>> I was talking about the number of different voltage levels that may be >> output from the sensor itself. [ ... ] >> > > I've thought about it for long enough ... a year or two ago. > > There's no point in considering the A/D conversion as separate from > the sensor *because you can't do anything about it*. Yes and no. You (or was that someone else?) were also talking about understanding the limitations of the equipment, and when you want to do that, you should always try to see the full picture. And the dynamic range of the actual sensor is part of that picture, too. Also, doesn't the fact that the signal is (potentially) amplified after it leaves the sensor, but before it enters the A/D converter, mean that you can do a little bit about it in one sense? > The RAW data > output is what you can manipulate and represents conceptually a > linear gamma space. [ ... ] >
> With regard to the data you can manipulate, using the strategy of > providing enough exposure to place Zone IX properly just under the > saturation threshold will *always* produce the best data to work with > regardless of ISO because the reason it works is founded in numerical > operations. Yes, I believe that is true. However, I also think it can be proved that your method will not give better results than certain others (like using the old good old 18% grey) at higher ISO settings. But it will never give worse results than other methods, either, so there you have it... > If you plot difference in the response curve of the > sensor at different ISOs, you could also use that information to > expand or constrain the dynamic range you want to apply to scene > capture, but that says nothing about proper exposure evaluation ... > which was the point of my explanation. > Yes. I think you did a much better job explaining it than that other article, which I initially read twice instead of reading two sources of information once (the other one being your post - talk about being absent-minded), by the way. I'm just trying to make a sort of addendum saying something about under what conditions doing a proper evaluation is most useful. My point is, among other things, that it seems to me you gain less and less by spending a lot of time on the exposure setup as the ISO setting increases (because you hit other limiting factors first, so as to speak.) But you still want to avoid doing things like deliberately underexposing by closing down or reducing the exposure time, of course... Or that's not *really* my point - I'm getting sidetracked here. I was originally just trying to comment on the impact of the conclusions (as opposed to their validity or the explanations), and how they related to what people are doing with film - which is what I'm actually using ;-) > I take the tactic that I use the lowest ISO that I can all the time, > consistent with achieving the focus zone and exposure time desired, > since I can compress or expand the dynamic range in software with > little constraint in post processing. The lower ISO setting will > allow more data and thus greater flexibility in rendering. > Very sensible... I also think it would be interesting to use a strategy of not changing the setting at all, but just shoot at the native ISO - and correct everything in software, for a while, though. Maybe I ought to try it myself, except that I would then have to get a DSLR first... - Toralf -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net
