Three sliders, one long chemical chain
When you invert a negative on your phone, you drag three sliders: black point, white point, gamma. But the frame under those sliders is the end product of light passing through three emulsion layers and then being corrected by a cocktail of couplers in the developer. Without the vocabulary, sliders are guesswork — once you know where the orange comes from, you know how it should be taken away.
This article contains no recipes and no manufacturer numbers (those belong in the official datasheet and nowhere else). It is about mechanisms: which layer each term refers to, what problem it solves, and what shape it takes when you meet it again in digital post.
The characteristic curve: contrast and latitude are drawn, not given
Log exposure on the horizontal axis, developed density on the vertical: the D–logE curve is a film stock's identity card.
It has three parts. The toe — underexposure, where density rises slowly and shadow separation gets squeezed flat. The straight-line section — the midtones, whose slope is the contrast (γ). The shoulder — the highlights, where density stops climbing and the negative rolls its highlights off instead of clipping them.
A negative's shoulder is long and gentle, which is exactly why negative film is said to have wide latitude and to prefer over- to underexposure. Slide film has a steep shoulder and blows out within a stop. A digital sensor, by contrast, responds almost linearly and clips highlights hard. So the "highlight rolloff" control in a film simulation is not decoration: it is standing in for that shoulder.
Develop the same roll a little longer and the straight-line slope gets steeper. Contrast is not fixed at the moment of the shutter — it is pulled out of the film in the tank.
The relationship between γ / CI (contrast index) and development time differs for every developer. Read the manufacturer's time–contrast chart for the combination you actually use; never transfer numbers between developers, and never fill them in from memory.
Three emulsions: who "sees" what, and who "blocks" what
Silver halide on its own is sensitive only to blue and violet light. Colour film uses sensitizing dyes to extend the other two layers into green and red — and that is what a spectral sensitivity curve describes: how sensitive each layer is at each wavelength.
That is why the top layer is usually the blue-sensitive one, immediately followed by a yellow filter layer that absorbs the remaining blue so the layers underneath — which are inherently blue-sensitive too — are not contaminated by it.
After development each layer holds not silver but dye: yellow, magenta, cyan. Now a second curve takes over — the dye absorption spectrum, describing which wavelengths each dye blocks. Sensitivity governs how light gets recorded; absorption governs how the record is seen again.
Only both curves together make a stock's "colour". A LUT is one slice through that chain under one illuminant — change the colour temperature of the light and real film and the LUT part company.
| Curve | What it describes | Digital counterpart |
|---|---|---|
| Spectral sensitivity | Each layer's response per wavelength | Colour filter array response + white balance |
| D–logE curve | Exposure → developed density | Tone curve / highlight rolloff |
| Dye absorption | Which wavelengths the dye blocks | Channel mixing / saturation and crosstalk |
The orange mask is not a protective coating — it is a deliberate colour cast
An ideal magenta dye would absorb only green, an ideal cyan only red. Real dyes have unwanted absorptions: magenta also eats some blue, cyan also eats some blue and green. Left alone, prints come out muddy, dull, and refuse to saturate.
The manufacturers' answer is coloured couplers, also called masking couplers: where the coupler has not taken part in development it is itself coloured (yellow and red); where development happened it is consumed. What is left across the frame is a mask complementary to the image, dense exactly where it needs to cancel the unwanted absorption. Overall it reads orange — that is the colour negative's orange mask.
The orange mask is not a uniform sheet of orange. Its density varies with picture content — which is why subtracting one fixed orange will always be an approximation.
That has a practical consequence. When scanning or camera-scanning colour negative, the more reliable move is to sample the film base first — the unexposed leader, or the clear strip between the sprocket holes — and use that as the reference before pulling each channel. It is exactly why the "sample film base" step exists in the Negative Lab.
Black-and-white negatives and colour slides have no orange mask, so their inversion chain is one whole step shorter. That is often the real reason behind "switching it to black and white suddenly fixed the colour".
DIR couplers: film that sharpens itself
A DIR (Development Inhibitor Releasing) coupler releases an inhibitor as it develops, and that inhibitor diffuses sideways, suppressing development in the neighbouring area.
The consequence: at the border between a bright area and a dark one, the bright side gets suppressed harder by its neighbours while the dark side barely gets suppressed at all — so the density difference across the boundary is amplified. That is the edge effect (or adjacency effect), and it reads to the eye as sharpness.
There is a second job: the inhibitor can diffuse between layers and suppress development in the others, which reduces interlayer crosstalk and improves colour separation. DIR is doing sharpness and saturation at once.
Simulating it with ordinary unsharp mask does not work. USM is symmetrical and density-independent; the edge effect is directional and tied to local exposure. This is why "add sharpening" and "look like film" have never been the same operation.
Halation and grain: two things a flat filter cannot fake
Halation: light passes through the emulsion, hits the base, and part of it reflects back from the far surface to expose a second time around the original highlight. Red light penetrates furthest and is absorbed least, so halation usually reads red — that ring around neon signs at night, that glow on a backlit outline.
Film normally carries an antihalation backing to suppress it, which dissolves away during processing. How hard it is suppressed is a design trade-off, which is why different stocks have distinctly different "red halo" personalities.
Grain is the random clumping of developed silver (or dye clouds), its size and distribution inherited from the emulsion crystals themselves. It is not uniform Gaussian noise: shadows and highlights do not look equally grainy, and the physical size of a clump has nothing to do with enlargement. Print bigger and you magnify the grain along with the picture — you do not create more of it.
Halation, grain and the edge effect are all luminance-dependent. Implement any of them as a global, luminance-blind filter and you have copied its shape, not its behaviour.
Those two filters in the enlarger head — and pre-flashing
Most black-and-white printing today is done on variable-contrast (multigrade) paper: one sheet carrying two emulsions of different contrast — a high-contrast one sensitive to blue, a low-contrast one sensitive to green.
The dichroic filters in the enlarger head — magenta and yellow — exist to set the ratio between those two kinds of light. More magenta = more blue = higher contrast (higher grade); more yellow = more green = lower contrast (lower grade). A "grade" is fundamentally a mixing ratio, not a different sheet of paper.
Filters block light, so exposure has to be compensated. Manufacturers publish the compensation per grade — and successive generations of the same family (Ilford MG IV versus MG V, for instance) do not share the same figures. Use the table for the paper in your hand, and do not mix the two in your head.
Where these terms live inside Optaxis
Negative Lab — "sample film base" is there to deal with the orange mask (§04); per-channel black point, white point and gamma are you rebuilding that characteristic curve by hand (§02). Black-and-white film has no mask, so the chain is shorter and it usually lands first try.
Film Simulation — "highlight rolloff" is the shoulder (§02); shadow tint and warmth are weighted by luminance precisely because these effects are luminance-dependent (§06) rather than a global shift.
Darkroom tools — the multigrade filter compensation is the table from §07; test strips and burning / dodging are you locating points on the curve.
Reciprocity tool — on long exposures film's response departs from the characteristic curve and needs extra time. That is §02's curve failing at very low illuminance.
Learn these words and those three sliders stop being trial and error. You are bending a curve that chemistry already bent.
Further reading: to see these mechanisms actually computed, look at the open-source project spektrafilm (github.com/andreavolpato/spektrafilm, GPLv3) and its author's long write-up on the PIXLS.US forum — it simulates the full exposure → development → enlargement → scan chain from manufacturer datasheet spectra. This article covers public-domain concepts only and contains none of that project's code or data.