Accuracy of flight time and countermovement-jump height estimated from videos at different frame rates with MyJump.

Abstract

Recent improvements in smartphone video technology may provide sufficient accuracy for estimation of jump height via flight time determined from video recordings of vertical-jump tests. The aim of this study is to evaluate the accuracy of jump height estimated from videos at different frame rates. High-definition videos of 10 young adults (6 males, 4 females) performing 5 countermovement jumps were recorded at a frame rate of 1000 Hz and transcoded to frame rates of 120, 240, and 480 Hz. Flight time in the videos was assessed independently by three observers at each of the four frame rates with MyJump. Flight time and jump height were analyzed with mixed models for estimation of means and of standard deviations representing technical error of measurement (free of within-subject jump-to-jump variability) at each frame rate. The four frame rates and three observers produced practically identical estimates of mean jump height. The technical errors at 120, 240, 480 and 1000 Hz were respectively 3.4, 1.8, 1.2 and 0.8 ms for flight time, and 1.4%, 0.7%, 0.5% and 0.3% for jump height. Assessed relative to either differences in jump height between elite football players (standard deviation of ~12%) or the smallest expected test-retest variability (typical error of ~3%), the technical error was substantial at 120 Hz but negligible at 240 Hz or higher. In conclusion, use of frame rates above 240 Hz to estimate jump height with MyJump will not improve accuracy substantially.