resolution measurements are always made using a loss of source contrast in the output. If you start out with high contrast source you will get a higher resolution measurement to a fixed output contrast than if you start out with a low contrast source. this is going to be true with lenses as well as film and sensors. So the "loss of resolution" with low contrast sources isnt what it sounds like, it's just that you reach a fixed output contrast at a lower resolution number when the source doesn't have much starting contrast to begin with.
JCO -----Original Message----- From: William Robb [mailto:[EMAIL PROTECTED] Sent: Tuesday, November 02, 2004 6:09 PM To: [EMAIL PROTECTED] Subject: Re: USAF target and resolution tests ----- Original Message ----- From: "Jens Bladt" Subject: Re: USAF target and resolution tests > Thanks. > I don't undserstand your note about a pictorial application > (1000:1), > William? When you measure resolution, what you are doing is determining how fine a detail the optical system can see. Test targets are the easiest objects for a system to resolve well, because they are, essentially, knife edges of black and white: Lots of contrast to resolve. Unfortunately, it is very rare for a pictorial subject to have that much contrast between two adjacent areas. This is why film and lens makers use two scales for determining resolution. One scale is 1000:1, where the two adjacent tonal values are at the extreme of possible densitys (full black and full white). The other scale, and the one that makes more sense (at least to me) to use as a performance indicator is the lower contrast scale (1.6:1). This scale uses tonal values that are much closer together in density (1 stop, I do believe), which more closely resembles the sort of tonal transitions one in likely to encounter in real life picture taking. The other thing that test targets generally don't account for is colour resovling ability. For example, a particular peice of equipment or emulsion may well resolve 100 lmm with a black and white test target (USAF standard), but we don't know how well it will resolve the difference btween a red barn and a blue sky, for example. > > All I really want is to be able to jugde my lenses - before or > after > purshase...:-). Since I mainly shoot digital now, I want to use my > *ist D > for thesting the lenses. And I want to be able to choose the right > lens for > the right job (of cource it would be a lot better to just have > access to a > wide range of state of the art/pro lenses for everything :-). Maybe it's because I spent so many years in the game, but I just don't care about this sort of thing any more. I just go out and take pictures and I don't go looking for trouble. If the film/ lens/ digital sensor/ whatever makes pictures that I find pleasing, then I am happy with the equipment. I can't be bothered with trying to squeeze the last possible drop of performance out of something. That's too much like work. If small format, be it digital or film, won't resolve enough for a situation, I pull out a larger camera. I just don't think the difference between doing it right, and going to extraordinary measures makes a whole lot of difference. > > Finally: > 90 lpm is not bad at all, is it. Many lenses can't reslove that > much on > film. I found one film on the Fuji website, I think it was a 50iso slide film, that was 88lmm at 1.6:1. For most film, 45lmm-50lmm is closer to what you will see in real life. I asked this once before, and I am pretty sure I got an answer, but I will ask it again, as it makes a bit of a point. The question was: Do digital sensors lose resolution when scene contrast drops the same way that film does? The answer I recall getting was no, they don't. If this is really the case, then a digital camera resolving 45lmm (Rob says the istD is capable of 46) is actually very close to what film is doing in pictorial applications. Discussion anyone? William Robb
