shazamalam
TEAM ZERO's killer code for finding copyrighted audio.
it takes two directories, a directory of original audio files, and a directory of audio files that may have used pieces from files in the first directory. The output is a list of potential matches, and the start time of those matches in the audio files that use the original audio.
For a summary of our strategy, check out notebooks/fingerprinting.ipynb
We do a nasty spectral analysis of the audio files, creating sick audio fingerprints that consist of rad pairs of peak frequencies. Check it out.
Our Awesome Team Members:
- Dan Calacci (calacci.d@husky.neu.edu)
- Lucas Haber (haber.l@husky.neu.edu)
- Eric Peterson (peterson.er@husky.neu.edu)
- Talia Swartz (swartz.t@husky.neu.edu)