Researchers Jacob Haqq-Misra and Ravi Kopparapu examined 112 videos released through the Pentagon's PURSUE program and reported that every public clip is missing at least one quantity needed to recover an object's physical velocity. Their preprint was submitted to arXiv on August 12 and was reported by The Debrief on August 20.
The missing inputs vary by clip but include range to the target, observer or platform velocity, line-of-sight history, aspect angle and the sensor's angular scale or field of view. Without those values, motion across a screen is an angular rate rather than a defensible estimate of miles or kilometres per hour.
The paper uses PR113 and PR149 to show the ambiguity. PR113's graticule permits an angular-rate estimate, but the same image could represent a small nearby bird or a larger distant object. For PR149, a ship in the scene provides a conditional scale; if the target is at the ship's range, the authors estimate an upper bound near Mach 0.4 rather than supersonic performance.
The authors do not claim that the clips are solved. They argue that extraordinary flight claims cannot be tested from the released pixels alone and call for unredacted telemetry, known sensor calibration and coordinated multi-sensor observations. The manuscript is a preprint and had not completed peer review when the report was published.