As I previously mentioned at least twice, at sea level, the earth's atmosphere blocks only about 2% of visible light.
No, I'm afraid that's not the right number. But let's let
@Vixen tell us the right one and then we'll figure out how many f-stops that works out to.
While Vixen is counting on her fingers, I'll follow up on yesterday's discussion. The reason some astrophotographers want slow film is that slow film means smaller silver halide crystals. Smaller crystals mean tighter grain and more resolution. This is an advantage to some, but only if you're shooting through a substantially expensive telescope. If you're just shooting through a typical SLR lens, the grain advantage is swallowed up by the shortcomings in the optics.
Obviously you need a longer exposure for slower film. One bad side effect with that is sky fog. Even with clear air and good seeing, a long exposure will still give you some exposure in the non-star parts of your photo. This may make it hard to see very faint stars in your photo. For this reason some people use very high ISO film, as the expected subject brightness falls closer to the expected film response.
But the other problem with long exposures is reciprocity breakdown. Reciprocity is a subject we alluded to before when we were expressing star brightness in terms of f-stops. For those properties of photography that are linear, we know we can compensate for a change in one with a reciprocal change in another. So if we switch from ISO 100 to ISO 200 film (a linear doubling of its sensitivity) then for ordinary photography we can compensate by halving the exposure time: say, going from 1/250 second to 1/125 second—also a linear change. The factor by which we multiplied the film sensitivity is the same factor we use to divide the exposure time—the reciprocal.
But this relationship breaks down at long shutter speeds. That is, it stops working that way. It works only for short shutter speeds.
Here's why. Photographic film is a matrix of silver halide crystals arranged in as close to an even distribution as possible in as thin a plane as possible to approximate a 2D surface. When a photon strikes the crystal, it causes a chemical change. The actual reaction happens at the molecular level. The crystal is composed of many molecules. So at the crystal level it takes many photon hits for the crystal to fully decompose into metallic silver. I'm simplifying the process slightly to get to my main point. We can talk later about the nucleation of the latent image and all that.
To further simplify, imagine that film is a very large wall set up with cork to be a dartboard. A number of ballots are fastened in a regular pattern on the wall. You stand an appropriate distance away and throw darts. You're aiming at a target, but as with most of us your skill at darts is only average. The darts are photons and your inexpert aim is the slight scattering effect of passing through a lens. The balloons are the silver halide crystals. When you hit one with a dart, it pops.
Now your first dart is almost certain to hit
some balloon and pop it. Your next dart has a slightly less chance, since one ballon is already popped. After 20 darts all hopefully aimed at the same spot on the dartboard, you're hitting the board but not any more balloons. In order for a dart to hit a new balloon—and thus change the state of the dartboard in terms of popped balloons—it has to be sadly and badly aimed and go way off target. After 50 darts you're simply not hitting the balloons any more. After a while it takes more and more darts to have any effect.
This is why reciprocity only works for a small number of
darts photons = a short exposure. The lens can keep directing photons all it wants to that same spot on the film, but if all the
balloons are popped silver valid crystals are reacted, the light has no more effect on the film.
To skip to the end, what this means is that stellar photography using low ISO film requires
ridiculously long exposures—hours in some cases. You have to allow enough photons for their natural distribution to spread to outlying areas around the primary star exposure.