Colour Science Behind Film Looks

"Colour science" is a term that gets thrown around in camera discussions without much precision. People say Fujifilm has "great colour science" or that Sony's colours are "clinical." What they are actually describing is the set of decisions a camera manufacturer makes about how to translate the light hitting the sensor into the colours you see in the final image. Every camera has to make these decisions, and they are the reason a photograph of the same scene looks different on different camera brands even with identical exposure settings.

These decisions include: how the sensor's colour filters are arranged and what wavelengths they pass, how the raw sensor data is demosaiced (converted from a grid of single-colour pixels into a full-colour image), what tone curves are applied to map the sensor's linear light response into a perceptually pleasing image, how individual colour channels are adjusted relative to each other, and how all of this changes depending on which colour profile or film simulation is selected.

Fujifilm's colour science is distinctive because it starts from a different place than other manufacturers. Where Canon, Nikon, and Sony developed their colour processing primarily as digital engineering problems, Fujifilm developed theirs as an extension of decades of film manufacturing. The question they asked was not "how do we make a technically accurate digital image?" but "how do we make a digital image that has the quality of colour that our films had?" That distinction runs through everything the camera does, from the sensor design to the final JPEG output.

Most digital cameras use a Bayer colour filter array: a repeating pattern of red, green, and blue filters arranged in 2x2 blocks (one red, one blue, two green) laid over the sensor pixels. This pattern is simple, well-understood, and supported by decades of optimised software. Fujifilm's X-Trans sensor uses a different arrangement entirely.

The X-Trans colour filter array uses a larger 6x6 repeating pattern with a more random-looking distribution of red, green, and blue filters. Every row and every column in the pattern contains all three colours, which is not the case with Bayer. This design has two practical consequences.

First, the more complex filter pattern reduces moiré and false colour artefacts without needing an optical low-pass filter (the anti-aliasing filter that most Bayer sensor cameras use). Removing the low-pass filter means the sensor resolves more detail from the lens. This is why Fujifilm's APS-C sensors often resolve detail that rivals higher-resolution Bayer sensors: the light reaches the photosites without being softened first.

Second, the X-Trans pattern requires a different demosaicing algorithm. Software that was written for Bayer sensors cannot simply be applied to X-Trans data. This is why some RAW processing software historically produced artefacts with Fujifilm files, and why the camera's own JPEG processing, which uses algorithms specifically designed for the X-Trans pattern, often looks better than third-party RAW conversions. The camera knows its own sensor better than any external software does.

The X-Trans colour filter arrangement also affects how colours are sampled and interpolated. Because every row and column contains all three colour channels, the camera has more colour information to work with at every point in the image, which contributes to the smooth colour gradations that Fujifilm images are known for. Whether this makes X-Trans objectively "better" than Bayer is debatable. What is not debatable is that it produces different results, and those differences are part of what gives Fujifilm images their particular quality.

A physical film emulsion is a complex chemical system. Multiple layers of silver halide crystals, each sensitised to different wavelengths of light, sit in a gelatin base. When light hits the emulsion, it creates a latent image that is revealed through chemical development. Each film stock has a specific spectral sensitivity (which wavelengths it responds to and how strongly), characteristic dye colours (the actual pigments that form the final image), a particular grain structure (the size and distribution of the silver halide crystals), and a unique tone response (how it handles different levels of brightness).

When Fujifilm creates a digital film simulation, they are not simply applying a colour filter or a preset. They are attempting to reproduce the entire chain of how a specific film stock responded to light and rendered a scene. This means modelling how Velvia's dye set amplified certain reds and greens while keeping blues deep and saturated. It means replicating the way Provia's neutral dye balance produced accurate colours across a wide range of subjects. It means capturing the way Classic Chrome's processing mimics the desaturated, shadow-rich quality of certain reversal films used in magazine printing.

