Bell Telephone Laboratories · conceptual instrument 1948/2026

The Noisy Channel

A playable Claude Shannon information-theory lab. Type a message, measure its uncertainty, compress it, add redundancy, throw it through noise, and ask the receiver to reconstruct what you meant.

MODEL 48 · INFORMATION TRANSMISSION BENCH CARRIER READY

INFORMATION
SOURCE

→

SOURCE
ENCODER

→

REDUNDANT
ENCODER

→

NOISY
CHANNEL

→

DECODER /
DESTINATION

CHANNEL MONITOR · transmitted bits moving through uncertaintyREADY

SOURCE → ENCODE → CHANNEL → RECOVERTIME →

The moving trace is a conceptual information monitor, not a literal electrical waveform. Corrupted bits flash red as the channel flips them.

Source entropy—

Source compression—

Channel capacity—

Rate vs capacity—

Bits put on wire—

Bits flipped—

Bits repaired—

Message fidelity—

Transmitted code stream

Received code stream · red = channel flip

Destination reconstruction

Shannon moment

Source symbol distribution—

Entropy measures average uncertainty per source symbol. A source that repeats the same character is extremely predictable. A broad, balanced symbol distribution carries more surprise.

The channel has room.Move the bandwidth, S/N, and requested-rate controls to discover Shannon's boundary.

0capacity

What this instrument is showing: Shannon separated the semantic meaning of a message from the engineering problem of communicating a selected symbol reliably. Entropy describes uncertainty in the source. Source coding exploits predictability. Redundancy can make errors correctable. Channel capacity sets a theoretical ceiling for reliable communication at a given bandwidth and signal-to-noise ratio. This page uses a binary-error toy channel and simple repetition coding for intuition; it is not a full modern communications stack.