You are here

Issues for digital artifact photo archiving [Part 1]

I'm becoming involved in a historical artifact photography project using digital images. As such, I need to develop some materials to use in describing collections of digital photos of artifacts. Here's my first cut. I'm trying an experiment here by breaking this into a multi-part series. I originally intended two parts, but it has gotten longer than this...

Background: it is easy to understand how a digital camera works. Light enters the camera, and hits a rectangular tile mosaic of microscopic sensors. Typically, each sensor is square. By measuring the light values on the mosaic of squares, one gets a mosaic of picture elements called "pixels", each one consisting of a light intensity level. Color digital photography is a bit more complicated: colored filters are put in front of some of the sensors so that they measure just red, green, or blue light.

The "resolution" of a digital image is the number of pixels horizontally and vertically over a given distance. Normally, the horizontal and vertical resolution are the same. Much confusion results from the fact that the term resolution is often used, incorrectly, to denote the count of sensors across the whole mosaic in the horizontal and vertical directions: these two numbers are typically different. Thus, photographers will talk (correctly) about "75dpi" (dots per inch, aka pixels per inch) images; they will also talk (incorrectly) about a camera having "2048x1360" resolution, or even of "2 megapixel resolution". This last is simply the total number of sensors on the mosaic, computed by multiplying the horizontal and vertical sensor count of the rectangular mosaic.

True resolution computation for an image requires knowing two things: the physical size of the image, and the number of pixels vertically and horizontally in the image. However, it turns out that the physical size of a digital image is not fixed; it can be easily scaled to a variety of sizes. More importantly, from an artifact photo point of view, the resolution of the image is not the most important thing. Instead, proper emphasis should be placed on the effective resolution on the artifact. One would like to have as many pixels covering the object itself as is possible; this makes the photo much more useful.

In part 2 of this series I will complete the discussion of digital images.


[OK let's see how my email address is handled. Of course this is in near-anonymous passer-by mode not my Rob account for which I need to locate the password, but that's another story. I'll try Filtered HTML this time.]

Good stuff Bart, I'm looking forward to Part 2. I'd never thought about resolution in terms of pixels/distance before, except when printing. That's when 75 dpi seems to come up the most, when you've got a fixed ratio for the printer's capability. When taking the picture as you point out it's a different story, since you can focus near or far and thereby change the distance divisor.

As I type it occurs to me, another way I perhaps apply fuzzy thinking to this topic, monitor resolution. There's a huge difference between 1024x768 on a 17" monitor vs. the jumbotron you encouraged me to pursue at work (and thanks for that too, worked just fine didn't it!). Somewhere in this topic is the well-known difference/equivalance between a 21" CRT and a 19" LCD panel. Pitch?

Shrug anyway enjoyed the entry, good luck with the project, oh yeah that's the last thing: do you need another project?


First, I'd like to have left this comment on your latest blog entry :-), but since I can't... You might mention this page on your blog; it contains a lot of ideas for how to get a stuck ignition key out of a lock (for less than $90). Sorry to hear the sad news, especially too late to help.

Second, I think the 21" CRT is less equivalent to a 19" LCD than most believe. The inarguably correct key notion is that the pixels on a CRT are somewhat fuzzy and overlapping, so the claimed "resolution" is actually kind of bogus. Further, CRTs tend to smear signals badly horizontally, tend to get out of focus and "deconverged" (a whole 'nother topic), etc. But my experience is that a properly-adjusted high-quality 21" CRT is actually quite a bit easier to read than a 19" LCD. Resolution is great, but I actually am not resolution-limited on my 21" 1600x1200 display: I am limited by my poor vision and the number of characters per inch I can tolerate.

Finally no, I don't need another project. Smile But it's a worthy cause, and likely to lead to grant funding down the road I can use to pay my starving students to do interesting work.