Which Professor Are You?
Built by Nathan Shelly and Sasha Weiss, for EECS 352 at Northwestern University, WQ17.
Voice Matching
Which Professor Are You investigates voice-similarity matching. Specifically, we investigated various audio features that could be used to match a speaker to others they "sound the most like". As an application of this project, our project records a speaker, and presents them with the Northwestern EECS professor their voice most closely resembles.
Our implementation relies on MFCCs, which it uses as features for training and testing using a GMM classifier. You can read our report here.
Dependencies
Most dependencies are reflected in requirements.txt. To install using Pip, run pip install -r requirements.txt.
This project uses Python 2.7.
Essentia
This project also depends on essentia, which is not reflected in requirements.txt.
Essentia and Virtualenv
To run this project in a virtual environment, it is necessary to build/install essentia to your machine, and copy the package into your <venv>/lib/python2.7/site-packages/ directory. See this link for help.