Image and text supplied courtesy of Dennis diCicco

"Spectra shot with a my homemade spectrograph using a light source that has mercury (Hg) emission lines and broad emission bands. There are, however, gaps in the color continuum, so "missing" areas do not indicate a lack of sensitivity on the part of the camera sensors. In addition to the marked Hg lines, I've also shown the position of the 4 astronomically important wavelengths. It's interesting to note that both the Canon 10D and NikonD70 with the IR filters stop recording at just about the extremely important H-alpha wavelength, while the Hutech modified Rebel goes well
beyond that wavelength. It makes a good case for not having the filter in a camera used for astrophotography. Also, with the filter removed, there is more light hitting the CMOS across the whole spectrum, so there's more "signal" recorded at all wavelengths.

All three spectra exposures are exactly the same within the limits of the individual cameras hitting their "target" ISO values (all were shot at an ISO setting of 400) and the actual lenses delivering there set f/8 values -- I used both Nikon and Canon lens on their respective cameras to record the spectra, and the lenses were manually set to the f/8 value. Unfortunately, f/numbers don't seem to mean as much to today's cameras with autoexposure systems as they did in the days of separate hand-held light meters (but I'm dating myself...)."