How to Transcribe an Interview
Interview transcription is not only about producing words. A reliable workflow preserves context, makes quotes easy to verify and reduces time spent searching the recording.
Record each voice clearly
Place microphones close enough to capture each speaker without clipping, and choose a quiet space with limited echo. For remote interviews, keep the best original recording available. A short sound check before the interview can reveal quiet speakers, intrusive background noise or a missing audio channel.
Prepare the file and language
Upload a complete supported recording no larger than 100 MB. Select the spoken language if it is offered, or choose automatic detection. Keep a separate list of names, organizations and specialist terms to consult while reviewing; AudText does not provide a custom vocabulary field in this upload form.
Verify quotations with word playback
Click a timestamped word to replay its part of the interview. Check the surrounding context as well as the sentence itself, especially for names, dates, figures and quotations you plan to publish. Overlapping speech can cause missing or inaccurate words; do not infer a statement from automated text alone.
Edit a readable transcript
Use the plain text view to correct wording and adjust paragraphs, or Find and Replace for a consistently misheard name. Copy the reviewed text or download a document. Speaker information may be returned in structured data, but confirm attribution by listening rather than assuming an automatically detected speaker identifier is a person’s name.
Export for the next step
TXT gives plain text, DOCX supports further editorial work and PDF gives a fixed document. JSON includes edited plain text alongside the original returned word timing and available speaker data. If you need interview captions, download SRT or VTT and correct them in a subtitle editor; plain transcript edits do not update their wording or times.