About

Keyboard Ghosting is a hardware limitation where simultaneous key presses are lost or misregistered due to the internal wiring matrix of the device. This phenomenon is critical for high-performance activities like competitive gaming or stenography.

Standard membrane keyboards often suffer from 2-Key Rollover (2KRO) or 6-Key Rollover (6KRO), meaning they can only reliably register a limited number of keys in specific zones. This tool visualizes your keyboard's signal processing in real-time, allowing you to identify dead zones (Ghosting) or confirm full N-Key Rollover (NKRO) capabilities.

Formulas

The Polling Rate (R) is estimated by measuring the time delta (Δt) between consecutive input interrupts processed by the browser.

{

R ≈ 1000Δt Hz

Where Δt is the rolling average of time in milliseconds between keydown events.

Reference Data

Keyboard Type	Switch Mechanism	Typical Rollover (KRO)	Ghosting Risk	Polling Rate
Standard Membrane	Rubber Dome over Membrane Sheet	2 - 6	HIGH	125 Hz
Gaming Membrane	Optimized Matrix	10 - 24	MODERATE	500 Hz
Mechanical (USB)	Physical Switch (Cherry/Gateron)	6 - ∞ (NKRO)	LOW	1000 Hz
Optical / Hall Effect	Light / Magnetic Actuation	∞ (NKRO)	NONE	4000 Hz+
PS/2 Legacy	Interrupt-based Hardware	∞ (NKRO)	NONE	Interrupt

Frequently Asked Questions

Ghosting technically refers to the registration of a key that *wasn't* pressed (a "ghost" signal). Jamming (or Masking) is when a key you *did* press is ignored. Most people use "Ghosting" to refer to both phenomena.

This is a common matrix limitation on non-gaming keyboards. These keys often share a column/row trace on the circuit board. Without anti-ghosting diodes, the controller cannot distinguish the third keypress.

NKRO stands for N-Key Rollover. It means the keyboard can correctly register an unlimited number of simultaneous key presses (N = Infinite). This is standard on high-end mechanical keyboards.

No. Browser-based tools measure "Input-to-Paint" latency. It includes USB polling, OS drivers, Browser event loop, and Frame rendering. It is useful for relative comparison but not absolute hardware benchmarking.