The result is more nuanced than a simple look-up table that shifts colours. A film simulation adjusts different colours differently depending on their saturation, luminance, and context within the image. A moderately saturated red behaves differently from a deeply saturated red. A green in bright sunlight shifts differently from a green in shadow. This contextual, non- linear colour mapping is what gives simulations their organic, film-like quality. It is also why a Lightroom preset that approximates "Classic Chrome" never quite matches the camera's output: the preset applies a simpler transformation that cannot capture all the conditional colour adjustments the camera's processor makes.

The simulations that are not directly modelled on a specific film stock, such as Classic Neg. and Nostalgic Neg., still draw on the same principles. Classic Neg. reproduces the general characteristics of consumer colour negative film: the warm shadows, cool highlights, and moderate contrast that resulted from the typical negative-positive printing process. Nostalgic Neg. simulates the colour shifts that occur in aged or improperly stored negative film. Even these "non-specific" simulations are grounded in real photochemical behaviour rather than arbitrary digital colour manipulation.

A tone curve is a mapping function that translates the brightness values captured by the sensor into the brightness values displayed in the final image. The sensor captures light linearly: twice as many photons produce twice as bright a signal. But human vision is not linear. We are far more sensitive to differences in dark tones than in bright ones. A tone curve compensates for this, compressing the highlights and expanding the shadows to produce an image that looks natural to our eyes.

Every film simulation uses a different tone curve, and the tone curve is arguably the most important factor in a simulation's visual character. Velvia's curve is steep and contrasty: shadows drop off quickly, highlights are bright and saturated, and the midtones are punchy. Classic Chrome's curve is softer in the highlights but drops shadows deep, producing that characteristic combination of restrained bright areas and rich, weighty dark areas. Pro Neg. Std's curve is the flattest in the system, with gentle transitions throughout the range that preserve maximum detail for post-processing.

The Film S-Curve

Physical film has a characteristic S-shaped response curve. At low light levels (the "toe" of the curve), the response is gentle: shadows gradually fade to black rather than clipping abruptly. At high light levels (the "shoulder"), the response also tapers off, with highlights gradually rolling into white rather than hitting a hard ceiling. The midtone section between toe and shoulder is steeper, providing good contrast where it matters most for everyday subjects.

This S-curve is one of the primary reasons film images "look like film." Digital sensors, by contrast, have a linear response that clips abruptly at both ends. A properly exposed digital image has hard edges to its tonal range: pure black at the bottom, pure white at the top, with no gradual transition. Film simulations apply a digital version of the S-curve to soften these transitions, which is why simulations look more organic than a flat digital rendering. The highlight and shadow tone controls in your recipe adjust the steepness of these curves at the top and bottom of the range, giving you direct control over how "filmic" the roll-off feels.

Highlight Roll-Off vs Digital Clipping

The most visible difference between a film-like tone curve and a raw digital capture is in the highlights. When film approaches its maximum brightness, the response tapers off gradually: white shirt detail, cloud texture, and skin highlights in bright light all transition smoothly towards white. A raw digital capture clips hard: once a pixel reaches maximum brightness, it is pure white, and the pixel next to it may still have full detail. This creates an unnatural, jagged boundary between "bright with detail" and "gone."

Fujifilm's film simulations, particularly when combined with the Dynamic Range settings, work hard to recreate the gradual highlight roll-off of film. This is one of the key reasons Fujifilm JPEGs look more pleasing than many other camera brands' output in high- contrast situations. The highlights do not disappear suddenly; they fade gracefully.

White balance is usually taught as a correction: set it to match the light source so whites look white. That is technically correct but creatively limiting. In the context of film simulations and recipe building, white balance is one of the most powerful mood-setting tools available, and using it purely for correction misses its real potential.

How White Balance Interacts with Simulations

A film simulation is designed and tuned assuming a certain range of colour temperatures. When you change the white balance, you are feeding the simulation different raw colour data, and the simulation responds to that data according to its built-in colour mapping. This means the same WB shift can produce very different results on different simulations. Shifting towards warm on Classic Neg. amplifies its already-warm shadow character and pushes the image towards a rich, amber-toned look. The same warm shift on Classic Chrome warms up a simulation that is naturally cool and muted, producing a more balanced but golden-hour quality.

