When I receive a recording for a new transcription client and tell them I’ve loaded it into my transcription management software, I’m sometimes asked if that does the transcribing for me. Um, no.

While there are of course software packages out there that are very good at working with a single voice dictating, even those can sometimes struggle. I know this, because I’ve edited work that has been dictated in this way – and it can often be rife with homophones.

Why use a human transcriber?

I’ve been providing transcription services for several years now. While a machine might be suitable for taking down the words of a single, non-accented speaker enunciating clearly into a good quality recording apparatus, with no background noise, no interruptions and no acronyms or jargon, the projects I’ve worked on have included some or all of these features:

• High levels of background noise – interviews in cafes with espresso machines whooshing and spoons clinking in cups
• People talking while they’re eating and eating while they’re talking
• Interruptions from waiters / room service / other members of the band or group
• High levels of tape noise leaving me straining to hear what anyone’s saying
• Multiple speakers including many people with similar voices around a conference table
• Overlapping speech
• Non-native English accents or heavy regional accents
• Very technical content – jargon and acronyms galore
• Creative content – album names, track names, novel titles, band names, author names
• Requests to provide the transcription missing out ums and ers through to smoothing the English to make it read as standard English

As a native English speaker specialising in working with music journalists and non-native speakers of English, I can cope with all of these, with some rewinding and checking. I doubt that the most sophisticated dictation software could do so, as yet. I might be wrong of course (let me know if I am!).

Understanding what’s being said on a transcription

The first issue is actually hearing and understanding what’s being said. I have a good ear and a native English speaker’s ability to predict what will come next in a sentence / how sentence structures work, plus my experience working with speakers of and texts in non-native English allows me to do this for native Arabic, Chinese, Eastern European language etc. speakers. My ear can filter out background noise where sophisticated software can only go so far. And I can hear around the clink of teaspoons or glasses chinking to grasp what’s being said.

Checking the content in a transcription

When one of my journalist clients sends me a tape, I check who the musician / band is and look them up (usually on Wikipedia for the general information, as their own websites are usually a bit harder to plumb for information). When I’m working on an international conference I will seek out or be given a conference schedule, list of attendees, etc. When I’m working with technical content I will look up information on that topic.

All this allows me to produce a transcription which the client will not have to check for themselves, or if they do check it (which I do recommend), there won’t be too much to change. And I won’t be embarrassed by too many mis-hearings. Just try popping a few names of country leaders, bands or albums into a Word document and running a spell checker and imagine what an automated dictation program would do with these terms!

Speech on a tape to words in a document

Very occasionally I’ll be asked to record exactly what the people on the tape say, including ums, ahs, repetitions and pauses. At the moment, I’m transcribing some roleplays for students learning how to operate a telephone helpline. Here it’s important to capture all the nuances of the conversation and I’m splitting the utterances into sections, numbering them, and including all the ums and ahs.

Normally, my clients will require some smoothing out.

• Most of my journalists like to have an indication of when their subject slowed down or had to mull over something and ask me to include notes of those pauses.
• Business people producing podcasts and telecasts often want a fairly accurate transcription, but smoothed out to eliminate ums, ahs, pauses and repetitions, so they have a good product to sell or include in packages for their clients.
• Some international conferences want to avoid embarrassment for their delegates by having their English rewritten as I transcribe to appear as close as possible to native British (or American) English
• I have worked with authors who start off with a tape and want it to turn into something they can publish as a book (this, unlike all of the other options, involves two processes: transcription and then heavy editing and rewriting).

Why should I use a human transcriber and not a software program?

I think I’ve answered that for you now. You should also consider using a human transcriber who’s a native speaker of the language you’re having transcribed: there are websites out there where you can find very cheap transcribers; they are often not going to be native English speakers and while they will get the gist of the tape down, I’d be unsure whether they could give you the service you needed.

I’ve written another article about how to get into transcribing as a job and how to prepare yourself for what is often a fun and rather creative area of work, and one about the technology transcribers use.

In the meantime, if you’re thinking of booking in a transcriber, do have a look at my page about my transcription services, and get in touch if you want to ask about what I can do for you. I hope you’ve enjoyed this insight into a transcriber’s work! Do leave a comment if you have …

If you want to learn more about Transcription as a career, buy my book: A Quick Guide to Transcription as a Career – buy from Amazon UK or visit the book’s web page for worldwide links and news.