Who is this for?

  • Researchers seeking to protect voiceprints of their cohort to facilitate sharing of the files with collaborators
  • No coding experience will be required to run the modules

How do you parallelize?

  • Run the module on several digital voice audio files at once.

Tool Category: De-identification

Function: Voiceprint De-identification

Tool type: Digital Voice Tool

Key Capabilities:

  • Maximizes the protection of a speaker’s identity (voiceprint) via voice alteration, while simultaneously maximizing the preservation of cognitive signal
  • Customizable parameters such as pitch, time scale, timbre, f0, and noise levels are utilized to alter the voice

Based on:

  • Open-source software packages such as librosa, soundfile, and pyworld

Works With:

  • Digital voice audio files

Potential use cases:

  • Protection against identification of a speaker via their unique voiceprint