estimate of the levels for the low and high states. So, these algorithms are the foundation for the proper estimation of pulse parameters. Some other names for the low and high state levels of pulses are “base magnitude” and “top magnitude”, as well as “base-line”, “bottom line” and “top-line”. Previous work described algorithms for estimating the magnitude of the state 1 and state 2 levels that are based on histograms of the pulse waveform. However, the algorithms leave many choices to the user. Due to these user choices, different results can be obtained for the state 1 and state 2 levels that are all compliant with the standard. We compare the results of several different algorithms