Skip to content

How well does speech recognition work? – Part 1

This learning scenario is about using different speech recognition software and identifying the weak points. Are there voices that are recognised better and voices that are recognised worse?

In the preparatory phase, students prepare questionnaires to collect specific demographic characteristics of the participants. In order to better understand how speech recognition software works, a brief basic introduction to AI tools is given.

The next step is to pick out different tools that are of interest. Then, speech recognition tests are carried out and the quality is evaluated.

CreatorAnnika Nowak
SubjectComputer Science & Physics
Length60 minutes
Pedagogical ApproachExplorative learning
Competencies from the framework

Additional competencies
Scientific inquiry cycle
Scientific thinking
Scientific investigation cycle
Formulate research questions
Link to the worksheet

Learning Activities

LA1: OrientationFirst, students learn about different speech recognition tools
LA2: ResearchIn the next step, students research different speech recognition tools and decide on one.
LA3: Plan In order to test the speech recognition software comparably, the text and the different characteristics of the target group (such as age, gender, dialect) are defined. In addition, the students define a suitable research question.
LA4: PerformStudents conduct the planned experiment, collecting data as they go.
LA5: Storage and backupThe data is then stored and backed up so that it can be retrieved.


Hint: You can create a copy of the worksheet and save it in your google drive to modify and change the worksheet. Please use the Creative Commons License and create a reference to the FabCitizen project.

Leave a Reply

Your email address will not be published. Required fields are marked *