Jim Bell, Andy Switala, Mark Robinson; 10/18/95
Preliminary comparison between NIS and MSI sample measurements
We used the preliminary MSI sample spectra compiled by Mark Robinson along with the NIS sample spectra compiled by Jim Bell and Andy Switala to assemble composite MSI+NIS spectra for each material.
The details of the processing of the MSI and NIS spectra can be found in the links cited above. The spectra used here are the "final" results from those still-preliminary processing steps. Each spectrum is a ratio to the same halon sample, and has then been multiplied by the reflectivity of halon to derive the reflectance of the sample. For NIS, the narrow slit, gain 10 spectrum was used.
Error bars for MSI represent the variability over the 40x40 pixel box used by Mark in each image; error bars for NIS represent the average over the 25 spectra measured for each sample. We note that both the MSI and NIS data shown here are still in a preliminary form of reduction, as additional calibration refinements are necessary. For example, some residual differences between NIS and MSI at the very shortest NIS wavelengths are probably due to incomplete characterization of the NIS spectral crosstalk correction, which is still being understood and developed.
The following plots show the merged MSI+NIS spectra at this stage of the data reduction. The filled squares are the MSI data points. Open circles are the NIS Ge data points, and open triangles are the NIS InGaAs data points.
What is remarkable is how well, in general, the MSI and NIS spectra match up in the overlap region. We note that THERE HAS BEEN NO SCALING OR OTHER MASSAGING of these data whatsoever: the fact that they overlap nicely is, apparently, simply a reflection of the inherent quality of the data and of our ability to measure the same samples with two different instruments using as close a setup as possible.
Last Modified by Jim Bell on 14 November 1995.