Introduction This project done for my CS221 class aims to classify classical music by musical era (Baroque, Classical, Romantic, Modern) with composers as a proxy. Using audio processing techniques, such as Short-time Fourier Transform, we extracted features such as the spectrogram and chromagram of the audio data from two datasets, Free Music Archive and MAESTRO. We used two ensemble classifiers, AdaBoost and Random Forest, and found that although Adaboost performed marginally better than Random Forest, the latter made more generalizable predictions....

Introduction I recently became interested in neural networks and sought interesting applications of what they could do. Neural networks are based on how the human brain works; the networks are made up of many nodes that take inputs and predict outputs. Over time, with more data, the networks improve by correcting for error, known as backpropagation. Neural networks become especially useful when there are many “layers” between the input and output that allow for more fine-tuned fitting of data....

This project done for my CS106X class mixes two given mp3 files together and has a visualization to go with it. Introduction I have always been awestruck by DJing and have had little idea of how it is done. I set out to create a python script to mix songs to understand the process a bit better. I first found that DJs use many effects in their mixes. These effects include changing tempo of songs, changing the pitch, as well as adding noises from other tracks....