Useful Tips When Transcribing Audio

Page content

An audio transcription is a written copy of pre-recorded audio, generally used for legal, business or medical purposes. By transcribing audio the client is able to keep a written record of the conversation which occurred. Audio transcription work differs from stenography or audio typing, where the transcription occurs in real-time or as it is taking place. This type of work is much more difficult and requires further training.

However, if you wish to learn of some tips when transcribing audio, this article will provide some basic pointers to get you started. This type of work is not hugely difficult, although it requires dedication and patience, and with the right guidance you should be off transcribing your first hour of audio-stream in no time.

Audio Quality Tips

These tips are not only useful if you are organizing the transcript-session yourself, but are also applicable generally even if you are receiving material recorded by someone else. If this is the case let the client be aware of these points.

  • It may seem obvious, but the first thing which you need to ensure is that the quality of the recorded audio is clean, the volume level of speakers is high enough and the background noise is minimal.
  • In the case of multiple speakers, (i.e. at a conference), ensure that the different speakers are easily recognizable and there are no overlaps or interruptions when speaking occurs.
  • Never sacrifice the quality of your audio, even if it means larger disk-space or more tape used.
  • Avoid using a poor quality dictaphone for recording audio or using the automatic features in a recorder which could cut-off sentences (i.e. Auto-Vox).
  • Ask for the speaker to repeat content if you are unsure, as it will save time when you are transcribing and will avoid you the pain of editing the transcription or simply guessing.
  • Finally save the recording in a high resolution format. WAV audio is preferred because although it is compressed, it is easy to work with and is of decent quality. For MP3s or WMA files, a bit rate of 128kbps or above is acceptable for speech.

Verbatim or Edited

Depending on the needs and flexibility of your client, you may be required to transcribe ‘verbatim’, which means a literal transcription of everything which has been spoken. You will not include ‘uhms’, clicks or phonetic sounds, but stutters and the repetition of words are usually included in the transcript. If there is something which isn’t clear, try not to guess but instead deduce logically from the sequence of the conversation and topic. As a last resort, including ‘inaudible’ in brackets should be considered if it has the potential to alter the meaning of the transcript or render it incomprehensible (i.e. a key word or phrase). For verbatim transcriptions it is important to have a good ear, attention to detail and accuracy since you will be putting-down pretty much everything.

Edited work requires you to have a good command of written English, (as opposed to spoken), grammar and sentence structure. Remember you are editing the work so that it is easily readable and flows, but the meaning of the transcript should not be altered. In edited work you will not include stuttering or similar.


As a last point, if you are working on a job involving technical words, make sure you are familiar with the subject being spoken or the field generally. This is particularly true of medical transcriptions where you would at least research the topic briefly so that the terminology is transcribed accurately.