How to prepare audio for transcription: practical steps before you record and send
The quality of a transcript starts with the recording. Clear audio means fewer unclear passages, more accurate names and speaker labels, and less time spent checking. At LessRec an editor listens to the recording and checks the full transcript, and good preparation helps the editor get every important word right.
Before you record
- Get the microphone close to the speaker. Distance is the biggest cause of muffled speech. A lapel or desk mic near each person beats a single laptop mic across the room.
- Choose a quiet space. Close windows, turn off fans and air conditioning if possible, and avoid cafés and open-plan offices.
- Record each remote participant locally when your platform allows it. Separate tracks make speaker labels much easier.
- Avoid background music. Music under speech makes words harder to separate, both for people and for software.
- Introduce the speakers at the start. A quick round of names and roles helps the editor label everyone correctly.
- Get consent. Recording laws differ by state; some require every party to agree. Check the rules that apply to you before recording.
During the conversation
- Ask people not to talk over each other on the passages that matter most.
- Spell unusual names and terms aloud the first time they come up.
- Repeat key numbers, dates and figures clearly.
- If someone speaks off-mic, repeat their question into the microphone.
Before you send the file
- Send the original file rather than a re-recorded or heavily compressed copy. Common formats such as MP3, WAV, M4A and MP4 work.
- Write a short brief: who speaks, their roles, the language, and the purpose of the transcript.
- Add a glossary of names, organizations, acronyms and technical terms. A slide deck, agenda or case caption is often enough.
- Mark priority sections with timestamps if only part of a long recording matters.
- Choose the output: clean or full verbatim, timestamps, speaker labels, and file type (DOCX, TXT, SRT).
Which service fits your recording
Different recordings need different conventions. Research interviews often need consistent verbatim rules and anonymization (see research interview transcription). Hearings and legal recordings need careful speaker attribution (see court transcription). Video content may need timed captions (see subtitles and captions). If you already have an automatic draft, it can be checked against the audio (see editorial transcript review).
Fixing a recording you already have
If the recording is already made and the audio is poor, do not try aggressive noise reduction on your own copy first; it can remove parts of speech along with the noise. Send the original and describe the problem (echo, wind, a speaker far from the mic). The editor will mark anything that cannot be heard reliably rather than guess.
Why preparation pays off
A well-prepared file reduces the number of passages marked unclear and the time you spend verifying names and numbers afterwards. It also makes the written quote more precise, because the scope and difficulty of the recording are clear from the start.