Higher Frame Rates Cut Dispersion in Free-Fall g Estimates but Yield Mixed Accuracy Gains in Tracker Analysis

The Echiburu preprint decimated original 120 FPS recordings uniformly to create matched 60, 30 and 10 FPS sequences, thereby isolating frame-rate effects from trajectory differences. Statistical outputs (mean g, SD, relative error, RMSE) were computed for each rate after Tracker fitting of position-time data. Free-fall dispersion fell monotonically with rising FPS, yet bias did not; pendulum period extraction proved insensitive to sampling density because the underlying frequency content lies well below even 10 FPS Nyquist limits.

This pattern aligns with earlier physics-education studies using Tracker that reported diminishing returns above 30 FPS for pendulum periods but continued precision gains for ballistic motion when acceleration is derived from second derivatives. The present work adds the critical control of phase-invariant subsampling at 10 FPS, demonstrating that phase jitter contributes less variance than inter-realization differences, a nuance absent from most classroom lab manuals.

Metrology implications extend to smartphone-based gravimetry and drone-borne structural monitoring, where storage and processing costs scale with FPS; the results suggest system-specific optima rather than blanket adoption of maximum frame rates. Future work should incorporate camera rolling-shutter characterization and sub-pixel centroiding to test whether hardware limits now dominate temporal sampling.