state machine startup seems messy from screen sensor perspective

[leoscholl]

like the framerate takes some time to settle after the fsm starts.

black dots are clock timestamps, orange is photodiode (analog) and blue is photodiode (digital). clock starts way before the screen does anything, and then the screen has several frames where it does nothing before jittering a lot.

could maybe add a special state at the beginning of the experiment while it waits?

[leoscholl]

added a 1 second delay where only two clock pulses are sent in c198e86819547757ce3a8cdf6fbff16175dab8a9

[leoscholl]

updated sync procedure to have a variable delay at reduced frame rate in ee837fae8853f87e45209d43a49b8499944a94b3 also changed sync procedure so that the clock always reflects the cycle count