We use cookies to improve your experience on our website. By clicking OK or by continuing to browse the website, you consent to their use.
Click here to review our Cookie Use Policy.

OK

by QSR

Supports robust qualitative and mixed methods research for virtually any data source and any research method.

Learn more

by QSR

Intuitive data analysis software designed for public policy experts analyzing surveys.

Learn more

Creating software to help you discover the rich insights from humanised data.

Learn more

How to get the best quality audio for transcription

25 October 2018 - BY Dan Watts

NVivo Transcription has arrived to help solve the problem faced by many researchers – hours of manual transcription of data sources like interviews, and focus groups, which take precious time and resources away from analysis of the data. NVivo Transcription is an automated transcription assistant, which utilizes natural language processing technology to listen to the audio uploaded, to produce an accurate transcript, ready to be coded and analyzed.

NVivo Transcription produces the best and most accurate results when the audio uploaded is of good quality. In this blog, I’ll explore what we really mean by that and give you some helpful tips for recording your audio, so when it comes time to transcribe, NVivo Transcription will deliver a great result for you.


What is good quality audio?

We define good quality audio for NVivo Transcription as:
  • having minimal background noise
  • clear speakers
  • close microphone placement and limited crosstalk
Achieving all of these elements can be a challenge depending on the environment that you’re recording in, the equipment you have to work with, and particularly when it comes to cross-talk, larger focus group situations can be trickier.

These are my top tips to get the best recording, the first time.


Tip One: Take time to set up and test

It pays to know what you’re working with. Whether it’s just your smartphone, or a more sophisticated set-up with microphones and headsets, taking the time to test your equipment, preferably in the environment you intend to record in will let you know ahead of time how you can expect it to perform. Listen out for things like ‘fuzzy’ audio, feedback, and echoes.


Tip Two: Minimize the background noise

Quick snippets of audio on the street are great for news grabs – but less so for natural language processing to be able to accurately understand what’s being said by the speaker. So, when selecting a location to record an interview or focus group, try to find a space which is away from particularly loud and busy areas, such as a carparks or lunchrooms. And, while they’re convenient meeting places, try to avoid locations like cafes and bars which typically have a lot of busy chatter or music.

Touching back on the previous point, it pays to record a few moments of ‘dead air’ in the space when you’re testing your equipment, as this will give you a good indication of the background noise your equipment is likely to pick up. It’s quite amazing what you’ll hear when you focus on it – in our busy lives we’re very good at filtering out what doesn’t matter to us, so the reversing of a truck and it’s warning signal might not usually register with you when going about your usual business, but it will certainly stand out as background noise on an audio recording.


Tip three: Ask your subjects to speak clearly and avoid colloquialisms

Let your subjects know ahead of time that you are recording the interview for the purpose of transcription, and that it would be greatly helpful to you if they are mindful of the pace of their speech, particularly if you are recording people with strong accents as they can be difficult to transcribe.

If possible, avoiding the use of slang and dialect specific words can also be helpful.

Another important tip to keep in mind when you’re uploading your file/s for transcription, is that if your audio file contains two or more accents, in order to achieve the best results you should select the language model of the weaker speaker.


Tip four: Strategically place your microphones

To get the most accurate capture of what people speaking are saying, as well as minimizing background noise, place the microphone close to them. If you’re recording an interview with multiple people, or a focus group, it’s worth considering having a microphone (or two!) that they can pass around.


Tip five: Do you best to avoid crosstalk

Crosstalk happens when two or more people speak over one another, and it’s difficult to transcribe. It is most likely to occur if a speaker is interrupted, so as an interviewer, do your best to ensure that your subject has finished speaking before moving on to the next point.

In a larger group interview or focus group, crosstalk may occur when a great discussion gets going, however, these important points and details may be missed if what is being said is not captured well. In this case, some ‘moderation’ and of your focus group comes into play, by asking individuals to repeat their point if they were not heard, and politely reminding others to please raise their hand to ensure you can hear from everyone.


You can achieve up to 90% accuracy with NVivo Transcription

By following these tips, you’ll be able to capture audio with minimal background noise and clear speakers, which is exactly what you need when it comes time to transcribe. With good quality audio, you can achieve up to 90% accuracy with NVivo Transcription.

NVivo Transcription is available now, ready and waiting to be your automated transcription assistant. You can try NVivo Transcription for yourself, when you sign up for a free trial, with 15 minutes of time included.