1) create a function which calculates the chroma (12-dimensional) of an audio file
- your output should have a chroma vector for each frame given hop and window sizes
2) Use your DTW function to find the optimal path between 2 audio files based on chroma similarity
3)Create distance and path plots for the two files below
- britneyIntroVox.wav, stonesIntroVox.wav,
4) Also compute self-similarity matrices and cross-similarity for the songs in their entirety
- britneySatisfaction.wav, stonesSatisfaction.wav,
5) play around with the window/hop sizes and the distance the cutoff frequencies are located from the mid frequencies of each octave in your chroma computation
if you don't finish your chroma extraction function in class finish it before next Wednesday b/c we will use it for ML and pattern classification |