22 Feb
2015
22 Feb
'15
11:37 p.m.
On Sun, 22 Feb 2015, Andrew Levine wrote:
I think it probably would. The original reasoning was to spot if the end
of a previous encode had been corrupted by a crash / power-cut or
whatever. On reflection it seems more sensible perhaps to work back from
the end of the existing data and find the first frame that has the correct
checksum, then continue the encode from there.
Best regards,
Carl
On 22 Feb
2015, at 19:32, Carl Hetherington <cth(a)carlh.net> wrote:
Yes.
Would it be possible to skip that step? ;-)
DCP-o-matic wrote 46158 frames to disk, 143758 were already created on a
previous run, so did not need to be encoded and written.
But verified with their md5-checksum, right?