How big are the files?

Just look at our files they're THIS BIG! Just imagine the quality - so how big are our files?

If you're looking for a measure of image quality checking the size of data files is a very poor guide, let me try to explain. The basic variables that determine the quality factor are the number of bits / dots per inch, the number of bits used to express the colour of each bit, well that's about it. The third factor that comes into this is the file type - typically jpg or TIFF.

I hope it's obvious that when scanning a slide, scanning a photo or scanning a negative the higher the number of bits per inch the better the quality of the image. Well, to a point anyway. If we scan a print a 600 dpi image is inherently more detailed than a 300 dpi scan. However there does come a point at which increasing the dpi rating achieves nothing more, you're just taking a more detailed picture of the grain or paper fibres. From tests we've done, using the scanning devices we have here, the highest scan quality for a photo is 600 dpi and a 35mm slide / negative is 4,000 dpi.

Second, there is the question of the number of bits used to measure the colour of the dot the scanner sees. You need (for a colour image) a red value, a green value and a blue value. A 24 bit scan uses 8 bits for each value - that yields 256 different points between darkest red / green / blue and lightest value. The next step up from that would be 16 bits per colour per point, 48 bits. However this will generate a much bigger data file and you'll need software which can display this additional degree of subtlety.

So away goes the scanner, it looks at the number of points you've specified and gives them a colour value. The scanner will then output the result of all those measurements as a data file - which is where the choice of jpg or TIFF comes into play. The key difference is that a TIFF file is uncompressed (and so very large) while a jpg is compressed (and smaller). Somewhere along the way a programmer has written code which looks at all those point values and decides how a file can be compressed, with the goal of achieving a smaller file with minimal loss of quality.

Whilst the specification behind jpg compression is open to all programmers and software suppliers, the results of their work varies. From our experience there are noticeable differences between the efforts of Kodak, Microsoft, Adobe and Apple. And there are differences over time, and between products. If you take a TIFF file and save it as a jpg using an early version of Photoshop, a recent version and Photoshop Elements, you get a different result (for the same degree of requested compression). Today's version of Aperture creates smaller high quality files than did the first version of Aperture.

Put most plainly, a 4000 dpi 24 bit jpg saved across these products gives files of different sizes. Which is the "right" one? Take your pick - mine would be for the latest versions of Photoshop and Aperture. Should you necessarily conclude that a scan made using the same scanner but output via one rather than the other is better just because the file is bigger? I don't think so, so don't be mislead by this single factor.