The Faculty BMS (Behavioral, Mananagement- Social-Sciences) from the University of Twente uses AmberScript’s Automatic Speech Recognition software to give researchers and students of the faculty the opportunity to automatically transform audio recordings into text. By using automatic speech recognition (ASR), researchers lose less time on creating transcripts manually. The time (and budget) saved can be reinvested into what really counts: Innovative research.

About the BMS Lab

The BMS Lab is the innovation booster for the Faculty of Behavioral Sciences, Management Sciences and Social Sciences at the University of Twente. The aim of the BMS Lab is to stimulate innovation within the research community by inspiring researchers to use new technologies and methodologies. By doing so the BMS Lab stimulates researchers to seek new insights and strive for excellent research results.

What is the goal of implementing AmberScript?

Transcription (converting audio into text) is a crucial but time-consuming activity in the process of qualitative research. Researchers need an accurate transcript in order to be able to carry out analyzes, but transcribing manually is a tedious process. Jan Willem van 't Klooster, Managing Director of the BMS Lab: "For some time now there were questions to the BMS Lab about whether technology could accelerate this process. So far we didn’t find one, but with AmberScript we were convinced of the quality and user-friendliness."

AmberScript uses state-of-the-art speech recognition technology to speed up the transcription process. With the help of artificial intelligence, audio recordings are transformed into a transcript within a few minutes. Subsequently, researchers can make adjustments where necessary. Eventually the transcript is further processed in analysis programs such as SPSS, MAXQDA or Atlas.

Until now, researchers at the University often hired students who had to manually type out interviews and group discussions. It turned out that transcribing and editing an interview of 1 hour could easily take up to 6 hours of manual work.

By automating this process, a considerable amount of time can be saved and more transcripts can be generated in less time. One student states: "The automatic transcripts from AmberScript need some editing, but it saves a huge amount of time."

How does this work in practice?

Students and researchers can request 'transcription hours' centrally in the BMS Lab. An admin user transfers transcription hours to users and in the admin dashboard the BMS Lab can easily identify which projects consume how many hours. 

Next to that, the BMS Lab supports users with lending out the right recording equipment and with advice on how to record high quality audio. The BMS Lab also advises on the right way to deal properly with confidential information from respondents. 

What are the possibilities of speech-to-text in the future?

At the moment researchers and students work with AmberScript's user-friendly editor.

In addition, the same technology can also be used for 'more challenging' purposes. The BMS Lab is now developing a research application that directly uses AmberScript's 'speech-to-text' engine via an API.

With this integration, researchers can immediately obtain results about the language used in experiments. The combination of several input channels in this app opens the doors to developing a more complete picture of respondents inside and outside of the lab environment.

Conclusion:

According to Jan Willem van 't Klooster, the project 'AmberScript’ is already a success. In recent months, the demand for automatic transcription has increased significantly. Researchers regularly request higher volumes of transcription hours and students and actively use AmberScript.

The reason for this: The old-fashioned way of manual transcription is time-consuming and can be frustrating to a lot of people. AmberScript helps researchers keep the focus on what really adds value: High-quality Research.