> > Please explain how that can be done. I know there are profiles for certain > types of film (Kodak has a well known one for E6, as does WCI), but tell us how > a digi cam can be profiled for a specific (or even a general) type of film.
Technically, of course, it can't. There will always be metamerism issues. You can't escape these; the sensor sampled the whole imput spectrum of light falling on it, convoluted it with the response curve of the sensor, and reduced the whole thing to a single value. There's no way to reverse that process to get back to the original stimulus, and then simultate the behaviour of a sensor with a slightly different curve. In practice, though, the overall shapes of most response curves are sufficiently close to each other, and the primary colour axes chosen are reasonably closely aligned in colour space (and to those of the final arbiter, the human eye), that for all practical purposes you can handle it with a simple colour space profile. If this weren't the case then there would be no way to colour balance monitors, printers, etc., or to adjust the white-balance of the sensor. Nor is this only an issue in the digital world; there are many analogies in the analogue world, too. You rely on the similarity of the shapes of those response curves each time you stick a filter on the front of your lens; you don't have a bag full of different filters, each matched for use with only one film type. > > Herb Chong wrote: > > > you can profile a film and a digital camera and directly match the digital > > camera to the film's profile over a moderately wide number of conditions. > > this is most easily done in a studio situation where the lighting is highly > > controlled. > > >
