Training pipeline for an osu!mania 7k next-event model, designed as the upstream predictor for audio-driven 7k map generation systems.

Cross platform audio feature extraction and sound classification tool

Java Implementation of the Sonopy Audio Feature Extraction Library by MycroftAI

Speaker recognition using Mel Frequency Cepstral Coefficients (MFCC) and Linde-Buzo-Gray (LBG) clustering algorithm

Urban Sound Annotation and Classification

Audio input -> real-time analysis -> OSC output. Takes in real-time audio, does feature extraction using smart algorithms then sends out OSC to be used in other programs.

Scratch for experimenting with audio feature extraction.

TuneSpy is a Python application that allows users to load audio files, generate spectrograms, extract MFCC features, and compare the loaded audio with a preprocessed database of songs to find the most similar match.

Convolutional-based supervised regression task for extracting high level timbral features from drums sound files, useful to condition a real time Neural Sound Synthesiser on continuous intuitive controls.

Drum Samples Clustering, Audio feature extraction and clustering audio files using data visualization and dimensionality reduction (PCA).

A CNN model for classifying whale calls

Built a Speech Emotion Recognition system that predicts human emotions from speech signals using Python and machine learning

Tooling and datasets for neural-network powered audio feature based synthesis

Various Neural Network Architectures for Supervised Tonic classification using the mridangam_stroke dataset, and supervised instrument classification on the TinySOL dataset.

A simple music feature extractor for Deep Learning models

Text-independent speaker identification system based on GMM

Created as part of Audio and Music processing lab assignment. Extracts and analyses features from an audio collection, and creates playlists based on various descriptors. Can create playlists based on music similarity too.

AudioInspect is an app that extracts audio features from uploaded audio files or audio files in a specified folder, providing insights into the characteristics of the audio.

Generation of music playlists based on audio features analysis using Essentia and the MusAV dataset

Developed a deep learning model using Multi-Layer Perceptron to recognize and classify speech signals into 6 distinct emotions. Extracted 160 audio features, enabling the model to detect emotions with around 75% accuracy on the training set. Implemented the model on a Streamlit dashboard.