Photography

Photographing carbonized papyri fragments is almost a science of its own. The most important variables affecting the production from the original to paper copies are:

To search the best ways to produce photographed pictures, we went through several variations. We concentrated on illumination conditions, film material, filters and exposure times.

Illumination

Optimal lightning conditions are essential for succesfull photography, especially in our case of an extremely dark object. A set of variables and restrictions define the limits and possibilities of illumination:

There is one limitation compared to direct physical study: the normal of the plate should point to the camera; i.e. if the plate is at an angle with it, some of the photographed areas may not be in focus. With a studio camera, the plate can be photographed also in a sligthly tilted position. When studied with the naked eye, the plate can be held at hand and examined from almost any angle. The human eye is also somewhat capable to reduce, or omit, the disturbing direct background reflections.

Experience shows that the readability of the writing seems to be best when the reflection of the light source is just on top of the text. Further tests emphasis this observation, and it seems that the readability varies according to the angular distance from the reflection line. When kept at a straight angle, the light source would be directly behind (or in front of) the viewer. Naturally, the reflection itself cannot be allowed to be in the picture. The optimal angle for a single light source can be found by moving the source so that the reflection just goes out of the picture boundary. With several light sources, symmetrically set at the optimal angle, a nearly uniform light field can be created. A large, ring-shaped light source could be nearly optimal. Use of diffuse filters or plane reflectors may further help to hide the reflections and improve the uniformity.

The contrast between the script and the background depends directly on the brightness of the lamps. The more light, the better the picture. Unluckily the black fragments absorb light very efficiently, and thereby heat up quickly. Heat erodes the pieces, so one must not expose the fragments to intense light for too long.

Tungsten light bulbs emit a wide range of electromagnetic radiation. Infrared films as Konica 750 or Kodak HiE are actually not sensitive to infrared radiation (from about 1000nm to 12000nm). Their sensitivity range has been extended from the normal visible light (400-600nm) to near-infrared, around 750nm. The typical spectrum of a tungsten bulb is a continuous, roughly bell-shaped curve, extending from below 400nm to over 1000nm. The spectrum of halogen light sources is weighted more in the shorter wavelength band, and therefore seems not as usable. Fluorescent light consists only of a few wavelengths, as shown in picture [9]. Because the fluorescent light sources produce a lot of light with very little heating, they cannot be completely omitted, but particularly for red-extended films, they are utterly unsuitable. In ultraviolet light the writing vanishes completely. So, this verifies the dependency between the relative reflectance and wavelength, presented in figures [5] and [6], in the shorter wavelength band. The infrared radiators, on the other hand, produce too long wavelengths, beyond the sensitivities of red-extended films and transparency of the glass.

The Photographs

The following film materials were tested:

Many other companies, like Agfa, manufacture similar film materials. These films were chosen to compare different film types, not companies. Technical Pan is a high-contrast B&W negative, T-Max a standard, high-resolution B&W negative, HiE and K750 are red-extended B&W negatives and EPT is a color film.

The pictures were taken with an SMC-Pentax M camera, using a Pentax 1:2 85mm lens with 12mm macro adapter. The distance between the film and the papyri fragments was 50cm. Three tungsten lamps was used as light sources symmetrically around the plate, near the camera. With 160 ASA as supposed sensitivity of film and using 1:2 lens, our Gossen Lunasix III gave 1/60 for exposure time and 8 for aperture, measured from a standard grey card. For black, carbonized papyri, strong overexposure is needed. Table 2 lists the exposure times and apeertures we used in the best photographs along with the film sensitivities. Without the macro adapter, the exposure times could have been halved. The ASA value is a rough estimate of the overall sensitivity, from the manufacturer. The IR filter used was equal to Kodak's Wratten 89B. 

	Film                    ASA  Aperture    Exposure Time (s)
	----------------------------------------------------------
	Konica 750              32      8             1
	Konica 750, IR filter   16      8             2
	Kodak HiE               200     8             1/30
	Kodak HiE, IR filter    125     8             1/15
	Kodak T-Max 100         100     8             1/8
	Kodak Technical Pan     100     8             1/8

Table 1: Exposure values

We developed the negatives with Kodak's D76 standard developer, shaking the developer container for 5 seconds every 30 seconds. The temperature of the developer was 20.0° C. The development times for the black&white negatives are given in table 2. 

	Film                 Time (minutes)
	-----------------------------------
	Konica 750           7
	Kodak HiE            11
	Kodak T-Max 100      (check)
	Kodak TP             12

Table 2: Development times for negatives with D76, 20.0° C

The paper copies were made with an Axomat projector with EL-Nikkor 50mm 1:2.8 lens, aperture 8. We tried contrast grades of 3, 4 and 5 (Agfa Brovira-Speed BH-310 papers RC 3,4,5). Number 3 is the standard paper, and 4 and 5 have higher contrast. The developer was Kodak's --, kept at 20.0° C. The papers copies were adhered for 15 minutes, and then kept under cool, flowing water for an hour.

Technical Pan is a very high-contrast film, and with grade 5 paper the resulting image is almost binary, containing only black and white, no shades of grey. If it were possible to separate the writing from the background only by thresholding with a certain grey shade, it could be done with this combination of high-contrast film and paper. Unluckily parts of the background are as dark as the characters, and the extreme contrast causes the dark background to merge with the characters, or parts of characters to vanish, depending on the exposure. The writing is quite readable from the best photographs, but it is also clear that much of the details of the original fragments is lost.

The color information of the EPT 160 film didn't seem to be of much help. There is really not so much color in the original fragments, and the measurements of the spectra show that there is no single peak in the contrast in any bandwidth (color). Still, the idea of using color should not be completely discarded, because in some of the carbonized fragments the background is still yellowish, and that information could be of use. One can use filters with black and white film to restrict the bandwidth between certain wavelengths, but from color film, one can roughly filter out unneeded wavelengths afterwards.

To produce a smooth, low-contrast and high-resolution grey scale image one should choose Tmax film and use a grade 2 or 1 paper. The resulting contrast is then so low (with grade 3 paper the outcome is near the real contrast - a very low one) that large magnification is needed to increase the readability to an acceptable level. Enhanhing the Tmax's contrast with a grade 4 or 5 paper gave quite good results. The resolution was good, background details were still visible and the contrast between the writing and background was enhanced.

Back to Tests Back to Contents Next: Scanning Photographs
Antti Nurminen, 34044T, andy@cs.hut.fi