More songs available for demo.
Updated Pitch Detection Algorithm.: Uses Machine Learning (Tensorflow) to determine the frequency of small intervals (~10 ms) in the each audio bite corresponding to a word in the transcription. After truncation and converting the frequencies to semitone units (musical notes), the most frequent musical tone is returned as the pitch of that audio bite.
Optimized dependencies for Sagemaker environment.: Reduced the execution time for fulfilling dependencies by migrating to Tensorflow notebook. This will make the demo much quicker. Local environment support in addition to Sagemaker coming soon.