FURI | Summer 2020

Violin Vibrato Video Imaging: A Novel Method to Generate Frequency Iterations Using Convolutional Neural Networks

Data icon, disabled. Four grey bars arranged like a vertical bar chart.

The violin vibrato is an artistic modulation of a violin’s pitch that gives the tone of the instrument a desired aesthetic effect. Using computer vision techniques, the purpose of this research is to analyze video and audio of a violin’s hands. The audiovisual data will be fused together to generate a more complete waveform to better understand how hand motion contributes to creating vibrato.  Currently, efforts in the research project are focused on refining a convolutional neural network to track finger joint locations on a large corpus of videos. This is the first step to an automated pipeline for determining vibrato with machine learning techniques.

Student researcher

Ryan Widjaja

Electrical engineering

Hometown: Chandler, Arizona, United States

Graduation date: Spring 2023