How Grain Forms
Grain is the part of the film image a LUT cannot carry. Color and contrast can be sampled into a cube; grain cannot, because it is spatial and it changes every frame. It is also not a layer added on top of a clean picture. It is the structure of the picture itself, seen close enough. Once you know how it forms you can read what the Grain tab is doing.
Grain is the emulsion
A film emulsion is silver-halide crystals suspended in gelatin, coated thin on the base. Kodak describes those crystals as coated in a deliberate range of sizes, because size is one of the things that sets how much light a crystal needs before it will record anything. Exposure does not fill a crystal part way. Kodak describes development as an all-or-nothing event: a crystal that caught enough light develops completely, and one that did not develop stays clear and is washed away. So the image is built from discrete sites that either fired or did not, and the texture of those sites is the grain. There is no separate grain to add; it is what the emulsion is made of.
Why color grain is a dye cloud
In a black-and-white film the developed crystals stay as metallic silver, and the grain is silver. A color negative ends up somewhere else. As the negative essay describes, development builds a dye alongside each developed crystal, and then the silver is bleached and fixed away. What is left where the crystal sat is a small cloud of dye, larger where the crystal was larger and tighter where it was smaller. That is why color grain is soft, a cloud with a fuzzy edge, rather than the hard speck of a silver image. It also means the grain you see is made of the same dye that carries the color, not a gray noise sitting over it.
Kodak also corrects a common belief: developed grains sit essentially where the exposed crystals sat, with little migration or physical joining. The clumped look grain sometimes has is largely the eye grouping a random field rather than particles that actually joined.
Where grain lives in the tone
Grain is not spread evenly across the brightness of a scene. Because the largest, most sensitive crystals are the ones that respond first, Kodak locates the coarsest apparent graininess in the shadows and the underexposed parts of the frame, where only those big fast crystals have fired. The highlights are the opposite. Lab accounts describe the dye clouds in a dense highlight overlapping through the depth of the layer until they average out, so a bright, well-exposed highlight is the quietest part of the image. Grain reads most through the shadows and midtones and thins toward the top of the range.
It also differs by layer. In published granularity figures the blue record is by a wide margin the grainiest of the three, which is part of why color grain has a character and is not neutral. None of this comes from a single number for a stock; the grain a stock shows depends on where you are in the tone and which layer you are looking at.
Why it moves
Which crystals catch enough light is a matter of chance at the scale of a single grain, so the exact pattern is different on every frame of a shot. Colorists describe this as the grain rerolling frame to frame, and it is what makes real grain read as alive rather than as a fixed pattern laid over the picture. Kodak makes the related point that at playback the eye gathers detail spread across successive frames that no single frame holds. Both are reasons to judge grain in motion, at delivery size; a paused, enlarged frame can misrepresent how it plays.
What the stock decides
Grain is stock-specific. The old rule is that a faster stock is a grainier one, because speed came from larger crystals, and larger crystals leave larger clouds. Kodak’s T-Grain technology complicated that: by growing crystals flat instead of chunky, it let a stock gather more light without a bigger footprint, so a modern high-speed stock is far less grainy than its speed alone would predict. Grain also is not one thing across a whole film chain. Colorists note that a negative’s grain and the grain of the print struck from it are different structures, and that what reaches a screen is the two combined. And apparent grain depends on format: a larger negative packs more crystals into a frame, so its grain is finer, while blowing a small format up magnifies whatever grain it has.
In the plugin
The Grain stage runs last, after the print and after halation, applying its texture to the finished color image to stand in for the emulsion structure, and the Overview shows where it falls in the chain. It carries its own Film Stock and Film Gauge, separate from the Negative tab, and weights the grain by the brightness of each pixel so the shadows carry more than the highlights, matching the tonal behavior above. The per-layer weights set each layer’s grain strength, and Layer Separation sets how independent the three layers’ patterns are; together they turn near-monochrome grain into the colored grain of a scanned negative. Silver Halide Mode switches from dye clouds to sharp silver, for a black-and-white or bleach-bypassed image where the silver is still in the film. Scanner noise is kept apart from grain, because it is electronic and not photographic. The Grain guide covers all of it. The stage does not model how development changes grain: Push/Pull reshapes the negative’s tone and the grain follows the image it receives, but no direct link runs from a push to the grain itself, so you match a heavier push by raising the grain yourself.
The account of crystal sizes, all-or-nothing development, the dye-cloud structure of color grain, its shadow weighting, and T-Grain is drawn from Kodak’s How Film Makes an Image, Color Films, and the Essential Reference Guide. The frame-to-frame character, the blue-record weighting, the negative-and-print combination, and the format scaling come from colorists who profile film rather than from a manufacturer.
Next: Why Photochemical returns to where the chain begins, now that you have followed it end to end.