See also reduce controls.
The reduce window is where you collect the various scans to create the reflectivity signal. This is a confusing screen.
The basic formula is (specular-background)/slitscan. In addition there is scaling by the transmission coefficient of the sample environment, polarization correction for polarized beam and footprint correction.
Select all specular lines. These will appear on the legend annotated with a plus, and the "specular" entry will show the average of all the specular lines. The highlighted line is the final reduction, which at this point will just be the average of the specular. So selecting one line will display it overlayed by the identical average overlayed by the identical reduced line.
Next select all background lines. These will appear on the legend annotated with a minus, the "background" entry will show the average of all the background lines, and the "subtracted" entry will show the difference between the specular and the background. The highlighted reduced line will now overlay the subtracted line. Note that psd data will already have the background subtracted from the specular, so no need to select any.
For NG-1 and X-ray data you need to divide by a slit-scan. This is a sequence of data taken with the same slit settings as the specular and background data but directly counting the beam rather than its reflection. Because the detector is directly in line with the beam for a slit scan, there is no Qz value intrinsic to the specific slit openings. Instead for each Qz in the specular data, you must lookup the value of the slit 1 opening for that particular Qz in the slit data and this is your slit scan. Linear interpolation is used if the slit values are not exactly aligned. NG-7 data does not use slit scans but instead the specular and background data is divided by the monitor count when the data is loaded. Note that NG-7's monitor does not count all incident neutrons. Instead, you will have a "slit to monitor" ratio by which you need to scale the reflectivity curve. You can enter this value as a factor in the "incident medium transmission coefficient" or if it is measured in a file, divided it through as a single point slit scan [untested as of this writing]. Note that the slit to monitor ratio decreases as the slit width approaches zero, so for careful experiments you may want to measure the slit scan across many points.
Select all the slit scan lines. These will appear on the legend annotated with a slash, but will not appear on the graph. If you select the legend entry, the actual data will appear but it will be counts versus slit 1 where slit 1 values go across the top axis rather than along the bottom axis, so there will be no correspondence with the Qz values of the specular or background curves. The slit scan entry represents Qz value of the specular interpolated into the average of all the individual slit scans. The divided entry is (specular - background)/slit, which will be overlayed by the highlighted reduced line.
The next step is to calculate the footprint correction. Click the Parameters... button to enter the footprint parameters. You can suppress the footprint correction by toggling the associated check box on this screen.
Finally, you can scale the data by the "incident medium transmission coefficient". If you "Guess" at the value, then it will scale the data so that the peak reflectivity is 1.0. The error on the scale factor will equal the error on the peak value.
Save your data as linear and/or print the graph and you are done. The header will contain details of all the files which went into the reduced (Q,R,dR) data columns.