Back to Malin's Homepage
A Brief Look at Techniques in Astronomical Photography
by David Malin
About the colour images - an outline
Most Australian Astronomical Observatory colour photographs were made with the Anglo Australian Telescope (AAT) or the UK Schmidt Telescope (UKST). A few in our collection have been made by plates taken with the Isaac Newton Telescope on La Palma and two pictures of the Crab Nebula were made from plates taken many years ago with the Hale 5m telescope (see Malin and Pasachoff, Discovering the Color of the Crab Nebula in the July, 1993 edition of Sky and Telescope (86 :1 43-46)
Although the AAT uses a mirror to collect and focus light, when it is used as a camera it is equivalent to a lens with focal length of 12.7 metres and a focal ratio of F/3.3. The square field used for photographic imaging covers one degree of sky (about two moon diameters across) and the images were recorded on special black and white photographic emulsions coated on glass plates 255 mm (10 inches) square. Photography is no longer used on the AAT, having been replaced by electronic detactors such as CCDs.
The UK Schmidt is primarily a photographic survey telescope and has a focal length of 3.07m with a focal ratio of F/2.5. It photographs a 6.6 x 6.6 degree field on plates or film 356 mm (14 inches) square. However, like the AAT it no longer takes photographs, but has been modified to make wide field spectroscopic surveys, using optical fibres to make many individual spectra in one long exposure.
The colour images are created by combining three separate, black and white exposures with combinations of plates and filters designed to record blue, green or red light. For the faintest objects exposures of 90 minutes are required for each colour. The monochromes are then manipulated, enhanced and combined photographically to produce the final picture, which show the object much as it might appear in the telescope if our eyes were much more sensitive to the colour of faint light. It is these techniques that are described here. For more background on the practicalities of please click here
About the images - technical detailsIntroduction
The techniques developed by David Malin have led to a unified system of photography designed specifically for the unusual imaging requirements of astronomy. The key elements are the use of diffuse light contact copying for all stages of image transfer except the last, which usually employs an enlarger, and a simple additive method for combining many plates into either very deep images or colour pictures -- or both. All the methods are non-destructive and employ simple equipment and readily-available materials. The are described in more detail in the Technical Appendix of Colours of the Stars by David Malin and Paul Murdin (Cambridge University Press, 1984), and an illustrated bibliography describing the techniques is also available on line -- and still under construction.
Unsharp masking, for exploring the images of bright objects
In a conventional negative on ordinary black and white film, the highlights (dark on the neg) transmit about 50 times less light than the shadows. This 50:1 density range (D max about 1.7) embraces all the image information on the negative and ensures that a normal print will retain detail in both the brightest and darkest parts of the scene, if the printer so choses. By comparison, an astronomical negative might have a density range approaching than 100,000:1 (D max 5) because the emulsions used are designed to have high information storage capacity. Bright objects recorded on such materials are completly unprintable in any normal way. However, the information in these high densities can be extracted by a technique which was initially used in the graphic arts industry.
The method involves making a blurred positive film copy of the original negative by contact and then copying the original and the out-of-focus positive together to remove the 'unsharp' information (that contained in the blurred positive) from the photograph. The blurring is usually achieved by placing the film for the unsharp mask in contact with the back of the glass negative and using a diffuse light source to make the exposure, the glass thickness of the plate acting as a spacer.
When the mask has been processed, it is replaced in position on the back of the original and a normal, emulsion-to-emulsion copy is made on another piece of film. The blurred positive cancels the low frequency information in the negative, acting as a kind of automatic dodging mask, leaving the fine detail untouched. This sounds more complicated than it is in practice, and a diagram should clarify the stages. Consistency, careful control of exposure and processing and a little experience are needed for success. However, the improvement in detail that the technique can reveal is quite dramatic.
Copying with an unsharp mask is useful wherever large scale, low information-content detail contributes to the overall density while obscuring finer structures, for example where structures in the dust lane of Centaurus A is hidden in the bright envelope of the giant elliptical galaxy. The illustrations show (a) the appearance of the original plate, (b) an unsharp mask made from it and (c) the result of copying (a) through (b).
