This is more of an idea than a feature request at this stage.
I was just watching a bit of CSI and noticed that skipping past the title sequence required one jump forward, then a minijump backwards (right arrow, then backspace) whereas, Yes Minister is almost exactly one jump forward, Kath and Kim, 2 jumps, then a bunch of minijumps back and so on.
Skipping title sequences is something we all probably do so frequently, we barely notice ourselves doing it, but, I see no reason that Plex shouldn’t do this for us - apart from the obvious implementation difficulties.
I’ve got 3 ideas for how this could be done:
I could raise a request on thetvdb for some extra fields - title sequence start time, title sequence end time, and from there, I’m thinking the work required on the Plex side shouldn’t be too dramatic. The guys over there have always been pretty good whenever I’ve dealt with them, and, they seem to be genuinely committed to providing the best service they can, so I think there is a good chance they’d agree to this.
Much less reliable than 1, but often title sequences are higher volume than the rest of the programme… have plex watching the volume, skip when it goes up for a bit, then start playing again normally when it settles. AFAIK this was/is the method mythtv used to detect advertising. I never used it in anger - does anyone know how well it worked? This would probably be a bit of a hassle to implement I reckon.
For the first couple of episodes, the user hits a button at the start and the end of the title sequence. A second or so of video & sound is captured, then when subsequent episodes are played, Plex watches for a fuzzy match of this start/end video and sound sequence and skips that way.
So…
Does anyone think this is a good idea?
Does anyone have any other ideas as to how it could be implemented?
One problem I can think of with options 1 or 3 - in a lot of my erm… aquisitions… whether or not the title sequence is even there varies between episodes depending on who ‘made’ the TV rip…
Good point, I should've thought of that... although that shouldn't be such an issue with 3... but 3 is also by far the trickiest to implement... we're effectively talking about a video recognition engine there.
I still think it is a good idea, but like many good ideas, it may just be that the cost:return ratio is just too high*.
* Edit - unless some clever pants comes up with a better way to do it :)
I must say… I watch a LOT of tv-series… and all in HD. And almost none of the series has shortened or removed intros. Am I the only one experiencing this?
Atrus, no we are all experiencing that.... Aaron's point is that scene releases may have slightly different start/stop points for the intro sequence, although, if allowed a few secs of either end of where we think the intro is, we'd probably get around that - most releases are pretty close to each other.
Actually - thinking about this a bit more... for scanned content, we know what the title sequence sounds like, so, maybe some sort of musicbrainz thing that scans the audio track and looks for it... throw a little logic in (it must occur within the first 10 mins of the program, it mustn't be more than X seconds long) and we'd be cooking.
It wouldn't necessarily have to be an on the fly thing either - a background task (or seperate app) that runs through the audio of all files and puts the appropriate bits of info in the database would do the job.
