On the BBC Internet Blog, Andy Quested has discussed the various issues around the addition of a test signal to BBC HD. It makes for very interesting reading, well worth it for anyone wanting to better understand their television and get a better picture.
I actually emailed Andy about one part of his article:
"The audio is actually two blocks of wood being banged once a second - nothing to beat the real thing!"
I emailed him to say that in college one of my much respected lecturers (Morgan Jones) proposed using a spark gap as a syncronisation source. Few things in nature are more instant than a spark and the correlation between the light and the light is absolute (subject to the speed of sound, etc).
His response was that he would look at it but he also raised another issue, how do you deal with the fact that an audio compression system based on the psycho-acoustic model might ignore such a short spike of audio? Frankly I don't know, I am not sure I know enough about compression systems, but it seems worth a look.