Peer-reviewed research
Evaluating Workflows for Creating Orthographic Transcripts for Oral Corpora by Transcribing from Scratch or Correcting ASR Output
Jan Gorisch and Thomas Schmidt, Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), 2024
Read the originalRead on October 17, 2026
- What it is
- A 2024 comparison of two ways to make orthographic transcripts of recorded speech for research: typing from scratch, or correcting automatic speech recognition output. It used recordings of interviews, colloquial talk, dialect and historic material.
- How much weight it carries
- Peer-reviewed conference paper. The transcription conventions were strict research conventions, and the corpora were research recordings rather than general business audio, so the timing may not match a routine transcription job. The abstract is the part we read.
Facts taken from this source
- The 2024 study's authors found that the more complex the recordings, the more time corrections take.
“The more complex the primary data, the more time has to be spent on corrections” (Abstract)
- In a 2024 study of research transcripts, transcribing from scratch and correcting automatic output took equally long, about 7 times real time on average, for the data and conventions tested.
“transcribing and correcting still take equally long with 7 times real-time on average” (Abstract)