Clive, the only thing that matters is the total number of pixels in the =20=
final image. The size or resolution of the sensor isn't important in =20 itself, only in how it relates to the physical size of the camera and =20=
lens. If you put a giant sensor in a view camera that had the same =20 number of pixels as a small one in a DSLR, you'd get the same result =20 (lenses, etc. being equal). Think of it being a sheet of graph paper, =20=
and you're filling in the same number of little squares - it doesn't =20 matter how how big the squares are, as long as there is the same number =20=
of them - the end result is always a mosaic of the same number of =20 squares in the same pattern. The fact that the Contax N Digital has its =20=
sensor spread over a slightly larger area is actually a very slight =20 advantage, compared to a similar 35mm-sized camera - a smaller sensor =20=
with the same number of pixels would essentially be cropping into and =20=
blowing up the center area of the lens slightly.
The N Digital also has a huge advantage because it doesn't crop off the =20=
ends of the image area - most other DSLRs effectively give each of your =20=
lenses a slightly longer focal length.
There has been a lot of talk about 250 dpi images for inkjet prints, =20 but 300 dpi is standard for commercial publishing. Images from DSLRs =20 used larger than about 8 x 10 or so start to run into problems - there =20=
are not enough original pixels in the image to go larger. This is true =20=
even if you're using a full-frame image. If you crop, the problem is =20 worse because you're cutting off pixels and lowering the resolution of =20=
the image at its finished size, but you still need to have 300 dpi in =20=
order for the color separator/magazine/etc. to accept your file.
You need to increase the resolution in some artificial way, perhaps =20 using one of the fractal-based utilities, but you're not adding real =20 data to the file. Stock photo agencies require film originals because =20=
you can scan them at resolutions high enough to print at larger sizes. =20=
As I had pointed out earlier, even a consumer-type scanner doing a 4000 =20=
dpi scan of a slide gives you almost 4 times as many pixels as the N =20 Digital does.
It's been my experience that clients will always underestimate how they =20=
will use a photo. All they see is their immediate need. Part of the =20 reason I still shoot on film, exclusively, is because this always gives =20=
me the option of getting bigger files if I ever need them. The client =20=
who insisted that they only use a photo for a tiny magazine ad will =20 come back six months later and want to use the same photo 8 feet tall =20=
on a trade show display - I can always re-scan the slide if this =20 happens.
Marc, your comments have been extremely interesting. However, several =20=
ad agencies that I've worked with enthusiastically embraced digital at =20=
first, but are now backing off and returning to film. I myself feel the =20=
pull of the tremendous convenience that digital would offer, as well as =20=
quick turnarounds (no waiting for processing! no scanning! no dust!) =20 and the ability to have the client view and approve your images =20 on-screen during a shoot. It would also be much cheaper in the long run =20=
- I buy and pay for the processing of 200-300 rolls of film in a year =20=
(which is a very modest number compared to what a lot of other =20 photographers shoot), but this would pay for an N Digital pretty =20 quickly.
A few clients have been surprised that I shoot on film, since I'm also =20=
an illustrator and graphic designer who's been doing computer graphics =20=
since the 1970s. Most of my professional photography is for projects =20 that I'm also the graphic designer for, and I've been involved in =20 digital prepress work since the very early days in the late 1980s and =20=
early 1990s. Until I switched to Macs in about 1987, I was working on =20=
computer graphics systems that cost a half million dollars per =20 workstation. I started using Photoshop with version 1 in about 1991. =20 So, people who know my background know I'm hardly a Luddite. Right now, =20=
I believe that I can provide a better file by shooting on film, and =20 will have more flexibility with how the image is used in the future =20 since I can always re-scan it again later. Since I seldom work with =20 files smaller than about 50 MB, the output of a DSLR seems pathetically =20=
small to me.
However, I'm not completely close-minded about this, and I do feel the =20=
seductive pull of digital, as I mentioned above. I have been surprised =20=
at some large photo reproductions that I know came from DSLRs - for the =20=
small resolutions of the original files, I'm astounded at the quality.
In the end, few clients/publications really care where the original =20 file came from, as long as they have a good, printable digital file as =20=
the finished product.
Another thing is that I save all my outtakes from a shoot - this means =20=
simply keeping all the slides. With digital, I think there would be a =20=
tendency to throw away all but one or two "keeper" shots from a shoot. =20=
I am surprised how often I have been able to find an image I need by =20 looking at unused shots that did not suit my client's immediate need at =20=
the time, but were still perfectly good shots. This is especially true =20=
with people shots - you'll often find slightly different poses or =20 expressions that you didn't use before. I also tend to shoot 'what the =20=
hell' shots if I'm on location and see something I feel the client =20 might want in the future, even if they have no interest in it at the =20 time. As I mentioned, many clients are so focused on their immediate =20 needs and don't see the possible uses of their images in the future. I =20=
figure as long as I've dragged myself and my gear there and set up, it =20=
takes little effort to do a few more shots.
Marc, I don't understand this comment: "NO ONE scans film to sizes of =20=
the actual application as it would cost a fortune per image and take =20 forever." I always make a scan at the resolution I need for a specific =20=
use of a transparency, and I don't understand what else would make =20 sense - perhaps I'm misunderstanding what you're saying here?
- Paul