I have recently learnt that many transcribers run their audio files through voice recognition software and if it is of poor quality audio, some actually listen to the file and read it aloud or repeat it into speech recognition software. There are so many dangers in doing this which can cause your business embarrassment, cost your business money or clients, and damage your reputation.
Machine transcription cannot interpret context and meaning, place accurate punctuation marks or identify multiple speakers accurately. Machine transcription cannot tell the difference between ‘complement’ or ‘compliment’, ‘a lot’ or ‘allot’, ‘miner’ or ‘minor’ or even ‘minus’, ‘aloud’ or ‘allowed’, ‘aide’ or ‘aid’, and ‘offices’ or ‘officers’, just to name a few. It can incorrectly interpret ‘radio’ as ‘rodeo’, may hyphenate or join two words to make one when not required, type ’16’ instead of ’60’. These errors can totally change the meaning or the context or the sentence. How do I know this? Experience. From years of transcribing, I have learnt so much more about the English language, and I am still learning. Using artificial intelligence to produce a transcript does not help the typist to learn and then become a professional, as with other professions.
I had the pleasure (not) of editing a transcript which had been submitted to a client. It took me longer to edit the document than it would have to type it in the first place. It was a fantastic learning as now I know what is involved. I had to edit it twice as during the first edit, I was correcting some serious errors. While focusing on those serious errors, I found it was very easy to miss those little things which would’ve been noticeable had there not been so many issues.
Automation makes our lives so much easier in so many ways and there are many things we are happy to be automated, as long as we receive the quality desired. Do you receive a better quality service when there is another person answering your questions while on chat, or if a chatbot is trying to decipher your questions? How many times have you had to hang up the phone and start again because an automated service tried to solve your problem instead of a human? So, when deciding who to use for your quality transcript, ask whether or not your transcript is going to be hand-typed or if a machine is going to handle this for you.
