spectro-aras.com

Christian, I was wondering if ISIS could be upgraded to provide an error estimate for the signal level across a spectrum. I had in mind that when stacking several spectra the variation in the level (spectra to spectra) could be used to estimate the noise (variance, standard deviation or standard error) by wavelengths. Then if an option to output this information in a text file, or include it in a plot of the spectra, would be excellent.

Thanks Andrew

Andrew,

A possible response. First select option "No" in the section "Erase automaticaly intermediate file" of "Settings" tab.

Note the avaialbility of @rawxxx.dat files -> raw profile just after binning directly in digital count.

Note also information about mean (and median) level of individual raw profile (new feature of very
new V5.1.0) - see attached screen copy.

Christian

Hi Andrew,

When trying this, make sure the individual spectra are well calibrated and aligned in wavelength, which of course they will be using the various calibration tools in ISIS particularly if the spectrum is "spiky" otherwise the misalignement appears as noise.

Cheers
Robin

Thanks Christian I will give it a go. Also thanks Robin for the advice. I think a better quantification of error (S/N and calibration) is the one missing element in our quest for profesional quality spectra.
Regards Andrew