Note: This article was published prior to the last total solar eclipse on March 29, 2006.
Since the beginnings of astrophotography, countless images of total solar eclipses have been taken at great expense that contain much information about the corona of the sun. What is the best way to salvage this treasure?
There are three main reasons why photographing total eclipses of the sun is among the most difficult tasks of astrophotography. First, there is the enormous contrast differential that makes it impossible to capture the entire phenomenon in a single shot. This is because photographing a total solar eclipse requires a brightness jump of 1:1,000,000, which no camera, classical or digital, can achieve. Second, there is little room for experimentation. If something goes wrong, it can take years before another experiment can be attempted. And third, processing images of total solar eclipses is a complex and time-consuming procedure for which highly specialized computer programs need to be developed.
Countless astronomers will have to make a decision before the eclipse on March 29, 2006: do I use film, or do I go digital? The answer is anything but simple. These days, most astro images are taken using digital cameras. Compared to classical film, CCD and CMOS chips have a variety of well-known advantages, but also a few disadvantages that may come into play when photographing total solar eclipses.
Classical or digital photography?
The most important advantage of classical photography is the extremely high dynamic range of modern negative films. (Slide films have a higher contrast and lower dynamic range, which is why they are less suitable for eclipse photography.) Modern negative films are capable of being vastly overexposed without reaching maximum image density. Even if the contrast in an overexposed image is very low, it is not zero. The problem with classical films has to do with underexposure. If the quantity of light does not exceed a certain threshold value, the film behaves as if it were not exposed at all.
The most important advantage of digital photography is how easy it is to use. It doesn't have to be chemically developed or digitally recorded. From a scientific perspective, the linearity of the detector elements is a very important characteristic. On the other hand, digital cameras have a low dynamic range and serious problems with overexposure. The chip's output signal is linear up to the point of saturation—after that, it contains no further information about light quantity. As a result, the overexposed portion of an image is simply a white, bleached-out surface.
This is why there is no clear winner in the battle between classical and digital photography. Neither digital nor classical photography can deal effectively with such extremely high contrasts, and as a result no single image—neither digital nor on film—can capture all of the phenomena that the human eye takes in during a total solar eclipse. The strongest argument in favor of digital photography is its ease of use, whereas negative film is unbeatable when it comes to representing phenomena that are highly contrastive (e.g., the pearl necklace phenomenon), and no parts of overexposed images are ever completely bleached out.
The disadvantages of both photographic methods can be overcome by overlaying several digital or digitalized exposures thereby creating a single image. But even if the digital image that is obtained were able to reproduce the brightness and color of the phenomenon correctly in all its parts, without subsequent mathematical processing, the results would still be far inferior to the overall impression made by the solar eclipse on the observing eye. Why is this?
Photography and visual observation
The principle of photography is completely different from that of human vision. Both film and digital detector register the absolute brightness of all image elements. By contrast, the human eye performs a differential analysis of an image—it can only compare the brightness of an image element with that of its surroundings; the structure of which, in turn, is largely dependent on the characteristics of the overall image. The image that we "see" is not the current brightness distribution that is projected onto the retina and transmitted to the brain—it is a virtual model of the external world that is constantly being updated based on new differential measurements.
Another important difference is the capacity to adapt. In an image registered on film or on a CCD chip, the technical parameters (focus, exposure time, sensitivity, color characteristics, etc.) are firmly established for each image element. By contrast, the eye is able to master the extreme differences in brightness that occur during a total solar eclipse, while at the same time perceiving the most finely nuanced local fluctuations in brightness within the corona. Even if a photographic image is taken under optimal conditions, these same details will remain invisible. No screen or projector can reproduce brightness contrasts in the range of 1:1,000,000. The brightness range that can be reproduced on paper is even narrower.
Mathematical methods of visualization
There is only one way to obtain an image that renders the experience of a total solar eclipse on the observing human eye, and that is to create an image by mathematical modeling that exhibits a representable brightness contrast on paper or monitor while at the same time doing justice to the visual experience. Such models are based on so-called adaptive filters that simulate human vision numerically.
Since the beginning of astrophotography at about the turn of the last century, a vast quantity of photographic material has with great effort been collected about total solar eclipses. These photographs contain valuable information that has not yet been made accessible about phenomena in the solar corona. Modern computers with their fast processors and almost unlimited storage space now make it possible to use these newly developed numerical processes to analyze this archival material, and to produce images of the corona whose quality would have been unimaginable a few years ago when the images were taken.
In 2002, the author of this article began the MMV Project ("Mathematical Methods of Visualization of Solar Corona") at the Brno University of Technology in the Czech Republic, with the goal of developing new and better numeric methods for evaluating the images. What is involved is the highly precise alignment of the individual images and their representation with the aim of coming as close as possible to the actual visual experience. Some of the results of this effort illustrate this article.
All professionals and amateurs who have good images of total solar eclipses are welcome to participate in the MMV Project. For more information, and to view already processed images of total solar eclipses, see our website at www.zam.fme.vutbr.cz/~druck/Eclipse/