Signals of Reality · Reality Is Not What You See

The Picture Is Not the Thing

An astronomical image is a chain of measurements and display choices. Understanding that chain makes the picture more informative, not less real.

Start with the small print beneath the spectacle. In NASA’s Webb image of the Tarantula Nebula, the color information identifies four instrument filters and the visible colors assigned to them: F090W to blue, F200W to green, F335M to orange, and F444W to red. That key is one of the most informative parts of the picture. It tells us how to read what looks, at first glance, like a richly colored view through a window. See NASA’s NIRCam image record.

The colors do not mean that a nearby human observer would see exactly that palette. They encode measurements made through different infrared filters. The image is a visible presentation of information that our eyes could not otherwise inspect in this form.

A suspicious reader might call that artificial. It is certainly constructed. The important question is what the construction preserves.

Before there is a picture

A scientific image starts with a measurement arrangement: an instrument, a target, a selection of wavelengths, an exposure, and a detector. The output is not yet the familiar color scene on a screen. It is data that must be interpreted in relation to the instrument that produced it.

NASA’s explanation of Webb image processing describes how separate filter images become a composite and how infrared measurements receive visible colors. Processing includes practical work on artifacts and alignment as well as choices about display. The finished release carries both information from the sky and decisions about how to make that information legible.

The ESA/Webb processing guide likewise shows separate observations being assembled into a color image. Neither account asks us to believe that the observer simply invents the scene. They describe a route from recorded signals to a representation people can inspect.

It helps to separate three questions. What did the detector record? What physical source best explains those readings? How should the readings be displayed? These questions interact, but an answer to one is not automatically an answer to the others.

Changing a palette may leave the recorded measurements intact. Changing the physical interpretation of a feature may leave both the measurements and palette intact. Altering the underlying measurements without disclosure is a different matter altogether.

A pixel with several possible histories

Consider a deliberately simplified example with invented numbers. Two detector locations record values of 10 and 100 after an appropriate calibration. On a linear grayscale display with a fixed scale, the second receives a much brighter shade. If the full image contains values extending to 100,000, however, both may appear nearly black when that entire range is squeezed onto an ordinary display.

A nonlinear brightness transformation can make the faint structures visible. The resulting image may show both dim and bright regions clearly, but a visual comparison of their screen brightness no longer reproduces the ratio of the original values. That ratio has not necessarily been lost from the data; it has ceased to be readable directly from the display.

This example is arithmetic, not a report about the processing settings of the Tarantula image. It illustrates why the instruction “show the data without doing anything to it” is incomplete. A display requires a rule mapping numbers to visible output. Even a linear rule needs a range and endpoints.

The choice should follow the question. If the purpose is to reveal faint structure, a transformation that compresses bright values may be useful. If the purpose is to compare measured intensities, access to calibrated values and the transformation rule becomes essential. One rendering need not serve both tasks equally well.

An image can therefore be faithful without being visually literal. The relevant fidelity might concern positions, boundaries, relative responses across filters, or a documented pattern of intensity. It is not the vague impression that the picture looks photographic.

The caption is part of the evidence

Suppose a bright feature appears red. Before inferring that it is hotter, cooler, farther away, or chemically different from its surroundings, the reader needs to know what red encodes. It could represent stronger response in a chosen wavelength band, or a derived quantity in another kind of visualization. Color alone does not determine the physical explanation.

NASA’s filter key for the Tarantula image offers a concrete starting point. It connects the composite to particular observations rather than leaving the palette as an unexplained aesthetic fact. Further physical claims still require the relevant analysis. A color assignment is not, by itself, a measurement of age or composition.

This is the role of a good caption: to prevent a compelling visual feature from acquiring unsupported meaning. Labels such as “artist’s concept,” “simulation,” and “composite observation” distinguish different routes by which an image came to exist. None is an insult. Each sets expectations about what the picture can establish.

Imagine two visually identical pictures of an unfamiliar planet. One is a proposed illustration, and the other is reconstructed from observations. Their resemblance would not make them equivalent evidence that the planet has a particular surface. Provenance matters because it describes the connection between the marks and the claim.

Nor does provenance make an observed image infallible. A real measurement can contain artifacts, and a plausible reconstruction can depend on assumptions that later prove inadequate. The advantage is that documented procedures expose places where the result can be questioned and improved.

An image worth returning to

The comparison with an ordinary photograph is helpful only up to a point. Both involve capture and display choices, but an infrared composite does not have to reproduce the appearance of a scene to unaided human vision. That is often not its purpose. Treating visible appearance as the universal standard would disqualify precisely the measurements that extend our reach.

A useful reading habit is to move from the whole picture to a particular claim. The image may immediately suggest a cavity, a cluster, or a boundary. What in the measurements supports that impression? Does it persist across filters? What does the accompanying analysis establish, and what remains an interpretation? These questions let visual curiosity become investigation.

The wonder need not shrink. Knowing that the Tarantula image combines infrared observations adds another scale of achievement to what is visible: an instrument made measurements, a processing chain preserved usable information, and a display translated part of it into a form that human vision can explore.

The nebula is not contained in the rectangle. But the rectangle can carry evidence from it, provided we keep the key.

Sources and image documentation

NASA’s Tarantula Nebula image page includes the filter-to-color assignments. NASA’s processing explainer and the ESA/Webb guide describe the path from observations to a public image. The numerical display example above is hypothetical.

Share this article
Facebook WhatsApp X LinkedIn Medium
Quick navigationEsc