Only the brightest stars in the sky will appear on high-speed image-intensified video since the light-integration period is very much less than that of the normal-speed image-intensified video. More often than not, no stars were visible in a high-speed video image sequence. Consequently, it was not possible to determine the high-speed video camera's orientation directly from the stars.
Fortunately, the sprites themselves could be used to determine the camera orientation. This was done by determining the azimuth and elevation angles of sprite features in a calibrated normal-speed video image and then adjusting the high-speed camera orientation parameters (and field of view, if not already set) to produce the correct azimuth and elevation values for the same sprite features. The error in this technique is estimated to be typically no more than a few hundredths of a degree above that introduced by the normal-speed video stellar alignment procedure.