My apologies for taking so long.
No worries - I know how that goes. And I'll apologize for the late response by me too.
I used the same SD card I had been using for all the previous test. In the video there does not seem to be the time shift I have been previously experiencing. Unfortunately it was a cloudless sky and the clouds moving was the easiest way to see the shift in the interval time, but the shadows moving and the traffic seems to be consistent.
I think the pretty much demonstrates that when the log file is "flogging" the SD card with log messages, in addition to the actual image data being stored, every 4 seconds, then the SD card bogs down. I'll look at doing a "delayed write" log file that only updates every so often - or maybe after shooting has completed. The issue there is making sure people shut the script down cleanly so that the log file has time to be written.
There is some flicker in the exposure. It is not the huge jumps I have seen in some of my other attempts but still there. This might be down to the limitations of the camera since it is one without an iris in the lens so if there is any breathing in the focus
In these later tests (unlike kap_uav.lua which tries for the highest shutter speed at reasonable ISO) the camera exposure is running in full "AUTO" mode. Nothing to do with the script. The flashing seems to be some sort of artifact of the camera AE operation - either it is not responding to legitimate exposure changes very well or the scene is actually changing like that as clouds roll over.
