Digital Image Studio wrote: > On 24/08/07, Toralf Lund <[EMAIL PROTECTED]> wrote: > > >> Actually, Wikipedia says: >> >> ISO Standard 12232:2006 >> >> <http://www.iso.org/iso/en/CatalogueDetailPage.CatalogueDetail?CSNUMBER=37777> >> ("Photography — Digital still cameras — Determination of exposure >> index, ISO speed ratings, standard output sensitivity, and >> recommended exposure index") defines ISO speed in terms of the >> amount of light needed to achieve a certain "quality" in the sense >> of a per-pixel signal-to-noise ratio. [ ... ] >> >> >> > > The problem is that the strict definition of ISO film speed can't be > directly applied to direct digital image capture. They record light > levels in a completely different fashion, film being non-linear and > digital very linear. > Yes, but I think that if you look at it in a somewhat more broader sense, the definition I was referring to makes the most sense. I mean, maybe it's in the understanding of how to interpret the ISO value of film I've missed something, but if I have not, the ISO rating tells you how much you have to expose the film to reach a certain point relative to the base; it says nothing directly about how much you can expose the film after that before it's over exposed. And if the maximum value isn't included in the equation for one medium, it seems unnatural to drag it in for another - regardless of the fact that curves are different etc... > The "ISO" bestowed on a digital sensor is an ISO that mimics the > sensitivity of a film of the same speed for a middle gray based on the > absolute white point (just before saturation) a black point (before > noise becomes the larger signal) from what I have read. > If you say so.
In any case, doesn't the real point still remain? Doesn't a lower native ISO mean that you have to amplify the data more at a given higher ISO setting? If it does, asking for a lower ISO is in reality asking for more noise at higher ISO values, since when you amplify the data, you also amplify the noise... - Toralf -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net
