How the Images were Digitized

The images were digitized using a semi-permanent copy-stand setup consisting of a Sony DXC-930P 3-CCD video camera feeding into a frame buffer.

Software written in-house allows the raw data from the framebuffer to be transmitted to a machine running Linux (we often use a laptop), where the image is written to disk in JPEG format, which can accommodate 16.7 million colours. (The same setup works well with greyscale images.)

The resultant image is about 1,600,000 pixels - that is, about four times the resolution attainable from a digitized "ordinary PAL video" signal. At reasonable compression (85%), it normally occupies just under 500KB on disk.

The Sony DXC-930P is equipped with a closeup lens, meaning that images as small as a small human fingernail can be digitized full-frame. Equally, the signal can be passed to a laserdisk recorder, which we find to be a convenient storage device for large quantities of video-sized images (the recordable platter holds 36,500 images).

As some of the images attest, the difficulties with photographing such small objects (which I did with tools that were positively al fresco for trying to hold the objects still) include accurate placement, and depth of field.