Definition
VideoText editorial team. Updated September 13, 2026.
Word error rate (WER) is the standard research metric for speech recognition. It counts substitutions, deletions, and insertions against a reference transcript, then divides by the number of words in that reference.
The sections below explain how word error rate is used in transcription and subtitle work, what people often mix up, and which VideoText workflow applies when the next step is a file or review pass.
Align the hypothesis to the reference with an edit distance. Each wrong word, missing word, or extra word is an error. Divide by N, the reference word count.
It lets researchers compare systems on the same audio and the same reference. The Whisper paper reports many WERs so readers can see robustness, not just LibriSpeech Clean.
Converting WER to “percent accurate” (100 − WER) hides insertions and treats every word as equal. A 2.5% WER on clean audiobooks does not mean 97.5% accurate captions on a Zoom call.
WER weights every error equally. Mishearing "the" as "a" and mistranscribing a drug name both count as one substitution, although only one of them matters to a reader.
It is also uncapped: a system that inserts a long hallucinated passage can score above 100%, because insertions are counted against a reference that never contained them.
Before scoring, both transcripts are usually normalised — case folded, punctuation stripped, numbers written consistently. Different normalisation rules produce materially different WER on identical text.
This is why a quoted WER without its test set and its normalisation is not comparable to another. Two vendors can report different numbers on the same audio purely through methodology.