8.9 Transmission Error

A transmission error is a discrepancy between a transmitted symbol, bit, or message and the corresponding received symbol, bit, or message, caused by the action of noise, interference, distortion, or other channel impairments on the signal during its passage through the communication channel. Transmission errors are the fundamental mechanism by which channel noise translates into informational corruption: the physical perturbations that noise introduces to the signal are manifested at the symbol decision stage as incorrect decisions about which symbol was sent, and these incorrect symbol decisions propagate into the decoded bit stream as erroneous bits. The statistical characterization of transmission errors—their rate, their distribution in time, their dependence on channel conditions—determines the reliability of a communication system and guides the design of error mitigation strategies.

The most fundamental measure of transmission error is the bit error rate (BER), defined as the probability that a transmitted bit is decoded incorrectly at the receiver:

For a binary phase shift keying (BPSK) modulated system in additive white Gaussian noise (AWGN), the BER is given exactly by the Q-function of the signal-to-noise ratio per bit:

where E_b is the energy per bit and N_0 is the noise spectral density. The Q-function decreases rapidly as E_b/N_0 increases, so the BER falls steeply with increasing SNR: doubling the received power (adding 3 dB of SNR) dramatically reduces the error rate. For most digital communication systems, target BERs range from 10⁻³ for voice systems (where occasional errors are tolerable) to 10⁻⁹ or below for high-integrity data communication systems where uncorrected errors would corrupt data or trigger safety-critical failures.

Transmission errors are classified by their temporal distribution as either random errors or burst errors. Random errors arise from AWGN and other memoryless noise sources: each bit has an independent, equal probability of error, and errors are distributed uniformly in time. Burst errors arise from impulsive noise sources (lightning, switching transients, Rayleigh fading in mobile channels) that produce correlated sequences of errors: when a burst event occurs, many consecutive bits may be corrupted, followed by long error-free periods. The distinction between random and burst error statistics is important for error correction code design, because codes designed for random errors (like Reed-Muller codes) perform differently against burst errors than codes designed specifically for burst error patterns (like Fire codes, or Reed-Solomon codes in block form).

Cyclic redundancy check (CRC) codes are among the most widely used tools for transmission error detection. A CRC encoder computes a remainder by dividing the message polynomial by a generator polynomial over a binary field, and appends the remainder as a checksum to the transmitted message. The receiver performs the same division on the received message-plus-checksum; if the remainder is zero, no error (at least within the code's detection capability) has occurred; if the remainder is nonzero, one or more transmission errors have been detected. The generator polynomial is chosen to maximize the set of detectable error patterns: the standard CRC-32 polynomial used in Ethernet and many file systems can detect all single-bit errors, all burst errors of length 32 or less, and a large fraction of longer burst errors.

Automatic repeat request (ARQ) protocols use error detection codes to enable reliable transmission over noisy channels by requesting retransmission of errored frames. In stop-and-wait ARQ, the transmitter sends one frame and waits for an acknowledgment before sending the next; if no acknowledgment is received within a timeout period (or a negative acknowledgment is received), the frame is retransmitted. Selective repeat ARQ allows the transmitter to continue sending subsequent frames while waiting for retransmission of only the specifically errored frames, improving efficiency in high-delay channels. The throughput efficiency of ARQ decreases as the BER increases, because more retransmissions are required—at high error rates, the channel is consumed almost entirely by retransmissions and little new information is delivered.

Forward error correction (FEC) codes add redundant parity information to the transmitted message that allows the receiver to correct transmission errors without requesting retransmission. The Hamming distance of a code—the minimum number of bit positions in which any two valid codewords differ—determines the error correction capability: a code with minimum Hamming distance d_min can detect up to d_min − 1 errors and correct up to ⌊(d_min − 1)/2⌋ errors per codeword. Modern capacity-approaching codes like turbo codes and LDPC codes can operate within a fraction of a dB of the Shannon capacity limit, achieving very low BERs even in channels with substantial noise by encoding the information across many symbols and using iterative decoding algorithms that progressively refine bit estimates by propagating soft probability information through a graphical representation of the code.

In human communication, transmission errors manifest as mishearings, misreadings, and miscommunications in which the received message differs from the sent message at the phonemic, lexical, or syntactic level. Acoustic transmission errors in speech—hearing one phoneme as another due to acoustic similarity—are common in noisy environments and are partially corrected by the brain's use of contextual information and the redundancy of natural language. Written transmission errors in handwriting, optical character recognition, and digital text transmission include character substitutions (reading a 0 as an O, or an l as a 1), omissions (skipping a character or word), insertions (duplicating a character), and transpositions (reversing the order of adjacent characters). Error-detection strategies in human communication—asking for repetition, reading back a message, using spelling alphabets (Alpha, Bravo, Charlie) for critical identifiers—are behavioral analogues of the technical ARQ and CRC mechanisms that manage transmission errors in engineered systems.

The economic cost of transmission errors varies enormously with the application. In casual conversation, occasional word-level mishearings are inconsequential and self-correcting. In medical communication, a transmission error in a drug name, dosage, or patient identifier can have fatal consequences. In financial communication, an error in a securities transaction value can produce massive unintended positions. In navigation and aviation, position or heading errors transmitted with noise can cause collisions and crashes. The appropriate level of error protection—the amount of redundancy, the sophistication of error detection and correction, the stringency of confirmation protocols—should be calibrated to the consequence severity of uncorrected transmission errors in the application, not to some universal standard. Applications with catastrophic error consequences invest heavily in multiple redundant transmission paths, end-to-end error detection, positive confirmation of critical fields, and procedural protocols that make transmission errors detectable before they have consequences.