Why Use Human Transcription

There have been many advances in speech recognition technology, and machine generated transcripts are certainly a quick and cost-effective way of getting a rough record of a conversation. But we’ve all seen how inaccurate those Instagram captions can be, and that’s often when just one person is speaking!

Why human transcription is the best choice:

• Unlike automated transcription software, human transcribers can ignore background noise (within reason!) and still produce an accurate transcript. (Though a good quality recording will still always be preferred by a human.)

• It is virtually impossible for automated software to recognise and identify multiple speakers in a group of more than three people.

• The most obvious way to spot a machine generated transcript is mistakes with homophones, such as sea/see or weather/whether.

• Automated software will always struggle when speakers have strong accents or difficult dialects.

• Sometimes a bit of Googling is required. A human transcriber can use the internet to research unusual technical terms, names and places.

• People do not speak in perfect grammatically correct sentences. But human transcribers can tidy up as they type, so that you end up with a readable yet accurate record of what has been said, without wading through all the ums, errs and false starts. We call this intelligent verbatim.

• In transcription, common sense often needs to be applied, and our transcribers know how to make judgements based on the context of the conversation.