companding

Companding is a signal processing technique that reduces noise by compressing the signal at the transmitter and expanding it at the receiver. The name, a contraction of compression-expansion, describes the complete cycle. In energy and utilities applications, particularly in remote telemetry and SCADA systems, companding protects weak signals traveling long distances over noisy channels, whether power line carriers, radio links, or fiber optics. A weak signal arriving at a substation or control center can be reconstructed with much of its noise filtered out, improving the signal-to-noise ratio without requiring more transmission power.

The mechanics are straightforward: a compressor at the transmit end amplifies soft signal details while limiting peaks, creating a non-linear scaling. Loud portions are suppressed; quiet portions are boosted. This compressed signal is transmitted as-is. At the receive end, an expander applies the inverse transformation, restoring the original signal's dynamic range while leaving noise largely behind. Because noise is typically added uniformly across the signal path, the compander preserves wanted signal information while attenuating the noise that was added during transmission.

Practical variants and applications

Companding schemes differ mainly in their compression law. Mu-law (used in North America and Japan) and A-law (used in Europe and most of the world) are the dominant standards in telecommunications and utility communications. Mu-law uses a logarithmic curve with a mu value of 255; A-law uses a piecewise linear approximation. Both were standardized by the ITU to allow interoperability between equipment manufacturers. In analog power line carrier systems, companding improves intelligibility of voice and data signaling at frequencies between 9 kHz and 500 kHz. Digital implementations apply the compression curve to sampled signals before transmission, then the expander after reception.

Problems arise when companders are mismatched: if a receiver uses the wrong expansion law for the compression applied, the output is distorted and unintelligible. System designers must ensure transmitter and receiver use the same standard. Companding also works best when the signal is weak relative to noise; if the signal is already clean, companding adds unnecessary processing and slight distortion. In modern digital systems with forward error correction and higher bandwidth availability, some legacy companding has been replaced by simpler noise-filtering schemes, though companding remains embedded in many older utility communication networks.

Companding differs from simple amplification because it is nonlinear and selective. Straightforward gain boost would amplify noise equally with signal. Companding trades peak fidelity for overall intelligibility, a worthwhile trade in noisy utility telemetry where the goal is to detect state changes and fault conditions, not to preserve waveform perfection. The technique is also distinct from data compression; companding does not reduce bandwidth or transmission rate, only improves signal quality in the presence of noise.