Picturing the Cosmos: Hubble Space Telescope Images and the Astronomical Sublime

3. Translating Data into Pretty Pictures

127 THREE TRANSLATING DATA INTO PRETTY PICTURES Hybrid Images How do Hubble images represent the cosmos? In chapter 2, I argued that they are more than pretty pictures, and in fact have scientific and aesthetic value. I have referred several times to their status as digital images, and I explained very briefly how that can affect their appearance. However, more attention and more sophisticated analysis is necessary to properly address the topic. The medium of the Hubble images coupled with their relationship to science makes them into complex hybrids. Their digital makeup brings together two modes of representation, number and image. Working in concert, the telescope and cameras extend perception by magnifying and recording light; they enable observations with incredible precision and across a range of wavelengths. The resulting data are indexical records of physical properties that exceed the limits of vision and of representation. The astronomers must use image processing software to make that expanded range of data legible to the human eye. Because digital data are numeric, astronomers can intervene by modifying a single value, or pixel, or an entire data set. But they must also determine , given the impossibility of a perfect translation, the elusiveness of a perfect reflection of numeric data as visual representation, when to maintain indexical aspects and when to convey attributes symbolically. Analog images, of course, also allow for specific and general alterations to their appearance. Digital images are distinct because of the ease of making such changes and the possibility of doing so in a manner that leaves no trace. Every astronomer 128 TRANSLATING DATA has the tools—a computer and image processing software—to make such changes, whereas changes to a photograph required a darkroom and the expertise to use it. (That is not to say that manipulating digital data does not require expertise, but the opportunity to do so is far more available.) Also, digital image processing can offer a far greater degree of control. To make certain types of modifications, astronomers apply algorithms based on mathematical functions, a process that can be exactly repeated or reversed. Software can automate mathematical transformations that astronomers might perform with a set of equations, as well as allow for pictorial expression of the modified data. As scientific images, which should accurately refer to the phenomena they represent, such changes might seem unnecessary for the Hubble images. Isn’t the best image, the truest one, the version that most directly transforms the phenomena of the world into a representation? Such a notion rests on the objectivity of the instruments , or what the historians of science Lorraine Daston and Peter Galison call mechanical objectivity. However, as they argue, twentieth and twenty-first century scientists have increasingly supplemented mechanical objectivity with trained judgment , thus producing an image reliant on the ability of its maker “to synthesize, highlight, and grasp relationships in ways that were not reducible to mechanical procedure.”1 For the Hubble Space Telescope, the larger cultural turn toward trained judgment has another rationale: the extension of perception exceeds the capacity of representation. Hubble images cannot imitate the appearance of celestial objects because some of their features are invisible to our eyes. They also cannot exactly imitate the data; the precise distinctions within the data also are beyond what our eyes can see. Therefore, astronomers have had to find a way to display the data that makes it possible to experience the cosmos with the senses without compromising the validity of the representation. Perception and representation, number and image, index and symbol—the hybridity of the Hubble images raises a question that haunts all scientific images: What is their relationship to the phenomena they purport to represent? When observing with the Hubble Space Telescope, astronomers are forced to consider what constitutes a legitimate image. What modifications to contrast, color, and composition are acceptable and which ones are not? Because the new medium of digital imaging upended estab- 129 TRANSLATING DATA lished conventions, astronomers who worked with Hubble data early in its history (for whom the Hubble Heritage Project often acted as spokesperson) found themselves groping toward a new set of guidelines for representing and displaying astronomical phenomena. In many cases astronomers sought direct correlations, a key that could be followed in multiple situations, but the complexity of the data, the diverse audiences for the images, and deeply engrained ideas about the visual culture of science made such an approach inadequate. Furthermore, it limited...