Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[rick.ca]

I actually don't agree here.  I mean, of course, that would be better than nothing.  But what I'd really like to see is the TVDB update to run continuously in the background, like AutoImport.  So that, if you change the tags on a file that was mislabeled, it would just "automagically" pulled down the other relevant tags for you in the background and within 30 seconds or so of the tag change the new data was just "there".

Have you read the API? I believe the data is only updated once a day, so a daily update is going to give you all the data available. That was the whole point of my recommendation. It's pointless to do separate lookups for every episode when all the data currently available can be downloaded in one, simple daily update. And yes, correcting a filename would obviously result in that file being correctly matched to it's meta data—instantly.

I figured.  Unfortunately, I think this will be of somewhat limited usefulness for most content for "many users" for the reasons outlined above.

Not only will it not do the additional things I was hoping for, I don't think it's going do what Matt expects it to do very well. It will be inconsistent, require too much user fiddling, and be more annoying than useful.

Even if it's only going to be used to fill fields for files as they're imported, it makes more sense to have the user specify the series they want and then maintain that data via the API separately. From a user perspective, that means they set-up a series once, and then never, ever have to concern themselves with episodes of that series not being properly identified or having the most current meta data.

My experience suggests very few users are interested in messing around with meta data for current episodes. They just want a little information (a screen shot, synopsis, guest stars and maybe director and writer), but they want it to be fully automatic.

[glynor]

Have you read the API? I believe the data is only updated once a day, so a daily update is going to give you all the data available.

That's not my goal.

My goal relates to a hypothetical file I have that are untagged, so it has NO information from TVDB of any kind (hypothetical because that never happens, right?).  Or maybe if I have a file that is tagged incorrectly (like it has the wrong episode number, or is mislabled as the wrong Series or something).  I'm looking for a system where when I fix the tags so that MC can look it up "properly" in TVDB, the related "cloud" tags just "appear" automatically and I don't need to manually resubmit them for updates or wait until the next "update cycle" that day for it to grab the information off of the cloud.

So, for example, imagine I have a file that is tagged as Lost, Season 2, Episode 11.  The TVDB system would auto fill the information from this page.  But then, after checking, I realize that because they aired the weird Lost "Revelations" recap show in between episode 9 and 10, my episode labeled 11 was really actually episode 10 (and what I had labeled as 10 was actually a mislabeled "special episode", and all of the other episode numbers from that point on were off by one).  So, in this instance, ALL of the episodes of LOST in that season after #9 actually have the wrong descriptions right now.

If I fix this by renumbering the episodes, I'd like it to automatically in-the-background go out to TVDB and fix all of the other related tags without having to manually "resubmit them" or wait for the auto-cloud-check process to happen in the middle of the night (or whatever).  Likewise, what about new content?  When Auto-Import picks up a new file, and I go in and tag it as Fringe Season 3, Episode 15, I want it to automatically "know" within a few seconds that the episode's title is "Subject 13" and fill the description field and whatnot.

So, I'm looking for more of a "live" connection to the database, rather than a pull-from-cloud on a scheduled basis type of system.... Much like the difference between POP3 email and Exchange email.  Not because the TVDB database is out-of-date, but because my information may change and I don't want to have to wait for a scheduled "sync".  Obviously, we're not really looking at having an auto-lookup system of any kind right this very second, but I'm looking towards the future.

[glynor]

From a user perspective, that means they set-up a series once, and then never, ever have to concern themselves with episodes of that series not being properly identified or having the most current meta data.

I don't know for sure at all... But I'm thinking that you could accomplish this on the fly via a "learning" type of system.

For example, what if you just get some episodes of a new show in the database and you tag them as [Series]=Game of Thrones, like you normally would.  Since MC has never seen that particular [Series] tag before, it shows a one-time "match" dialog, where you pick what fields you'd like it to import from TVDB for that series, and makes sure that what it finds on TVDB actually matches the files you have on disk (by asking the user, like the Wikipedia match dialog).

Then, once you've filled this out, it automatically "remembers" the settings and just handles any episodes tagged similarly in the future.  If you later tag an episode as "The Game of Thrones" instead of "Game of Thrones", it will throw up the match dialog again (cluing you that something is "wrong"), and perhaps even suggesting an alternate tag from the files already in your MC database.

[MrHaugen]

So, if you have to start it out slow I would suggest you focus on a few things first. This is all from a point of view that I think would benefit most users. It is not by any means my dream setup or my goal, but it might be a starting points and something you can work on.

