It's become plain from various experiments that a method of avoiding (or at least significantly reducing the chance of) single frame skip in synchronized captures, is to initialize the camera components simultaneously. Need to experiment to see whether the Python based startup is sufficiently consistent to make it worth implementing this feature, but if it is additional protocol commands should be introduced to permit shutting down and starting up the camera synchronously.
Come to think of it this may be worthwhile anyway as a debugging tool (e.g. the ability to have cpid relinquish control of the camera temporarily in order to perform some debugging then have it grab it again).
It's become plain from various experiments that a method of avoiding (or at least significantly reducing the chance of) single frame skip in synchronized captures, is to initialize the camera components simultaneously. Need to experiment to see whether the Python based startup is sufficiently consistent to make it worth implementing this feature, but if it is additional protocol commands should be introduced to permit shutting down and starting up the camera synchronously.
Come to think of it this may be worthwhile anyway as a debugging tool (e.g. the ability to have cpid relinquish control of the camera temporarily in order to perform some debugging then have it grab it again).