Measurements

The contrast between the characters and the papyrus background varies depending on the wavelentgh of light. By contrast we mean the ratio between the reflectivities within the chosen spectral band. In blue light, black text upon red background would produce black text and dark background, characters being difficult to read. In red light, the background would be bright while the characters would remain dark, meaning much better contrast. If there is an optimal spectral band for the papyrus writings, that should be used.

The spectra of the papyrus background and lampblack (carbon black) were measured using Monolight Instrument's [xx]. Direct measurement of the ink with the instruments at hand wouldn't have been reliable due to the thinness of the characters; parts from the background would inevitably have been included in the measurements. Insteadt, we used self-produced lampblack substitute. Lampblack can be produced in small quantities simply by the ancient method of burning a candle and collecting the carbon from the flame onto a surface, in this case, onto a surface of a small 2x5 cm microscope glass plate. Measuring the spectra from the small plate is quite easy. The ink in carbonized papyrus has only its lampblack pigment left, so our substitute should give quite reliable results.

Figure [6] shows the relative reflectances of the papyrus background and the lampblack within 400-1100nm, measured using a silicon detector. A germanium detector was used to measure the reflectances in the 800-1500nm band in figure [7]. The raw spectra were normalized to one at 400nm (fig. [6]) and 1000nm (fig. [7]). The differences in absolute intensity values were big, the reflected intensity of the background being tenfold compared to the lampblack. The absolute values are not presented here, as the relative reflectance is a better measure of contrast differences as a function of wavelength.

Because the reflectance of lampblack is nearly zero in wavelengths over 1000nm, measurement of the relative reflectance becomes very noisy. The Ge detector used in that area is also less sensitive than the Si-detector, so the overall noise is larger, although the spectra of figure [7] were gathered over a longer period of time.

The overall impression of the relative reflectance at wavelengths 400-1600nm is clear: at longer wavelengths, the papyrus background reflects relatively more than the script. Therefore, the contrast should be better at longer wavelengths. We should be able detect the effect at 750nm with red-extended films, and it should be clearly visible around 1000nm, at the limits of our CCD camera. These observations concern only the measured bandwidth. The absorption of IR radiation by the glass plate will cause noteworthy alteration to spectra in longer than 2000nm wavelengths.


Figure 5: Reflectances at 400-1100nm, Si-detector


Figure 6: Reflectances at 1000-1600nm, Ge-detector


Back to Previous Work Back to Contents Next: Recording
Antti Nurminen, 34044T, andy@cs.hut.fi