Jobs, Lifestyle, Student Life November 07, 2019
Automatic Transcription: How to save time while writing your thesis

Automatic Transcription: How to save time while writing your thesis

There are a lot of things that make you waste time in your masters (or bachelors) programme: the way too extensive essay you don’t even get a grade for, or your weaker self that makes you watch fail compilations on YouTube over and over again instead of starting with your paper. With that in mind it’s nice to have at least some tools that help you with other time-consuming work at university.

Transcribing, meaning typing out recorded audio, is such an annoying and time-consuming task. However, since it’s an important part of qualitative research, almost every student has to deal with it at some point. Especially when doing your masters it will probably become an issue, as research there is much deeper and more detailed. In most cases, transcription is necessary for recorded interviews, because as basis for an analysis you also need those interviews in text form.

Tools for Transcription

Although transcribing can be very exhausting, it is at the same time a great introduction to your own research. It’s a helpful step to become familiar with the content of what has been said. However, typing out recordings also means staring at your monitor for hours and skipping through your audio file hundreds of times.

It can take up to six hours to type out one hour of audio, depending on your transcription skills and the quality of the audio. This transcript of one hour of audio will take up the space of 20 to 35 DIN A4 pages. If you interviewed ten people for one hour each, it will take up to 60 hours to type out everything. In comparison to that: In 60 hours you could watch all Marvel movies and would still have time for the extended version of Lord of the Rings!
Luckily, there are some tools that can help you!

Software without automatic speech recognition

Software without automatic speech recognition offers convenient editors where you can upload your audio file. Then you can use shortcuts to play the audio faster or to repeat the last few seconds. It’s also possible to insert time codes or speaker names by using shortcuts. Don’t underestimate how many times a recording has to be played and paused. Shortcuts can save a lot of time with this. Still, you will not be spared typing out everything by hand.

A significant upgrade to that: Software with automatic speech recognition

Technology has come a long way: there is also transcription software that uses automatic speech recognition to automatically create transcripts. This software also automatically differentiates between different speakers and inserts time stamps. So you don’t have to type out everything anymore, you just have to correct the text that the software creates within minutes. Here, too, a editor offers speed regulation and a rewind function through shortcuts.

The automatically created transcript is linked to the audio which makes it possible to find any word or piece of audio directly. All of this reduces the working time to 1-2 hours per hour of audio. This way, in 60 hours you could manage to transcribe ten hours of audio AND watch all the Marvel movies! That’s a good deal, right?

General Tips for Transcription

1. Audio Quality

If possible, use a good recording device during your interviews. It doesn’t matter if you use transcription tools or not: good audio quality is always an advantage and can save you a lot of frustration afterwards. With good audio quality you often even understand incomprehensible words and the error rate of automatic transcription software is much lower. In many cases you can rent good recording devices at your university. For most recording devices you get the best result if they are not laid down on a table (especially for smartphones!) and at a distance of about 10 centimetres to the speaker. Avoid overcrowded places like the cafeteria or bars.

2. Avoid Interruptions

It may sound obvious but interrupting each other in an interview situation is more difficult to recognize and avoid than you think, as it’s such a natural part of conversations. Specifically pay attention to this issue and you‘ll thank yourself afterwards.

3. Distinct File Names

This sounds even more obvious than the previous advice but software and recording devices often assign quite ambiguous file names like „track1“. To differentiate between your files more easily, assign unique names to your files. It’s particularly annoying if you accidentally upload it to a transcription software. This happens more often than you think!

4. Scheduling

Start with your transcription as soon as possible after the interview and, of course, long enough before the deadline. This way you still have the content in mind and words and contexts are easier to understand. Transcribing is unfortunately a step that can’t be done quickly at the last minute.

5. Breaks

Never transcribe for too long at a time. After two to three hours at the latest, most people start making more and more mistakes, as it’s difficult to stay focused on such a monotonous task for a long time.
Take enough coffee/tea/snacks breaks!

6. Emergency Solution: Outsourcing

In case you underestimated the effort and time you have to put into transcribing your audio files – even if it’s just about the very last step – and the deadline is coming closer and closer and you become more and more desperate and nervous minute by minute – don’t worry – we all have been there! Luckily, with most automatic transcription provider, as for example Amberscript, there is the option of outsourcing the editing to other human transcribers after having your files automatically transcribed, so that in the end you receive a „perfect“ text file. Now all you have to do is focusing on the other things that are still on your list and that need more attention – or just relax!

At https://www.amberscript.com/de/blog you can learn more about transcribing and automatic speech recognition in AmberScript. You can even try 30 minutes of transcription for free!

Fun Fact: did you know that the idea for Amberscript was born due to this exact issue? Founder Peter-Paul de Leeuw wanted to save time transcribing interviews for his thesis but couldn’t find a satisfactory solution – so he founded Amberscript.

Related posts
Let's face it, appropriate clothing for a job interview can be a very broad term these days. Of course, fashion
Music has become an integral part of everyone’s lives. According to Forbes, individuals today spend about an average of 32
Are you unsure about your future career path? We’ve all experienced that part of life where we’re unsure about our

Leave a Reply

Your email address will not be published.