For those of us who dont know (me), what is the difference between a digital photograph of a negative made with a very high quality camera, and a scanner that images the negative line by line?
The line-by-line method is also in very close proximity to the film. Typically this results in much higher 'dmax' which is a fancy way of saying it gets more dynamic range out of the scan than does a photograph made by a typical tiny camera sensor.
Dynamic range for anyone who doesn't know, is basically the ability to capture the darkest to the lightest detail present in scene. Not the sensitivity (ISO) but the overall dark-to-light ability to capture and retain detail. An image sensor with low dmax will tend to block up dark bits and blow out light bits rather than capture the detail present in the original.
I am guessing that this Kodak scanner uses a very low quality camera. Would this concept be acceptable if Kodak used a high quality camera? Or are scanners still better?
It's not a terrible image sensor, based on my eval of a Wolverine brand, which I would bet is very much similar if not identical (it seems all these units are based on the same internals, just look at the specs). A 14 mp digital camera type sensor - very small as most point-and-shoot digicams are.
The DSLR method of copying negatives that others have discussed here (and which I am keen to try) is essentially the same thing on a larger scale. The idea is simple - you mount your negative behind a sufficiently bright and diffuse light source, you focus on it with a DSLR using a tripod or other stable platform and a macro-type flat-field lens stopped way down to increase focus accuracy, and Bob's your uncle.
Why would the DSLR method work better than the Kodak scan-in-a-box method? Primarily, I think, due to the lens quality of the DSLR (potentially) and the sensor size and dynamic range.
When I tested my recently-purchased used Wolverine, I noted that it does an "OK" job on some things - mostly low-contrast images that just did not contain that much dynamic range.
To have a high-quality unit like the Kodak (or its many clones), one would have to make use of a higher-quality sensor, which would probably be larger. Not necessarily larger, but with a higher ability to capture dynamic range. This could also entail a higher-quality integral lens, although with a sufficiently bright light source and very small aperture, this might be overcome to some extent.
There is such a product on the market, I forget the name of it, which is basically a framework for your negatives/slide and your own DSLR to do just what we're discussing. It's quite expensive as I recall. I am planning on trying a lower-budget DIY test myself. I am using an old enlarger with one of my DSLR cameras mounted where the enlarger head was, and the enlarger head turned upside down where the old photo paper base was located. We'll see how it goes.
I will also note that it appears from my own results as well as some published tests that these units seem to crop the 'scans' they make, which might be objectionable to some.
My plan for the moment with my Wolverine is to use it for quick scans of B&W negatives as I go through my archives, to see which of them I might like to do a higher-quality scan on with my Konica-Minolta ScanDual IV.