Threshold
This calculation determines the intervals where the Data Series exceeds the specified threshold. The result is a new Data Series of the same length with zeros indicating samples that did not exceed the threshold and ones indicating samples that did exceed the threshold.
First (if specified) the data is rectified and then a low pass filter (if specified) is applied using the parameter Low Pass Cutoff Frequency. This filter is a finite impulse response (FIR) filter using a Blackman window with a filter kernel width of 101.
At this point the threshold calculation is performed, to determine the samples that meet the threshold level. These results are then subject to suppression (if specified) and scaling (if specified).
This calculation uses fixed or percent threshold levels. A Fixed threshold level is a user specified value compared directly to each sample in the input series. A Percent threshold level is a user specified percentage of the total range of the data. This is useful if processing multiple series which have different amplitudes.
If specified, the resultant data’s spikes and gaps can be suppressed to get rid of unstable regions of the data. The user must specify a Suppression Time which is the minimum amount of time that the data must remain in one region (active or inactive) in order for that region to appear on the output (in other words, a span in the data that is above/below the threshold level must be longer than Suppression Time in order for it to appear in the output as a region that was above the threshold level).
In the script this calculation is performed by using the built in function thresholdsplit(…). This function takes in each of the user specified parameters and returns an array of windows that correspond to the chunks of the data that are above the threshold level. The script then iterates through these windows, checking their start times, and creating a new series (the output) containing the ones and zeros that indicate the samples of the input data series that are above or below the threshold level.
Finally, the output can be scaled to be equal to the standard deviation of the input signal when exceeding the threshold criteria, when Scale Output is selected. This is useful for creating an output that can be overlaid with the original data.
Parameters