Steven C. Barr wrote,

 > Is not the analog signal. . . essentially defined by the path
 > followed by the exact centre of the groove? Thus, if an image
 > could be obtained which would allow the determination of that
 > path. . . then the original waveform as recorded could be
 > established and replicated.

That's how I see it.  The waveform is preserved in analog (or at least
continuously variable) form within the scan, and could conceivably be
maintained in that form if desired.

 > In this case, there would only be problems if either the
 > actual surface of the record had been distorted in such
 > a way that the groove no longer followed its original
 > path, or if a portion of the groove was actually missing
 > due to substantial damage to the record surface.

This is why I feel that the scan must be three-dimensional and fine
enough to capture the smallest deviation of the waveform represented by
the groove.  Since a coarse groove is relatively deep, and contains the
waveform information along its entire depth, duplicated on each side, a
three-dimensional scan would allow one to seek an undamaged section of
the groove along its entire length and thus restore the complete
original contour of the groove, and the unaltered original waveform thereby.

Rob Spencer