RESEARCH
THE HANDSOLO
The HandSolo is a musical interface I created, that allows any surface to become a playable virtual drum kit, using hand-drumming techniques. The system is a fully blown MIDI controller, integrating seamlessly with music software, and allowing real-time rhythm composition and performance for computer music producers. The HandSolo was presented at NIME and was featured in the MIT Computer Music Journal.
VIRTUAL REALITY
Audio spatialization forms part of the sensory bedrock of any immersive experience today. I led the technical design and development of the 3D Audio Engine that powered the Virtual Reality Platform at Verizon XR. My research included problem solving to find techniques to optimize sound spatialization, realism, and aural immersion in augmented and virtual reality.
CHORUS EXTRACTION
In the field of Music Information Retrieval (MIR), I developed a chorus extraction system that uses novel preprocessing and feature extraction techniques to classify and locate chorus sections within songs. The detection works through machine learning by studying the energy contour of songs, and using targeted spectral analysis on the first order differential contour characteristics to model chorus presence and location.
VOICE ACTIVITY DETECTION
As part of my work at the Music Technology Program at NYU, I built a highly optimized and efficient system that utilizes periodicity measures, energy thresholds, and hangover smoothing to perform significantly effective Voice Activity Detection.
SOUND SYNTHESIS
At the Shared Reality Lab in Montreal, in conjunction with CIRMMT, I worked on interactive realtime software sound synthesis of walking footstep models. The resulting publication was presented at the EuroHaptics Conference in Spain and included in the 5th International Workshop of Haptic and Audio Interaction Design.
ACOUSTICS
In the field of Acoustic Echo Cancellation, my work at McGill University's Department of Multimedia and Signal Processing focused on the realtime implementation of the Normalized Least Mean Squares (NLMS) Algorithm to study and eliminate problems with echoes in acoustic environments.
