Ivie IE-45 STI-PA - Speech Intelligibility Analysis - Manual (Page 4)

--AVERAGING--
The IEC Standard 60268-16 (2003-05)states the following:
"The measurements should be repeated several times and the results averaged. Also, an estimate
of the standard deviation should be derived from them and included with the results."
The Ivie STI-PA software allows for multiple measurements at a single location to be averaged,
and the standard deviation (of those measurements) calculated and displayed.
--HOW TO AVERAGE--
Upon completion of a measurement you can do one of three things with the data.
A. Press the Save button to immediately save that single measurement. This precludes averaging
it with future tests.
B. Press the "Clear" button to discard the measurement. (Should you believe it invalid)
C. Press the "Avg" button to average this measurement with successive measurements.
Procedure for Averaging:
1. Press the "Start" button to institute a measurement.
2. View the measurement result and if you consider it valid go to Step 3. If the you consider the
data invalid then press the "Clear" button. The data will be discarded and NOT averaged with
other measurements. Now start another measurement.
3. Press the "Avg" button. Do NOT press the SAVE button. You will see a "1 test avg-0.XX"
displayed on the screen.
4. Repeat steps 1 through 3 for as many measurements you wish to average. Press "Start",
Review data, Press "Avg".
5. After pressing the "Avg" button for the concluding measurement press the 'Save" button. This
saves the average of the preceding group of measurements and clears the average buffer. You
can view the results by pressing the "File" menu and then selecting "View Job File..."
--HOW AVERAGING WORKS--
The data from the Averaged STI-PA measurements are summed and then divided by the total
number of tests. It is an arithmetic average of the STI-PA values.
The CIS is calculated from the Averaged STI-PA value.
The Std. Deviation is the Standard Deviation calculated from individual STI-PA values used in
the average.