search

fabiocat93 / fab

voice analysis
MIT License
0 stars 0 forks source link

Maintaining Desired Audio Features Post-Anonymization #1

Open fabiocat93 opened 1 year ago

fabiocat93 commented 1 year ago

Issue Description: As part of our ongoing efforts to ensure user privacy and data protection, we are in the process of implementing and comparing state-of-the-art audio anonymization techniques. While doing so, we want to make sure that certain key audio features are preserved even after anonymization. This will help us maintain the functionality of our audio data while still upholding privacy standards.

This issue is intended to track the discussion of maintaining some desired audio features post-anonymization. Please use this issue to list your desiderata and why.

EnriqueDiaz2023 commented 1 year ago

First of all, congratulations on the excellent work you are doing on the topic of voice anonymization. My first contribution is the following, if you consider the scenario where the speakers present dysphonia, then their anonymized voice would need to preserve this characteristic. One way to measure this preservation would be with the study of statistics derived from Fundamental Frequency, Jitter, and Shimmer (to mention a few) before and after anonymization.