This interaction is not additive in a simple way. Because simulations apply non-linear colour transformations, a warm WB shift does not just add orange to every pixel. It changes which colours fall into the simulation's various adjustment zones, which can produce subtle secondary effects: a warm shift on Velvia might push certain greens from "vivid" into "oversaturated," while the same shift on Astia keeps everything within a flattering range. Learning how your preferred simulations respond to different white balance settings is one of the most valuable skills a recipe builder can develop.

Daylight WB as a Creative Default

Many recipe creators use Daylight white balance rather than Auto, even when shooting in mixed or artificial light. The reason is consistency and intentionality. Auto WB neutralises the colour of the light, which sounds correct but often removes exactly the quality that makes a scene interesting. Late afternoon golden light becomes neutral. Warm indoor tungsten light becomes neutral. Cool blue-hour light becomes neutral. Daylight WB preserves the natural colour of the light source, allowing warm light to actually look warm in the final image.

This is how film photographers worked: you loaded daylight-balanced film and the colour of the light became part of the image. The film did not "correct" for the colour temperature of the scene. Shooting with a fixed WB setting recreates this approach and produces images with more consistent colour character across a session, which is exactly what recipe shooters want.

The WB Shift Axes

Beyond the main colour temperature setting, Fujifilm's WB shift grid lets you fine-tune the colour cast in two dimensions. The axes are labelled R and B, each adjustable from -9 to +9. R runs from red at the positive end to cyan at the negative end; B runs from blue at the positive end to yellow at the negative end. Between them they reach any hue on the grid: red-plus-yellow gives you orange, cyan-plus-blue gives you teal.

It is worth being precise about this, because Fujifilm's grid is often described as though it carried a green-magenta tint axis like Lightroom's. It does not. The only place Fujifilm exposes a green-magenta control is Monochromatic Color, which applies to ACROS and Monochrome only and tones a black and white image on Warm-Cool and Green-Magenta axes. If you want a magenta cast in a colour recipe, you reach it by combining R and B, not by a dedicated tint slider. Small adjustments on these axes (plus or minus 1 to 3) are among the most common recipe tweaks, and they can transform the mood of a simulation without changing its fundamental character.

Understanding colour science is not about memorising technical details. It is about developing an intuition for how your camera produces colour, so that when something looks wrong or you want to achieve a specific mood, you know where to look and what to adjust.

Predicting How Simulations Respond to Light

If you understand that Classic Chrome desaturates and deepens shadows, you can predict that it will thrive in overcast light (where its desaturation turns grey into atmosphere rather than dullness) and struggle in already-desaturated scenes (where it pushes colours into anaemic territory). If you understand that Velvia amplifies saturated colours non-linearly, you can predict that it will make autumn foliage glow but push already-vivid flowers into unnatural neon. This predictive ability means fewer wasted shots and faster creative decisions in the field.

Diagnosing Colour Problems

When a recipe does not look right, colour science knowledge helps you identify the cause. If skin tones look too warm, is it the simulation's inherent colour mapping, the white balance, or the WB shift? If shadows look muddy, is it the tone curve compressing shadow detail, noise reduction smearing fine tones, or the simulation's shadow colour cast? Each of these has a different solution, and understanding the system helps you reach for the right adjustment rather than randomly tweaking settings until something looks better.

Understanding Why Certain Combinations Work

Popular recipe combinations are not arbitrary. Classic Neg. with warm WB shift and slight desaturation works because the simulation's warm shadow character is amplified by the WB shift while the desaturation prevents the overall palette from becoming cloying. ACROS with a red filter and grain works because the red filter increases contrast in skies and skin (mimicking the effect of a physical red filter on panchromatic film) while the grain adds the organic texture that makes high-contrast black and white feel authentic rather than digital.

Once you understand the colour science behind these combinations, you can create your own rather than copying someone else's settings. You can ask yourself: what is this simulation already doing well, and what do I want to push further or pull back? That question, informed by an understanding of how the system works, is the foundation of every good recipe.

For the practical application of these principles, the Recipe Creation Tips article walks through the process of building recipes from scratch, and the Film Simulations Comparison describes each simulation's character in detail.