TBC off-by-one output error.

[happycube]

From #789, this is an off by one (or maybe even 0.5?) error I need to fix. I think the intention was to log the beginning of the line at 1/2*SYNC IRE, but it should be outputting once SYNC level is reached.

"(Note that the output is still offset to the right when compared with ld-decode's output for both PAL and NTSC - this is because it's following SMPTE 244M/EBU Tech 3280 by having each line contain all the blanking samples, then all the active samples, whereas ld-decode starts each line with the sample immediately before 0H.)"

[atsampson]

I think it does behave as intended at the moment (the first sample on each line is the nearest sample immediately before 0H, halfway down the leading sync edge - you can't sample on 0H because it falls between two samples for both NTSC and 4fSC PAL).

This is related to #242 - ideally both NTSC and PAL would be sampled at 4fSC exactly, and the output would conform to the SMPTE/EBU standards in terms of the contents of each line and how the samples are phased relative to the subcarrier. The tools can handle output like this already.