Adaptive Surface Deduction & 120Hz Fallback

Refraction delivers crystalline aesthetics, but during 120Hz (ProMotion) window dragging, heavy displacement shaders must gracefully degrade to guarantee zero dropped frames.

Why Runtime Auto Deduction is Essential

Client runtimes exhibit extreme volatility:

• The user may suddenly drag a window across monitors at high velocity.
• The operating system may throttle clocks under Low Power Mode.
• Integrated GPUs may hit memory bandwidth bottlenecks during rapid renders.

Forcing developers to sprinkle manual checks like if (isDragging) across views invites bugs. TuffEx abstracts this into the autoDetect perception engine.

The autoDetect Sensor Matrix

The system continuously tracks runtime physical signals:

1. Pointer Velocity Sensing: Samples positions $\vec{s}(t)$ to compute instantaneous velocity $|\vec{v}| = \frac{|\Delta \vec{s}|}{\Delta t}$. Sustained speeds exceeding $1.2\text{px/ms}$ trigger the motion state.
2. Window Move / Resize Events: Hooks Electron main-process native window movement events, dispatching immediate fallback signals.
3. Frame Budget Probes: Monitors composite duration; frames exceeding 8ms trigger lightweight fallbacks.

Instant 0ms Fallback & 140ms Asymmetric Recovery

TuffEx applies an asymmetric clock transition:

• Fallback (0ms Instantaneous): As soon as rapid motion is detected, heavy shaders drop to an ultra-lightweight monochrome mask within one frame. GPU load drops immediately, sustaining full 120Hz tracking.
• Recovery (140ms Damped Hysteresis): When motion ceases, the engine does not switch back immediately (which would cause chatter if pointer jitter occurs). It waits 140ms to confirm stability, then smoothly restores refraction.

State Machine Diagram

This architecture harmonizes high-fidelity material simulation with unflinching desktop responsiveness.