The science behind Voice Äpp

Voice Äpp is a science communication project (SNF Agora). The development of the app took two years. It took this time because we developed the following components from scratch:

  • Automatic speech recognition for Swiss German, based on audio collected with Dialäkt Äpp
  • Creation of a corpus of Swiss German dialect speech
  • Development of pitch analysis algorithms for mobile devices
  • Speaking rate analysis for mobile devices
  • Hearing loss simulation on mobile devices
  • Tinnitus simulation on mobile devices 

Potential future analyses with Voice Äpp data

Once we receive user data, Voice Äpp will allow for phonetic analyses. Results will provide valuable insights for dialectology and for population statistics of numerous speech parameters:

  • Evaluation of automatic speech recognition for Swiss German
  • Population statistics of pitch
  • Population statistics of speaking rate
  • Between-dialect, -age group, -gender and -speaker analyses of speech