Running the Reality Kernel to act on a scene: make it look like a target, through the physical channel.
P.I.G.M.I.E. Filing 1 · patent pendingThe Reality Transform is the controllable-rendering regime of the Reality Kernel. A rendering regime acts on a scene rather than only observing it. Its objective is simple: make the scene look like a target.
The regime projects adaptive light onto people, objects, or buildings, a son-et-lumiere mapping. The projection is constrained to the physical channel, meaning the optics, surface response, lighting, sensors, and environment through which the signal passes.
In some embodiments, the loop minimises divergence between the achieved output and a target. Divergence is the measured difference between those two states. The optimisation is subject to meter envelopes, which are operating limits for brightness, colour, exposure, and related quantities. Sensor feedback continuously refines brightness, colour, and pattern.
The Reality Transform shares the Reality Kernel front-end and convolution-bundle record with the verification and perception regimes. The front-end is the sensing and projection interface. The convolution-bundle record is the joint, time-ordered record of what was emitted and what was observed. This makes a rendering run auditable. It also allows regimes to be interleaved, for example sense first, then render to match.
The earliest rendering loops were single-shot loops. Each cycle learned from its own previous output, so any error became part of the next target and was amplified across cycles. The loop could drift away from the target.
The operating envelope depended on style, environment, and training data.
One route to stability was multiplexed paired acquisition, disclosed in the parent patent. Multiplexing means placing more than one signal on the same scene in a coordinated acquisition. A neutral probe emission and a styled emission are projected onto the scene and read together as a stereo pair.
The neutral probe emission is a reference illumination. The styled emission is the target rendering. The paired acquisition recovers two captures bound to the same scene instant: one neutrally-lit read and one projected-style read.
This re-anchors each cycle on an independent neutral probe. The loop does not have to learn only from its own previous rendered output. This reduces the compounding-drift failure mode of the single-shot loops.
With edge-preserving style transfer, the paired captures form a well-conditioned training corpus from the physical pairs. Edge-preserving means that scene boundaries and object structure are retained while style is applied. The dataset is the deterministic part: it falls out of the multiplexed capture. Agents trained on it are only stochastically stable. They are usually steady, sometimes drifting, and occasionally settle into resonances. They are not inherently stable.
The implementation used a slow-teacher, fast-student pattern. A slow teacher is an expensive styler that generates emissions offline. Those emissions are projected onto a stable scene and captured through the physical channel.
A fast student is a runtime model trained to reproduce the captured physical response. It learns the channel transfer function, meaning the relation between projected signal and observed result.
The styler training data was mixed. It included single-shot loops, the paired stereo corpus, and modified neutral projections captured through the physical channel. Some data could be screened by a verisimilitude check, meaning a check that the result remains visually plausible for the target condition.
The working stable demonstration from the parent-patent era used a basic, edge-preserving fast style-transfer student. The available compute scoped it to simpler styles.
Later experiments reached richer styles. They learned the scene's inverse response with pix2pixHD and used human-body pose-conditioning through OpenPose ControlNet, guided through ComfyUI and Stable Diffusion. These experiments ran as a single-step slow loop with weaker stability. Probe textures were drawn from the Describable Textures Dataset.
Both lines are part of the lineage into the current Reality Kernel formulation, which generalises them and absorbs their lessons.
The Reality Transform was built and shown as a stable rendering loop in the parent-patent era. That stability was stochastic, not inherent. Richer styles were explored with weaker stability.
The regime is described and enabled in the filings as part of the Reality Kernel. The separate digital Truth Beam verification regime is the single demonstrated, recomputable instance of the whole project.
The Reality Kernel · the apparatus and formalism.
Regimes · the three objectives.
Limager · the perception regime.
truthbeam.com · Truth Beam, the demonstrated verification instance.