Filter relationships

A second question (after binning in another thread).

I am assuming that any given filter except clear cuts out a certain amount of light. If the clear lets it all in, the red lets in about half, the green about half, and the blue, say, about two fifths or something. This will vary from manufacturer to manufacturer, but it could be measured and calibrated for any given set of filters/camera. And the measurements would change according to the color of the target.

With that in mind, if I could figure out the "efficiency" of my filters, and knew the perfect exposure time for "clear," could I multiply the inverse of the efficiency of a given filter by the perfect clear exposure time to get a good starting point for a color filter exposure?

Thanks for the advice.

Alex

Unfortunately, you are missing on key element. Knowing the efficiency of your filters doesn't tell you the quantum efficiency of your imaging chip! Different chips have significantly different efficiencies at different wavelengths. In the best of all worlds, the filters are matched to the particular chip being used so that you will get perfect "white" (G2V) balance with identical exposure times. For example, Astrodon makes both E-series and I-series RGB filters for use with Kodak's blue enhanced chips and the interline transfer chips respectively.

Based on your assumptions, you would expect that you would need to collect more color data than luminance data since the luminance filter will let in more light. In practice, that isn't true. The physiology of the human eye/brain enter into this as well. The eye/brain combo is more sensitive to noise in luminance than in chrominance channels, so most users find that collecting lots of luminance data is the key to a good image. As a starting point, you might try a 3:2:2:2 ratio with the L:R:G:B channels and see how that works for you. There is no single right answer, though, and getting the same signal strength in all four channels does not necessarily result in the best image.