Things I think are essential:
- Add standard fields for series, like "Series Overview" (relational field), "First Aired date" (relational field). Things like Season, Episode, Genre are standard
- Add a way to set Series on a Per Folder basis under Auto Import. Mark video files as Media Sub Type: TV Show. So it can be applied to a view directly.
- Scan this video files under auto import, for a predefined season and episode pattern. Example: sXXexx (not case sensitive). People have to rename files if they want it to be picked up automatically. Also scan for Series name. Strip ".", "_" and " - " and the previously named S/E pattern and everything after it.
- Match it to theTVDB, and include only info for files imported (for now at least. If that's simpler than adding whole series and marking things as local or not)

For better Theater View experience, please consider:
1. Allow the info pane to be active on all levels in Theater View (series and season), if set in view options. Adding your own info panes, adding different meta data to each, and selecting which levels and in which views they are used, would be the ultimate dream I think. Then you could add different panes to different media. Or possibly expression fields so you can add different info depending on the media. It would not make a difference in Series and Season pane though, as the media is the same, only the hierarchy level changes.
2. A Logic to count Series and episodes, together with a watched not watched number, so we can get a small summary (including Series Summary, Genre etc)
3. Changing backgrounds when you either select a Series, or when you hit enter on a show. Images can be random if there are more than one?

You COULD also base the theTVBD lookup on Series, Season and Episode tags, but that really don't make things any more automatic. You can get a summary, aired date and perhaps some nice images, when you first have tagged it. So it's not a complete loss! But a bit half hearted in my opinion.

[glynor]

- Add a way to set Series on a Per Folder basis under Auto Import. Mark video files as Media Sub Type: TV Show. So it can be applied to a view directly.
- Scan this video files under auto import, for a predefined season and episode pattern. Example: sXXexx (not case sensitive). People have to rename files if they want it to be picked up automatically. Also scan for Series name. Strip ".", "_" and " - " and the previously named S/E pattern and everything after it.

I think Matt was hinting that this type of system is a bit out of scope for right now, though certainly not impossible in the future.

A system that was able to parse filenames on import to pull in tags automatically would be, of course, awesome.

[Matt]

I think Matt was hinting that this type of system is a bit out of scope for right now, though certainly not impossible in the future.

A system that was able to parse filenames on import to pull in tags automatically would be, of course, awesome.

Yes, I think that could be a really neat feature.  I'm not a big downloader of videos, but I know a lot of our users are.  This would help them out.

With that said, I think metadata lookup is a separate thing from filling file properties based on rules, possibly at auto-import time.  We already have a 'Fill Properties From Filename' tool, and I think improving that (expressions, profiles, automatically guessing, showing a preview, running automatically, etc.) would be great someday.

[rick.ca]

Not because the TVDB database is out-of-date, but because my information may change and I don't want to have to wait for a scheduled "sync".

It seems you're still ignoring what I've been saying from the beginning. The processing of each file to add/update meta data is obviously completely separate from the once-a-day downloading of the bulk meta data. If all the meta data for requested series is already available on your HDD, any means by which files are updated (as they arrive, as you request and update, or, if you prefer, as differences are automatically detected) is going to be faster and more reliable. You're unlikely to detect any difference between data pulled from the download versus that from the cloud, other than the latter is much more likely to include user-contributed crap the moderators (or system processing) have not had time to correct.

But I'm thinking that you could accomplish this on the fly via a "learning" type of system.

Obviously it's possible. It's just not necessary, considering the frequency at which users typically add new series and the value of user-specified configuration settings in ensuring that series is always correctly handled. Such a capability is certainly not a substitute for the latter, and it seems that's how it's repeatedly but put forth here and how Matt intends to implement it. It's never going to be more than a novelty—like the Wikipedia system.

[rick.ca]

Yes, I think that could be a really neat feature.  I'm not a big downloader of videos, but I know a lot of our users are.  This would help them out.

What you mean by this is unclear. Surely you still intend on looking up based on the filename. Requiring a file to be tagged before getting meta data to tag it with seems rather pointless. Perhaps all you're referring to is what some of us call "auto-tagging." I agree that's a completely separate matter (for future consideration). I don't see it as unique to video—it has useful applications regardless of media type.

The persistent requests to automatically set [Media Sub Type] suggests an unfortunate reality. Many users seem unable to "think outside the box" and understand the program already offers plenty of capability such that there's no need to be dependent on a tag been set to handle the file properly. Even an auto-tagging system would have to get the data from somewhere. In most cases, that same data is accessible from existing MC functions. The obvious example of this is the insistence [Media Sub Type] be set based on the folder a file is in. If the folder a file is in indicates the media type, then that attribute can be used directly to organize the data. If the file is to be moved from it's incoming folder, then the destination folder is probably the one to use. But in the case of files that have been recognized as episodes based on their filename pattern, even this is not essential. Surely these files would at least be tagged with [Series], [Season] and [Episode]—and the existence of those tags would be sufficient for organizing them. This is not to suggest an auto-tagging feature would not be useful and make things overall easier to configure. But it's certainly not essential in the context of this feature.

