Transcription of audio data

Are you recording spoken language? And would you like to avoid transcribing the audio manually? Then an automatic audio transcription tool may be useful for you.  

Why would I transcribe my audio data? 

  • Enable searching and quickly identifying subsections within the audio data. This also allows for running textual analyses (e.g., topic discovery, semantic modelling). 
  • Create captions for the hearing-impaired. 
  • Lower the privacy-sensitivity of the audio data by taking out the voice and enabling the removal of other unnecessary personal elements. 

Which tool to choose? 

There are many tools available to automatically transcribe audio data. In this Zenodo record, you can find an overview of several common tools, including a test audio file (in Dutch) and raw transcripts from these tools, enabling you to check their quality.  

If you work at Utrecht University, Amberscript and Microsoft Word 365 are recommended tools to use for audio transcription.

For more information on how to access and use the various tools, visit the intranet page about software or the tooladvisor. If you wish to use other transcription tools, first contact us or your faculty data manager to make sure the appropriate security measures are in place. 

Please note that not only the tool, but also the quality of the audio recording will determine the quality of the transcription. For example, you can increase transcription quality by reducing background noises (e.g., when outside, use wind caps on the microphone), preventing speakers from talking simultaneously, and avoiding jargon or foreign language.