Tobias, You are right. Perhaps the trim operation is not the place where this check should live. Would you say that the "inspect" operation should be a mandatory operation in every workflow. I'd rather put some kind of duration check there if it makes more sense.
Re. making the MediaPackage duration be that of the shortest track, I can't agree with you. Imagine the shortest track is abnormally short due to some kind of problem. You would be favouring the bad tracks over the "sane" ones. Perhaps the inspection operation should include some kind of correction to make all the tracks of the same duration, e.g. repeating the last frame in the shorter tracks until they match the desired duration. What do you think? 2012/5/7 Tobias Wunden <[email protected]> > Ruben, > > > With my proposal, you can specify the "tolerance" or "threshold" you > want. The point is, that threshold doesn't apply to the difference between > the initial duration of the tracks, but to the difference between the > duration AFTER the tracks are trimmed. In other words, I'm implicitly > assuming that the important tracks are those resulting from the trimming > operation, while the original ones are somewhat less important. Then, it > doesn't matter that the tracks differ 5 minutes, if then I set the trimming > point 6 minutes before the end of the longer track, which means those 6 > minutes are unimportant or not wanted. The difference between this and > taking the shortest track duration as the "Mediapackage duration" is that > in this latter case we don't control the interval that's being (implicitly) > trimmed. > > what happens if trimming is not part of the workflow? We need to come up > with a solution that is robust for *every* case, not just for those that > happen to use trimming. From my point of view, choosing the minimum track > length is a valid one, since every piece of code inside Matterhorn has been > designed to expect tracks with equal length. > > Tobias > _______________________________________________ > Matterhorn-users mailing list > [email protected] > http://lists.opencastproject.org/mailman/listinfo/matterhorn-users >
_______________________________________________ Matterhorn-users mailing list [email protected] http://lists.opencastproject.org/mailman/listinfo/matterhorn-users
