When I add a Measurement, I have the option of applying it to the Whole Trace or Between Rulers.
For the Whole Trace, is the data used in the Measurement calculations limited to what is displayed on the screen at the moment of each calculation? Or, are the Measurement calculations applied across all the data captured during a session (since Go was pressed), which would be the data currently displayed plus the data that has "rolled off" the screen?
A measurement is always performed on the data that has been captured, and then stored in a buffer, which is then represented by the trace drawn, or being drawn on the screen.
In fast sampling mode, a measurement for the whole trace applies to the trace that has already been drawn at the time of the measurement.
For slow sampling mode, the measurement is for the trace that is currently being drawn, (so if you stop the trace you could have only a partial trace, and partially filled buffer, which will be the data used). If you have set the 'Slow Sampling Display' to display the previous waveform buffer, so that you have a rolling display, then the measurement only applies to the new trace being drawn behind the separating bar. Any data represented by a previously drawn trace that has rolled off the screen is not included in a measurement.
Gerry, thanks for your comprehensive answer. As I understand your answer, the Capture Count indicates the number of buffers of data used for Measurement calculations of Whole Trace.
A related question for a Measurement between rulers: In this case where rulers define the range of the data series on the timebase, does the Capture Count value displayed have any relevance? I would think that a Measurement between rulers is always based on the most recent capture which is displayed on the screen.
Sorry I misunderstood your question, I thought you were referring to the instantaneous measurement. If you are capturing measurement statistics, rather than just making an instantaneous measurements then the capture count relates to the number of buffers used for the calculation where the measurement is performed on the complete buffer. However, not all measurements use the complete buffer and the capture count then indicates other units of data collection used for the measurement, such as pairs of values for 'peak to peak', and numbers of crossings through thresholds for 'frequency' (particularly noticeable at slower sample rates).
Measurement statistics are still done between rulers, if you have specified that the span is between rulers, and this does have relevance (particularly where you're using a trigger to capture a repetitive waveform, so that the waveform segment between the span is always the same).