VidSync uses a Brown-Conrady distortion model with seven radial coefficients (k1…k7), three tangential (p1…p3), plus a flexible center of distortion (x0, y0). It fits these parameters by trying to straighten “plumblines” (lines of chessboard corners) in an iterative minimization. VidSync report two metrics to the user which reflect the quality of the distortion correction:
- Remaining distortion per point (RMS error in pixels after correction).
- Distortion reduction achieved (percentage between initial and final RMS error).
Values of 80–90% reduction are good. Some error is random noise or user imprecision.