> >>>>> Negative film gives the lab about a stop of underexposure and about 3 >>>>> stops >>>>> of over exposure before a good print can't be pulled from it, a jpeg has >>>>> about half that latitude. >>>>> >>>>> >>>>> >>>>> >>>> Doesn't this mean that what we ought to keep in mind is >>>> >>>> It is better to overexpose than underexpose >>>> >>>> for digital *and negative film*? (But maybe some of you have been >>>> thinking that way all along?) >>>> >>>> >>>> >>>> >>> That's correct, I rate Superia 400 at 320; others go even further than >>> that. >>> >>> >>> >>> >> Ah. Yes. There you have it. I heard people recommend this a number of >> times, of course, only I didn't think of it in this context... >> >> But, but, isn't a similar trick available for digital? Can't you just >> reduce the gain a bit and try to get an exposure "in the middle" (with >> the same exposure) rather than aiming for an exposure "to the right" >> (with a somewhat higher gain setting)? >> >> - Toralf >> >> >> >> > > No, because of the way digital sensors work(See Godders' superb > explanation upthread), you get more usable information with an exposure > 'to the right' on Digital. So you almost always want to get right in > below the clipping point with digital (At least when shooting RAW). > I think you missed my point. As far as I know, a digital sensor's sensitivity to light is constant. The camera will usually have a "sensitivity" or "ISO" setting, but what that really adjusts, is a gain applied to the signal after it leaves the sensor. Furthermore, I'm quite sure it is correct that you get the best data quality if the exposure etc. is set up so that the sensor is nearly saturated, but not quite. But, that is not the same thing as saying that the digital data, or even the analog signal after the gain stage needs to be that way. The way I understand things, the camera's histrogram function shows the digital data - not the actual sensor output - and I don't necessarily need to keep that "to the right". Rather than "overexposing" the digital data, then making it darker in Photoshop or whatever, you can do the "darkening" already in the gain stage simply by reducing the ISO value.
Differently put, I'm sure it is a good idea to keep the *sensor's* histogram (if you can make one) as far to the right as possible, but rather than moving the *digital data's histogram* to get what you want, you can just change the ISO setting of the camera. Essentially it is just a question of whether you want to do the exposure correction (if you can call it that) on the digital data or in the analogue world. I would intuitively choose the latter approach in most cases, but there could of course be some specific reason to prefer the former... Even more differently put, isn't what people are saying here just a rather convoluted version of "you get less noise with a lower ISO setting"? I don't know if I'm making it entirely clear what I'm talking about here... I'm writing this quickly during some natural breaks at work, so I don't have time just now to think a lot about how to explain things ;-) (Seems like I've posted a lot here the past few days. This have to do with the fact that I'm doing some testing that involves a certain amount of waiting...) - Toralf -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net
