Let's do a quick review. Remember that resolution is measured as pixels per inch which is the same as dots per inch or dpi. Resolution is also relative. You can't really talk about resolution without also considering the medium on which the image appears. For our purposes there are two resolution that we need to worry about: the resolution at which the negative is scanned and the resolution of any reprints we want to make from that image.

A 35mm frame is 24mm x 36mm (or it's supposed to be anyway). So a 2000 dpi scan should result in an image that is 1890 x 2835 pixels. Most labs that make prints from digital images will tell you that a print will only look good if its resolution is least 200 dpi, and many recommend at least 300 dpi. If this image (1890 x 2835) is printed at 4 x 6 inches the result would be about 472 dpi, which is great. At 8 x 10 inches it would be about 236 dpi, or marginally acceptable.

Here is a list that compares the three common scanning resolutions and what enlargements they are able to support:

- 2000 dpi: 1890 x 2835, good to 5 x 7, marginal to 8 x 10
- 3000 dpi: 2835 x 4252, good to 8 x 10, marginal to 11 x 17
- 4000 dpi: 3780 x 5669, good to 11 x 17, marginal to 16 x 20

If you understood all that then you need not read any further. If you're still scratching your head then let me try to clear the confusion. Remember that resolution is a relative measure. It depends on the size of the physical medium. The size of a jpeg image is always best expressed as actual pixels just for this reason. Imagine you have a jpeg that is 1500 pixels wide by 1200 pixels high. If you print this image at a size of 5 inches by 4 inches it will be printed at a resolution of 300 pixels per inch (dpi). That's because 1500 pixels divided by 5 inches is 300. Now if you take that same image and print it at a size of 10 inches by 8 inches, the resulting resolution will only be 150 dpi (1500 divided by 10). The same image printd at different sizes yields different resolutions.

Now let's go back to the negatives. When a negative is scanned the equipment picks up information at a predefined density, or resolution. If a negative is scanned at 2000 dpi, that means it measures levels of light and color at 2000 different, equally spaced, distinct points for every inch of the frame. A frame is supposed to be 24mm high by 36 mm wide, although some cameras might not get that exactly correct. Convert those millimeters to inches and then multiply by the density and you get 1890 pixels high and 2835 pixels wide. Once that scan is made and converted in to a jpeg, you cannot add information to it. Sure you can make it larger but you would have to extrapolate (or in other words "guess") the extra data. So for all practical purposes that image is 1890 x 2835 and no larger. Take that image and put it on a sheet that is 4 inches by 6 inches and it will be at a resolution of 472. You can get this number using simple math: 1890 pixels divided by 4 inches (or 2835 pixels divided by 6 inches). Well, actually its 472.5, but I rounded down.

It gets a little more interesting when you print that same image on a 5 x 7. The reason is that the frame doesn't quite fit on a 5 x 7. If you let the 2835 horizontal pixels exactly fill the 7 inches, then the 1890 vertical pixels will only cover 4.6 inches and you will end up with ugly white stripes at the top and bottom of the print. So the image has to be cropped on the sides to make the vertical pixels fit. The highest resolution you can achieve, then, is 1890 divided by 5, or 378. Got the hang of that now?

When I start scanning my collection of negatives I will probably stick with 2000 dpi. First, I need to hold the costs down. Second, nearly any reprints I will want will be 4 x 6. I am also planning on keeping my negatives, so on the rare occasion when I feel I want an 8 x 10 or larger I can always go back to the original negative and have it printed from the source.