There were two algorithms investigated by this project. The first
used a tokenization routine to express the time variance
of the dominant frequency of a given signal as a token string that could then be
searched. The second used a variety of neural networks
for pattern recognition of the audio recordings.