Step 1–2: Reconstructing Light

Reality: Film doesn’t see “RGB.” It reacts to the full spectrum of light (360–780nm).
Simulation: We reverse-engineer your digital RGB into a 43-band spectrum using Non-Negative Least Squares (NNLS). This ensures physically correct light distributions.

👉 Why it matters: Different films “see” different parts of the spectrum. This step is why Fuji, Kodak, and Agfa stocks look distinct.

Step 3: Film Sensitivity

Reality: Film has three stacked layers (red, green, blue), each with unique sensitivity.
Simulation: We model these as Gaussian sensitivity curves across the spectrum. Peaks and widths match the stock you select.

👉 Why it matters: Sensitivity defines a film’s “color signature.” Kodak Portra green isn’t the same green as Fuji 400H.

Step 4: Exposure & Activation

We integrate the spectrum with sensitivity curves, then apply a sigmoid (logistic) response to mimic how emulsions activate with light.

👉 Why it matters: This is why film has graceful highlight rolloff and rich shadows, instead of the linear clipping you get from digital sensors.

Step 5: Latent Image Formation

We compute the invisible latent image — the raw “exposure memory” of the film.

👉 Why it matters: This invisible stage determines how much dye forms during development, just like in a real negative.

Step 6: Development (H-D Curves)

Reality: In development, activated sites turn into visible dye. The relationship follows an H-D curve with toe (shadows) and shoulder (highlights).

Simulation: We implement adjustable H-D curves with toe/shoulder shaping.

Push/Pull Development (Single Slider):

• Pull (–) → flatter curve, muted saturation, more shadow detail.

• Push (+) → steeper curve, denser colors, stronger contrast.

👉 Why it matters: You can reproduce the classic darkroom technique of push/pull processing with one intuitive control.

Step 6.1: Bleach Bypass

Leaves silver in the negative for harsher contrast and muted colors.

Step 6.5: Film Grain

Grain is modeled as stochastic silver particle clumping, scaled by ISO, grain size, and image content. Each emulsion has it’s own unique fingerprint and responds uniquely to the settings.

👉 Why it matters: Unlike digital “grain overlays,” this responds organically to shadows and highlights.

Step 7: Dye Formation

Dyes (cyan, magenta, yellow) are simulated as asymmetric Gaussian absorption curves, matched to real film chemistry.

👉 Why it matters: Real film dyes don’t absorb symmetrically — cyan has sharp cutoffs, magenta rolls off slowly, etc. We capture this.

Step 7.1: Interimage Effects

This is where magic happens. Dyes interact during development, creating unique non-linear responses:

• Midtone Expansion – richer midtones

• Developer Competition – subtle saturation boost

• Chroma Expansion – better color separation (like Technicolor)

• Dye Inhibition – prevents colors from collapsing into mud

👉 Why it matters: Digital cameras can’t reproduce this naturally. These interactions are why film color looks so “alive.”

Step 8: Scanner Simulation

We simulate a real film scanner using Beer-Lambert law for light transmission and CIE Status M observer functions.

👉 Why it matters: Ensures the simulation ends in a scanned negative that behaves like actual processed film.

Step 9–16: Print Stage

The print stage re-runs the entire pipeline with different film characteristics, just like making a print from a negative in the darkroom.