SNN / DVS Projects
An event-based sensing project that captures the high-frequency vibration of guitar strings using a DVS camera and estimates the fundamental frequency for real-time tuning — no microphone required.
Event Stream / Vibration Demo placeholder
Replace with a DVS recording of string vibration or an interactive frequency visualisation.
Guitar strings vibrate at frequencies ranging from ~82 Hz (low E) to ~1318 Hz (high e). A Dynamic Vision Sensor, with its microsecond temporal resolution, can capture these vibrations as dense streams of polarity events — far beyond what a standard 30 fps camera can resolve. This project uses those event streams to estimate the fundamental frequency of a plucked string and display real-time tuning feedback.
Events generated by a vibrating string are projected onto a 1-D spatial axis perpendicular to the string. The resulting event-rate signal is then analysed in the frequency domain — using an FFT or autocorrelation approach — to extract the dominant oscillation frequency, which corresponds to the string's pitch.
Unlike frame-based approaches that require high-speed cameras and complex optical flow estimation, a DVS camera natively captures motion at microsecond resolution with minimal data redundancy. This makes the sensing pipeline lightweight, low-power, and suitable for embedded or real-time applications.
Placeholder for frequency estimation accuracy, comparison with a reference tuner or FFT of a microphone signal, and latency measurements. Add spectrograms or event-rate plots here.