Character error rate: word error rate computed on characters

Character error rate is word error rate computed on characters instead of words. Autrans' language page figures are word-based.

Character error rate is word error rate computed on characters instead of words: substitutions, deletions and insertions of characters divided by the number of characters in the reference. It is gentler than WER on a near-miss spelling, and it shows whether errors are small slips or wrong words.

Also searched as "cer".

What the sources say about character error rate

  • jitsi.github.io, jiwer. jiwer computes WER and CER from minimum edit distance between reference and hypothesis; for an empty reference each hypothesis character counts as an insertion.
  • lightning.ai, char error rate. The character score is (S + D + I) / N over characters, with N the characters in the reference; lower is better and 0 is perfect. Read on 22 October 2026.

Character error rate in Autrans

Autrans' language page figures are word-based. CER matters for languages whose spelling depends on marks and special letters (tone marks in Yoruba and Igbo, special letters in Hausa and Twi), where a missing mark is a single-character slip but a whole wrong word under WER. Autrans' own app evaluation scripts test diacritic restoration on Yoruba and Igbo outputs.

A worked example

Reference: 'ọkọ' (3 characters, with the dot-below letter ọ stored as one character, Unicode U+1ECD). Transcript: 'oko'. Two of three characters are substituted, so the character score is 2 divided by 3, about 67 in 100. The same single word is wrong, so the word score is 1 divided by 1, a whole word. If the text is stored with a combining dot instead (NFD), the reference is 5 code points and the same slip counts as 2 deletions, the character score is 2 divided by 5, 40 in 100: normalise text before scoring.

Terms next to character error rate

  • Word error rate: Word error rate is the standard score for a speech recognition transcript.
  • Diacritics: A diacritic is a mark added to a letter, such as an acute accent, a grave accent, a tilde or a dot below, that changes how the letter is read.
  • Orthography: Orthography is the set of conventions for writing a language: spelling, punctuation, word division, capital letters, hyphenation and the marking of emphasis.
  • WER vs CER: Both scores count substitutions, deletions and insertions against a reference and divide by its length.

Questions people ask

What does character error rate mean?

Character error rate is word error rate computed on characters instead of words: substitutions, deletions and insertions of characters divided by the number of characters in the reference. It is gentler than WER on a near-miss spelling, and it shows whether errors are small slips or wrong words.

How does character error rate show up in Autrans?

Autrans' language page figures are word-based. CER matters for languages whose spelling depends on marks and special letters (tone marks in Yoruba and Igbo, special letters in Hausa and Twi), where a missing mark is a single-character slip but a whole wrong word under WER.

What is the difference between WER and CER?

Both scores count substitutions, deletions and insertions against a reference and divide by its length. The full answer, with sources, is on the page for WER vs CER.

Where does this definition of character error rate come from?

It rests on jitsi.github.io, jiwer and lightning.ai, char error rate. What each source supports is listed above, next to its link.

Can you give an example?

Reference: 'ọkọ' (3 characters, with the dot-below letter ọ stored as one character, Unicode U+1ECD). Transcript: 'oko'. Two of three characters are substituted, so the character score is 2 divided by 3, about 67 in 100. The same single word is wrong, so the word score is 1 divided by 1, a whole word. If the text is stored with a combining dot instead (NFD), the reference is 5 code points and the same slip counts as 2 deletions, the character score is 2 divided by 5, 40 in 100: normalise text before scoring.

Related

Start transcribing free

Send Autrans into your next meeting, use it in the room, or upload a recording. Your first 30 minutes are free, given once, and no credit card is needed to start.

Last checked 22 October 2026. How we measure accuracy. Spotted a mistake? Write to hello@autrans.online.