A quality‑control process that compares a produced transcript of spoken material (interviews, speeches, broadcast audio/video) against the original audio or video source to correct factual errors, speaker attribution, timing, punctuation that affects meaning, and to note nonverbal cues where relevant for accurate reporting.

Principle

Automated speech recognition and human shorthand produce errors of omission, substitution and misattribution; explicit verification against source media is required to ensure verbatim accuracy or to document departures when a nonverbatim summary is used.

Demonstration

Illustrative scenario → A reporter uses automatic speech recognition (ASR) to generate a draft transcript of an interview. A verification pass listens to the audio, corrects homophones and proper names, fixes speaker labels where overlapping talk occurred, and annotates pauses and emphasis for editorial use. Result → Quotations used in the story match the recorded source and speaker identity is defensible.

Misapplication

Assuming an ASR or unreviewed human transcript is verbatim and publishing quotes without checking the audio; the error is treating machine output as authoritative rather than as a draft requiring confirmation.

Consequence

Proper verification reduces misquotations, clarifies attribution disputes, and strengthens legal defensibility and editorial credibility; it requires additional time and resources and can delay publication when tight deadlines exist.

Reversal

In live captioning or immediate breaking coverage, full verification is often infeasible and a trade‑off accepts lower verbatim fidelity for speed; for indexing or thematic summaries, nonverbatim transcripts may be sufficient if documented as such.

Boundary

Clearly within: comparing and correcting a verbatim transcript of a recorded interview before publication. Boundary case: spot‑checking selected quoted passages rather than full verification. Clearly outside: informal reporter notes or summaries not intended to serve as verbatim records.

Semantic Tension

Timeliness ↔ Verbatim Accuracy — the level of verification appropriate depends on the editorial need for speed versus the legal and factual sensitivity of exact wording.

Synthesis

Transcription verification converts provisional text into defensible evidence by reconciling machine or note outputs with source media; choosing the appropriate verification depth is an editorial judgment tied to legal risk and journalistic standards.