Horizon to Horizon
The Milky Way on Film
By Thomas Wright, Bill Scheffler, and Chris Huff

It is fair to say that the Milky Way is our night sky. Surrounding us, the denser parts of the disk form a continuous band. With few exceptions, everything that can be seen without optical aid is part of our galaxy. Only the Clouds of Magellan, the nearby galaxy M31, and, for those with sensitive eyes, M33 are naked eye objects outside of our spiral home.

Until recently, the only photographic portraits of our galaxy were the Mt. Stromlo H-alpha atlas and a black and white photographic mosaic done by the European Southern Observatory. Also, there is the image done by the Lund Observatory, which is drawn from photographs. The Milky Way deserved a new high-quality portrayal.

In 1997 we began a project to capture on film large areas of the sky at high resolution. Our first subjects were constellations. Since some of them are quite large we knew we needed a wide-angle camera. We also wanted to show the beautiful and interesting objects within the constellations, so we needed focal length.

The two needs led us to large-format. After studying and testing we chose 4" x 5" film and a 90mm lens. This combination can resolve to less than an arcminute across 70 degrees of sky. A practical result is that large-format images can be enlarged far beyond 35mm. The 4x5 format is standard for commercial product photography for just this reason.

After looking at our constellation pictures, it was a natural step up to the Milky Way project. It is the largest object in our sky, a source of endless detail as well as broad sweep. We knew that our camera could give us the galaxy in just a few shots. But how to do it?

Combining pictures into a mosaic is an art that has been practiced in darkrooms, but the computer can achieve far more seamless results. An experienced retoucher with an artistic eye can provide the visual fine-tuning needed to stitch frames into a single image. We therefore decided to scan, edit, and then combine our filmed segments into a panorama.

Distortions and perspective differences are inherent in wide-angle images. These can be reduced greatly, however, with the right lens and frame selection. A good large-format lens is distortion-free in that it records straight lines as straight. However, the stars can be thought of as an infinitely distant sphere (Aristotle was correct for this purpose). Therefore, the camera projects from a sphere to the flat film. This means that objects in the corner of the frame are stretched away from the center. If overlapping frames are in a line, without different tilts, there will be a line along the edge of each frame where the mapping agrees. What we had to do was ensure that all the frames were in a line-the galactic plane. That was not easy!

In order to track the sky the camera rotates around the Earth's axis. The galaxy has a different axis, so each shot would need to be rotated to agree with the galactic plane. The stretching effect in the corners makes polar alignment critical, especially since the galaxy comes very near both celestial poles. The camera must be very rigidly supported to avoid sagging during long exposures, so we did not use a tripod head.

Typical large-format cameras use a bellows and rail system for focusing, which can sag, so we made a rigid box with a helical focuser. A frame sitting firmly on the declination cradle had a circular cutout for the camera, marked in degrees. For the chosen fields, each 45 degrees wide by 60 high, the angle between the galactic plane and lines of declination was measured from two or three charts, and the camera set to that angle. The center of each frame was the galactic plane at 45-degree positions, four from the Northern Hemisphere and four from the southern.

If we achieved alignments accurate to one degree, the differences in star displacement would be so slight that no stretching or constriction of the images would be needed. In this we were fairly successful, so that the overall shape of the galaxy is accurately rendered. Simply cutting together darkroom prints would be easy, but the computer allows us to adjust contrast and color balance to a degree not possible in the darkroom.

A frame of 60 degrees is one-third of the sky, and demands perfect, cloud-free conditions. Choosing where and when to start meant balancing sky conditions, frame size, the availability of particular parts of the galaxy, and our own time off from other work. The available time proved to be fall or winter in each hemisphere. Weather is quirky in winter, so some research had to be done in choosing reliable locations.

Hawaii, with excellent climate and an advantageous southern location, proved to be too far south to include the Milky Way in Perseus and Cassiopeia. Weather records suggested that west Texas would be useful in November, and it has the advantage of being quite far south, roughly 30 degrees north latitude. At the corresponding southern latitude, we chose the town of Broken Hill, New South Wales, Australia. Both locations offer dry air, moderate winter temperatures, and unobstructed horizons.

The McDonald Observatory Visitors Center, in Fort Davis, Texas, is an excellent location for photography. Nice people, too. Our first night of shooting was during the Leonid shower of 1998. The staff opened the center at 1:00 a.m. and stayed open until 3:00 a.m.! Conversation with many enthusiastic meteor-watchers was pleasant diversion during the 80-minute shots. Two good frames were captured that night, and the other two successfully shot a day later. Six or seven meteors were caught on film but removed for the final image.

Broken Hill has no professional observatory, but near-perfect conditions. At the edge of a great expanse of outback desert, large enough for pleasant accommodations and isolated enough to show a very dark sky a short distance from town, it is an astronomical location hard to beat.

There is at least one good-size telescope in town, a 16-inch at the private observatory of Trevor Barry. He and his family made our stay a friendly and rewarding one. Bringing home the developed film from each site was the prize. Knowing that we had half the Milky Way in a small box was exciting, a sort of secret pleasure.

Most 4x5 film is slow, with the fastest being ISO 160. Newer black and white films became available with speeds of ISO 400, but these films had poor red response. We considered tri-color B&W but rejected it because of its longer total exposure times. If the target field sets before you're done you don't have a shot. Returning to reshoot a red or green frame would mean accepting the risk that slight aiming differences would produce large distortions in the corners, making it impossible to combine images into one color frame.

We selected a new Kodak negative film 400MC. This film performs well after hypering. We always processed film on site to be sure we had a good shot and to preserve the speed of the hypered film. Unfortunately, this film was taken off the market after only a few months. It was replaced with the new Portra series. This film change posed challenges in matching the backgrounds and star colors in the final image.

Our approach performed very well, recording past 12th magnitude with hypered ISO 400 film, with resolution as good as 30 arcseconds. Every Messier object, and many NGC ones, are not only recorded but resolved! We calculate that one frame can in principle show 5 million stars. In the final images, a good guess is 500,000 stars in each of the eight frames. This project has been an education in galactic structure, mapping geometry, weather around the world, film processing, computer capabilities, and camera technology.

Each frame was scanned to yield a 250Mb file in 8-bit RGB. The combined file size of the eight panels was 2Gb, and too many pixels long to handle in Photoshop. Combining the individual frames from each hemisphere, two large segments, north and south, were prepared. The resolutions were then decreased for final assembly. The image still resolves open clusters, nebulae, and globular clusters, and is a treat to peruse with a magnifier. The long focal length shows bright stars without making fainter stars grow into fat blobs. This is essential to keeping a sense of depth to the sky.

Traveling with camera equipment precluded bringing a telescope, but those long nights under a desert sky have reminded us of the pleasure of naked-eye stargazing. In Australia, the center of the galaxy is overhead. One sees it as a galaxy, not just a blur of stars. Dust clouds stand out from the bright star fields. A side benefit of spending hours studying the captured image is a sense that one can see these things in imagination while looking at the dismal urban sky. Large-area photos also aid in finding objects in the field, since one gets an intuitive sense of their positions.

Since most of us live in large cities, blind to the starlit sky, this image may restore a sense of wonder in that other half of the world, the night. For us stargazers it makes a cloudy night less disappointing, since we can now travel through the galaxy from our armchairs.