Nyquist - 625/50 analog to digital question
Not sure if this is the place for this, but its sort of a converter question.
As far as I understand the horizontal resolution of an analog picture is determined by Nyquist's theory and the vertical resolution is determined by Kell's theory.
With 5.5Mhz bandwidth the minimum vertical line width is 91nS, so according to Nyquist you need to sample at least every 91ns. So far so good, you will detect that a line is there but what about position? The line could have started, had its leading edge, up to 91nS earlier than the sample and a trailing edge up to 91nS after the sample, even if it is only 91nS wide.
To put it another way, you know there is a 5.5Mhz component but do you know its phase?
Or do you have to oversample to recover the phase information? Kell's theory tends to suggest this is the case. According to Kell with 575 active lines on a TV and a factor of .75, you can only represent 431 vertical areas.