Film has a rather limited range of exposure...in most negative films, that range is somewhere in the 9 f-stop value.
Do you know how this compares with CCD or CMOS sensors?
That's the first time I've seen anyone claim 9-stop exposure latitude for color negative film, most I've typically read is 7. Digital sensors are commonly quoted to have about the same range as what is usually stated for slide, i.e. about 5 stops.
My apologies, but I'm short on time..you might Google Kodak's site, and the Rochester Institute of Technology.
I've noticed the wildly variant claims and don't think anybody's off their rocker. A lot of the issue is what folks are willing to tolerate...in other words, they're describing a latitude of acceptable exposure, not the ability of the film to produce a modulation transfer or grain characteristic, etc. If they were, I think the numbers would soar simply because you can get a chemical response with very little light, but of course, it would be nearly worthless aesthetically.
On the night end of things I (and a number of other photographers) are quite comfortable with underexposure in some areas and gross overexposure in others, particularly with lights. See the lights in my image above...in a standard photo, that sort of thing would be objectionable and certainly an issue of overexposure, but in the night image it works quite well. At least, to my way of thinking. There still is "some" detail in the hot spots...
Also, the film's performance is heavily dependent upon a number of factors including but not limited to the film's condition (sitting on a hot dashboard for a couple of days?), the condition of the chemicals used in processing, the processing itself (do not push or pull)...even the very nature of the film (there are more contrasty "saturative" films...I use Portra NC for this very reason when I shoot my aerials).
Also, scanning...PhotoMultiplier Tubes and multi-sample, enhanced bit depth scanning can produce a usable shadow detail well beyond a few stops down. I offer as an example:
This was multisampled 16x and there is extremely little, in fact, insignificant grain in the periphery, in spite of being shot on ASA800 and technically "underexposed".
Then, of course, there is the empirical acceptance of extreme contrast image which suggests that people may be willing to tolerate (inside of "artistic latitude") even more under- and over- exposure than film can produce significant detail in!
On the CCD/CMOS issue, We'd have to treat those more carefully. It's very complicated...for example, CCD sensitivity is weighted in the near-infrared...pretty heavily. Signal to noise in the blue is hostile, and then you cannibalize the green sensor to provide total luminance and run calculations to produce an image (Bayer Array antialiasing, et. al.)...I believe that the electronic sensors are actually more tolerant than film, and in spite of all the digital manipulations (curves, etc.) that cameras will apply, you can do the same to a scanned film image in Photoshop, with respect to generational loss...but it is much more tedious.
I'll try to dig up some data when I come up for air. I do recommend: dpreview.com, and Googling "Bayer Array Antialiasing."
Oh...and on the night stuff, my observations are that reciprocity can actually work both ways, depending on the colors being imaged (increase or decrease range). Very sloppy observation, I know...sorry.