Photographic Amplification, for the faintest features
Deep exposures in astronomical photography are ultimately limited by the uniform fogging of the plate from the night sky airglow -- 'sky-limited'. The exposure of a sky-limited plate is carefully controlled so that the minimum photographic (sky) density obtained is that where the emulsion exhibits maximum contrast, usually at a developed density around 1.0 (10 percent transmission) above chemical fog. This is thus the minimum image density on the plate. Faint images are therefore at some small density greater than the fog produced by the airglow. Because the sky density represents a small fraction of the total storage capacity of the material, both the fog and the faint images tend to be in the upper part of the developed layer.
If a low density contact copy of a developed plate is made on to high contrast film, using a diffuse light source, the grains of the original are recorded on the copy with greatly increased apparent size, as shown on the diagram. High contrast prints made from this thin-looking but contrasty positive reveal objects which are less than one percent as bright as the night sky, equivalent to a surface brightness of about 28 mag per square arcsecond, or 5 to 6 magnitudes fainter than the Siding Spring night sky. These are some of the faintest astronomical images ever detected. Giant shells around elliptical galaxies and Malin-1 were discovered in this way. In combination with unsharp masking, contrast enhancement also revealed the internal structure of some shell galaxies, such as NGC 3923.
Combining many images to improve the signal-to-noise
Even with photographic amplification, there are images that are too faint to be detected against the grain noise of the developed emulsion. This is particularly true where photo-amplified images are greatly enlarged. Fortunately for the astronomical photographer, the subject matter remains the same night after night, year after year, so repeat exposures record the same stars and galaxies but have different pattern of granularity. Combining the data from several different plates therefore strengthens the signal while smoothing out the random noise inherent in the emulsion layer.
The improvement in signal to noise is evident in this series of photographs of the peculiar galaxy NGC 4672. Photograph (a) shows a greatly enlarged image of it as it appears on a UK Schmidt plate, (b) shows this image after photographic amplification, and (c) combines the information from photographically amplified derivatives of seven sky-limited plates of the same field. Some well-known objects such as M100 become unrecognisable as deep images reveal unexpected interactions with faint companions.
Colour photography of faint objects
Experiments with ordinary colour films at both the AAT and the UK Schmidt gave good results for only a few relatively bright objects, and most of these had already been photographed in colour in this way. Fainter objects were hardly visible above the the night sky airglow, which of course reflects the true situation and explains why the special black and white emulsions designed for astronomical use must be so contrasty.
Since these black and white emulsions can have their long-exposure speed increased by factors of 10 or 20 times by special baking and gas soaking techniques (hypersensitisation), and are available in a range of colour sensitivities, a 3-colour photographic system has been devised to take advantage of these properties. It is based on the earliest of all the colour photographic processes, James Clerk Maxwell's additive colour system, which was first demonstrated in 1861.
In this modern variant, positive black and white film contact copies are made from three separate negatives taken to record the red, green and blue light. The positive images are combined in register on colour film or paper using three separate exposures in an enlarger with the appropriate red, green or blue colour filters. This method had several important advantages for astronomy, the main one being that the image enhancement processes described above are easily incorporated at the positive-making stage. In addition, the method uses conventional, high-efficiency hypersensitized plates and the normal plate/ filter combinations used for standard astronomical photometry.
The power of this approach is demonstrated if we compare colour photographs of the same field photographed on film and using the three-colour process. This is the well-known region around Rho Oph and Antares. Image (a) was photographed with the UK Schmidt telescope on colour negative film, while (b) is the same field photographed with the same telescope on three black and white plates which were combined into a colour image.
The plates and filters used for this process are selected so that the colours are reproduced a close to 'true-colour' as possible, much as the eye might see them if it were much more sensitive to faint light. However, the dyes used in the colour materials that make prints and negatives, and the phosphors in the screen you are looking at now (to say nothing of the RGB transformations in the electronics) cannot precisely reproduce the colours of nature, only a representation. This is no different from any other kind of photography, except that here we are revealing colours that have never been seen before.
The colour photographs that appear elsewhere in these pages are the result of sporadic activity and development over about 23 years. I hope they give you as much pleasure as making them has given me.
Scan David Malin's general publications. Back to Malin's Homepage