Preprint
Disfluencies and Human Speech Transcription Errors
Vicky Zayats et al., arXiv, 2019
Read the originalRead on October 17, 2026
- What it is
- A 2019 paper on where human transcribers make errors in spontaneous speech, with a review of published human error rates for telephone conversation and other genres.
- How much weight it carries
- Preprint from an established speech group. It reviews figures from other studies, so the numbers are secondhand and depend on those studies' protocols.
Facts taken from this source
- Zayats and colleagues report that a quality checking pass improved the average transcription error rate by 5 to 20 percent.
“with a quality checking pass improving the average error rate by 5-20%” (Introduction)