Every pixel becomes a body with four parts: Pattern, Narrative, Behavior, Adaptation (PNBA). Its brightness is its load. Black sits at rest (Noble, τ = 0) and pure white sits exactly at the Torsion Limit (TL).
Bright bodies rise and dark ones stay low, so the picture becomes a landscape you can turn, sculpt and export as a 3D model.
A pixel that hits pure white has no detail left. It reads Shatter (red): a blowout. Autotune pulls those bodies back under the limit.
Load several bands of the same scene (ultraviolet, visible, infrared, or the R, G and B of a photo) and IMagine stacks them. That is broadband phase mapping: finer transitions and far fewer spikes.
Band floor and ceiling. Display: 0 and 1. Sky: sigma-clipped sky median − 2σ, and the 99.9th percentile. Percentile: 0.5th and 99.5th. Values below the floor are Noble; values at the ceiling reach TL.
Stretch. Linear, or asinh with softening 3σ: x = asinh((v − floor)/3σ) / asinh((ceiling − floor)/3σ).
B (Behavior) = Y, the weighted stack of normalized bands after adjustment. τ = TL × Y / Y_white: the critical value is the white point, so white sits at TL. Noble: τ = 0. Locked: 0 < τ < TL_IVA. IVA Peak: TL_IVA ≤ τ < TL. Shatter: τ ≥ TL.
P (Pattern) = Sobel gradient magnitude. N (Narrative) = 3 × 3 mean. A (Adaptation) = 3 × 3 standard deviation. Blowout regions are 4-connected Shatter components. TL = 0.136899099984016; Ω₀ = TL × 10.