Have begun playing around with -1's GOUT shot data. The idea is to incorporate at the minimum the number of frames per shot. With that data anyone can extrapolate just about everything else.
https://docs.google.com/spreadsheet/ccc?key=0AtjXdIjRQo5NdHlSeExLUExQUGJ2YkdhY2JEWGV1TXc
Row 'T' are the magic numbers at this point. Am in the process of verifying them thru random spot checks. Will most likely use the method -1 did, but was initially trying programs i'm used to and am getting slightly different results. For example, dropping the NTSC GOUT vob into Quicktime, i'm getting a -15 frame difference early on, but the numbers layout accordingly, until I get to one of the points where additional verification is needed. I noticed that often when i'd shift back and forth between frames sometimes the count wouldn't change. So am leery of Quicktime going forward, even though it's one of the few applications which can display frame data, which i've worked with. [the other known issues are three sections are returning negative shot lengths, and there is an issue with the frame counts at vob change points, some may have an additional frame to be added]
Another goal of this process is automatic subtitle recalculation, which requires being able to get the h:mm:ss.fff where the frames are often displayed as a fraction of 1. negative1's list had this data, but was attempting to recalculate it, but am slightly off. There are columns for 23.976, 24, 25, 29.976 fps values, want to see if they match video programs returns. Another problem with macs is they don't play well with FPS, most truncate 23.976 to 23.97 and some to just 24, but who knows what's being calculated internally. (AfterEffects is a pain sometimes because of this rounding shit)
Once one version of the movie is verified, then can begin the process of other versions. Which can then be consendced to just the values of where things change.
If you are curious what CatBus' Project Threepio looks like in a spreadsheet: https://docs.google.com/spreadsheet/ccc?key=0AtjXdIjRQo5NdHBwMXRkbU9ORzY1UWZPY0w0UTVpd0E Similar idea, could all the languages use one set of times, so that it's easier to calc SEs and version variations?