Start with the engineering

Before treating a strange sound on a phone call as anything unusual, it is worth knowing how many ordinary causes exist. Telephone networks are complicated, and early digital networks in particular produced noises that surprised even the people who used them every day.

A careful reading of any anomalous-call account begins by asking which of these explanations could produce the reported sound, and which details they cannot account for.

The GSM “buzz-click”

Second-generation GSM networks sent data in short timed bursts, using a technique called time-division multiple access. The burst pattern repeats at about 217 times per second, which falls inside the range of human hearing.

Unshielded audio equipment near an active mobile phone can pick up those pulses and turn them into a rhythmic buzz or click. Many people have heard it from speakers, radios and analog phones just before a mobile phone rings. It is a well-documented, repeatable effect.

Crosstalk and routing faults

Faults in switching equipment can let the audio of one call leak into another. A listener may then hear faint voices, tones or fragments of conversation that belong to someone else.

Because the leaked audio is often distorted and out of context, it can sound unlike any ordinary conversation, especially when a few words are heard without the rest.

Bit errors and narrow bandwidth

Digital voice is compressed and sent as packets of bits. When bits are lost or damaged, the decoded speech can turn robotic, warbling or metallic. Older equipment that narrows the audio range can add a thin, tinny quality as well.

In a 2007 article on possible causes of metallic voices on telephone calls, Ted Twietmeyer described these categories of technical fault and argued that they should be considered before anything more exotic. That is a sensible rule for any investigation.

What these explanations do not cover

Technical causes account well for odd audio. They are weaker for other details that some reports include: an origin number that is invalid or missing, calls that repeat on a fixed schedule, or several people independently describing the same sounds.

Noting that difference is not a claim that something extraordinary happened. It is a way of keeping the explanation matched to the evidence, and of recording clearly which parts of a report remain unexplained.