Error Detection and Correction·Lesson 2 of 9
Types of Errors
Two Ways Bits Get Corrupted
Data travels as a stream of — 0s and 1s — and noise on the line can flip some of them, so a 0 arrives as a 1 or the other way round. When we talk about the shape that corruption takes, there are just two kinds to know: a single-bit error, where exactly one bit flips, and a burst error, where two or more bits flip within a short stretch. Almost every question about transmission errors comes down to telling these two apart.
The difference in one line
A single-bit error changes exactly one bit. A burst error changes two or more bits, and we measure its length from the first corrupted bit to the last.
A typo vs a smudge
Imagine a handwritten note. A single-bit error is like one letter written wrong — 'cat' becomes 'cot'. A burst error is like a coffee splash that smears a whole run of letters at once. One is a tiny slip in a single spot; the other is damage across a stretch of the message.
Single-bit Error
A single-bit error means only one bit in the whole data unit changes value, and every other bit arrives perfectly. For example, if the sender transmits 00001010 and the receiver reads 00001110, exactly one bit (the fourth from the right) flipped from 0 to 1. This kind of error is actually quite rare in normal serial transmission, because noise usually lasts long enough to disturb several neighbouring bits rather than just one.
Single-bit errors are more likely in parallel transmission, where several bits travel at the same instant along separate wires — a glitch on just one of those wires flips just one bit. We look at this case in detail in its own section.
Burst Error
A burst error means two or more bits in the data unit are corrupted. The key detail is how we measure it: the length of a burst is counted from the first corrupted bit to the last corrupted bit, and — importantly — some bits in between may actually be fine. Burst errors are the common case in real networks, because a spell of noise almost always spans several bit periods at once. This is the error type we most need to defend against.
Comparing the Two
Here's how the two types stack up side by side — how many bits they touch, what tends to cause them, and how often they show up in practice:
| Single-bit error | Burst error | |
|---|---|---|
| Bits affected | Exactly one | Two or more |
| Measured by | Nothing — it's a single bit | Length from first to last corrupted bit |
| Typical cause | A very brief glitch on one bit | Noise lasting several bit periods |
| How common | Rare (mainly parallel links) | Common in real transmission |
| Hardest for | Almost any check catches it | Weaker checks like simple parity |
Why the Distinction Matters
Different detection methods are good at different things, and the error type is exactly why. A single bit reliably catches any single-bit error, but it can be fooled by a burst that flips an even number of bits. A , on the other hand, is designed specifically to catch bursts. So knowing which error you're up against explains why one topic has several tools instead of just one.
“"In a burst error, every bit in the affected stretch is flipped."”
Q:How is the length of a burst error measured?
A: From the first corrupted bit to the last corrupted bit, inclusive. If bit 3 and bit 8 are wrong, the burst length is 6 — even if some of the bits between them (4, 5, 6, 7) were received correctly. It's a measure of the span of the disturbance, not the number of bits that changed.
Quick Revision Cheat Sheet
Single-bit error: Exactly one bit flips; the rest are correct
Burst error: Two or more bits flip within a stretch
Burst length: From the first corrupted bit to the last
Middle bits: Inside a burst, some may still be correct
More common: Burst errors — noise usually spans several bits