DC restoration of an AC coupled video signal can be performed by clamping to some reference voltage, during the horizontal sync periods of the video. This all seems fine in a running system, when the hsync is detected, clamp to the reference for a short time and everything runs smoothly.
My question is: How is the first sync detected?
Without having DC restored the signal when the circuit starts operating, isn't the ADC's output unreliable, hence the sync pulses can't be detected?
In fact, the input to the ADC may well be out of specification: Consider the case of driving an AD9200, with a 1v peak to peak AC coupled video signal. The absolute minimum voltage of the analog sense pin (relative to analog ground) in this chip is -0.3v, with an unrestored video signal isn't it quite possible that a -0.5v is present on the analog sense pin? This could be prevented by adding a diode to ground to maintain a positive difference between ground and the analog sense pin; is this commonly done?
For reference there exist ICs to perform sync detection, for example the LM1881. Is it necessary to use one of these ICs or equivalent circuitry to detect sync pulses prior to any ADC, or can a more simple solution help with the clamp-bootstrapping?