Monthly Archives: December 2015

Although transcribing interviews by journalists or students that only involve two voices is the most common work I do, transcribers often have to work with tapes with more than two voices on them. How do you tell the voices apart so that you can differentiate them on your transcript? This article shares some tips I use to tell different voices on a tape apart.

What’s the problem with transcribing multiple voices?

I was transcribing an interview with two fashion designers today; my client had said it was OK not to differentiate them and the women themselves said that their voices were often confused. How did I tell the voices apart so I could produce a transcript that had the correct words attributed to the correct people?

Although it’s sometimes very easy to tell the people who are talking on a tape apart, for example if they’re a man and a woman, or one has a very strong accent, sometimes it can be difficult. Because it will help my client to know who said what, it’s important for me to try as hard as I can to differentiate the voices and make the transcript as useful as possible.

Before the interview: who are the interviewees?

If you know it’s going to be an interview with more than one participant, you can ask your client to help you from the very beginning.

Either they could ask their interviewees to introduce themselves by name at the beginning of the tape. Even if they are all, for example, young male voices, you can pick up a lot of information from this that will allow you to differentiate between them …

Or they could ask people to introduce themselves every time they make a point (this works in a more slow round table discussion at a conference, for example)

Taking the first option once led to a very sweet tape where the musicians in a band introduced themselves by name to me, mentioning my name, at the start: “Hello Liz, my name’s … and I hope you can understand me”. Aww!

After the interview but before you start typing: checking who is who on the tape

If you didn’t get the option to ask your client to get the interviewees to introduce themselves, it is OK to ask them who is who – for example, who speaks first, who has a voice that is distinctive in some way. They might also mention that, for example, the lead singer talks most and the person who only talks about one track is the drummer.

If you’re working on a discussion at a conference, you might be able to get some information from the conference website. For example, there might be a video up already that time stamps each person’s speech with a note of their name. Play the video, check the speech against your tape, and there you go.

When you’re transcribing: how do you differentiate between the different voices?

If you have no clues about who is who or who says what, there are still ways in which you can differentiate between voices on a tape. It can take time to get used to doing this, but it is useful.

1. Check the video. This one sounds obvious, but if you have a video to transcribe, do look at it carefully. There may well be captions stating who is speaking, at least for the first time, and you can recognise who is who by their appearance. If there’s the option of a video for a conference or marketing meeting / focus group, do take it, even if it takes longer to download.
2. Check where people are in space. In the tape I’ve been working on most recently, the speakers were sitting either side of the tape recorder. So, even though their voices were similar, one came from the left and one from the right. Result!
3. Check the sound level/volume. If one person is sitting further away from the recorder, they will sound fainter.
4. Check for even slight accents. There may be a non-native-speaker or regional accent on the tape: listen for different vowel sounds or intonation.
5. Check the ums, ers and filler noises. These really vary across speakers and can make a difference. Person 1 might say “like” constantly, while Person 2 “ums” and “ers”.
6. Check for clues in the environmental context. Does Jane order food but Sally just have a coffee? The one talking through her dinner is likely to be Jane.
7. Check for clues in what they say. I often switch off from the content when I’m transcribing, just letting the words come into my ear and out of my fingers. But people will refer to each other by name, and this gives you a good clue to who is who, or reinforces your first thoughts (If the person you think is Pete refers to “Pete”, unless you have several interviewees with the same name, he’s unlikely to actually be Pete!).

I have two other handy hints to add, which I use all the time …

• Draw a plan or write notes! When I work out who is who, I will write a little diagram out or make notes – “Bella … Jean” for the left/right ones, “Jim: high-pitched. Bob: rumbly and quieter” etc.
• If you can’t put a name by each participant, at least try to break the text up into paragraphs spoken by different people. You might be able to go back and add the names if Julie says, “As Veronica said earlier, it’s difficult opening a tin of Spam”, for example.

It can be challenging when you find you need to transcribe a tape with more than one or two voices on it. As you have seen, there are things you can do to make this easier before the interview even starts, once you receive it and during the transcription process.

If you’ve found this article useful, please click to share! If you are a transcriber and have any tips to share on this topic, please do comment below!

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.

Related posts in the series:

Why do transcribers charge by the audio minute and not per word?

How do you start a career in transcription?

Why you need a human to do your transcription

Being a professional transcriber – software to use to help

Ten top tips for transcribers