On Wed, 09 Jan 2008 12:17:13 -0000, Roderick Stewart
 wrote:

Rod,

>If the picture only has a temporal resolution of 25 samples per second, or one 
>sample per frame, then how can a visual event be fixed in time with greater 
>accuracy than that?

Andy has laid out the reasons that you might want to adjust the
clapperboard in a frame.  

On 35mm syncing of magnetic stock,  you had a choice of four perf
positions in each frame for sync and you were expected to use them to
achieve the best results.  You can see quarter frame differences in
sync on close shots.  In the digital and timecode world,  I can lay
easily to any one of the eighty timecode sub frames in a frame,  but
normally work to 20 sub-frames (1/4 frame) similar to 35mm synching.

On documentary and other work,  sync to the nearest whole frame is
normally adequate and in a fifty-fifty choice of which frame,  you
should go late.   You never want sound early since we humans can be
adversely affected by quite small amounts of early sound sync,  but
can accept up to two frames of late sync with no problems.

Jim.