Start With a Better Recording
The single biggest factor in transcription accuracy is the audio. Choose the quietest room available, switch off air conditioning if you can, and put the recorder between you and the interviewee rather than in front of yourself. For remote interviews, ask the other person to use headphones to avoid echo, and record a short test before you begin.
AI First, Human Review Second
Modern speech recognition handles most clear interview audio well, which changes where your time goes. Instead of typing every word, let the AI produce a full draft and review only the passages that matter: the quotes you will publish, the figures you will cite and the names you need to spell correctly.
VoiceMaster links every segment to its timestamp, so reviewing a quote means tapping it and listening — no scrubbing through a timeline. Edits are safe, with undo and one-tap restore to the original machine transcript.
Protect Your Sources
Interview recordings can identify sources, contain off-the-record remarks or include personal data covered by research ethics approvals. Uploading them to a cloud transcription service creates another copy outside your control. On-device transcription keeps the recording and transcript on your phone or Mac, which is often the simplest way to satisfy newsroom policies, IRB protocols and GDPR expectations.
Formatting the Final Transcript
Decide early whether you need a verbatim transcript (every “um” and false start, common in research) or a clean read (light edits for readability, common in journalism). Label speakers consistently, keep timestamps for anything you may need to re-check later, and store the transcript alongside the audio.
- Markdown export: readable drafts with speaker names and timestamps.
- Timestamped text: easy to share with editors and fact-checkers.
- JSON export: segments, speakers and bookmarks for analysis software.