You'd be doing many users a huge favour by processing episode files based on their filename, and providing default views which categorize and display the files based on this information. It would be difficult for anyone to believe the program is lacking any essential capability if that much happens automatically.

[glynor]

It seems you're still ignoring what I've been saying from the beginning. The processing of each file to add/update meta data is obviously completely separate from the once-a-day downloading of the bulk meta data. If all the meta data for requested series is already available on your HDD, any means by which files are updated (as they arrive, as you request and update, or, if you prefer, as differences are automatically detected) is going to be faster and more reliable. You're unlikely to detect any difference between data pulled from the download versus that from the cloud, other than the latter is much more likely to include user-contributed crap the moderators (or system processing) have not had time to correct.

Sigh.  Please stop accusing me of ignoring you.  I read what you've written very carefully.

I just think my ideas are subtly different.  Whatever... We've all made our points and J River has gotten a lot of the feedback they were looking for.

[rick.ca]

Sigh.  Please stop accusing me of ignoring you.  I read what you've written very carefully.

I'm sorry you find that annoying. But you referred to a point of my proposal, stated you disagreed (which is fine with me), but then provided a reason for your disagreement that had nothing to do with the point. In case you haven't noticed, I'm not inclined to ignore such things, and that's not going to change. In hindsight, I suppose you were trying to express a completely separate, independent option. If that's the case, you might have presented it as such—without making reference to my proposal.

[fitbrit]

I think that my daughter, who is going through her terrible twos, would argue just like this were she a seasoned user of MC.

"Do you want some milk?"

- "Yes! Wah!! I want milk!"

"Here you go, here's some milk."

- "No! Wah! I want it in THAT cup!! Wah!"

"Here's the cup you wanted."

- "No! No! No! I don't want milk! Wah! Wah!"

[glynor]

I think that my daughter, who is going through her terrible twos, would argue just like this were she a seasoned user of MC.

 

For the record... I like rick.ca, and I certainly respect his opinions.  I just tend to argue passionately for my ideas, and I assume that he does the same.  Hopefully he looks at it that way too.

[rick.ca]

I do. And I would never consider hiding the fact I love a healthy debate. But I do think we could try harder to stick to reasoned discussion, especially when we know there are bound to be differences in opinion. That's why I suggested outlining a proposal in it's entirety. That way, everyone can look at the pro and cons of alternate proposals. Debating those is quite different than arguing over what may be a keystone in one proposal—because it's inconsistent with or unnecessary in some other (not fully disclosed) proposal.

This is not to suggest that one can't question, comment on or even reject some aspect or part of another proposal. But in doing so, they should respect the fact any discussion about it should be in the context of the proposal from which it came. And when a question is raised (e.g., as many were about pattern recognition and regular expressions), the answer should be accepted unless inconsistent with one's personal knowledge. Effective debate requires that both parties have some knowledge of the subject matter.

[raldo]

Effective debate

-- is less wordy.

[rick.ca]

-- is less wordy.

Not always. I didn't fully understand either of your cryptic posts, and found it annoying to have to guess what you might have meant. I would be happy to make an effort use fewer words—if you'll make an effort to use more.

[raldo]

Not always. I didn't fully understand either of your cryptic posts, and found it annoying to have to guess what you might have meant. I would be happy to make an effort use fewer words—if you'll make an effort to use more.

Here we go again. So why didn't you just ask what I meant in my first post instead of blowing it off? Weird.

Interpreting the filename contents and pushing the metadata to tags is one of the things that must be solved first. Once that's been done, MC can apply some logic in their lookup to match the file with online metadata. (Or, match the file with previously downloaded metadata). That's what Matt said in his last posts too, I guess.

[rick.ca]

Here we go again. So why didn't you just ask what I meant in my first post instead of blowing it off? Weird.

I didn't blow anything off. I responded as best I could to those and other posts at the time. As to whatever point you were trying to make, I'm sure that I made my thoughts perfectly clear in one or more wordy posts.

[MrHaugen]

Yes, I think that could be a really neat feature.  I'm not a big downloader of videos, but I know a lot of our users are.  This would help them out.

With that said, I think metadata lookup is a separate thing from filling file properties based on rules, possibly at auto-import time.  We already have a 'Fill Properties From Filename' tool, and I think improving that (expressions, profiles, automatically guessing, showing a preview, running automatically, etc.) would be great someday.

Then it will not be automatic at all then  Back to the tagging each time I rip or download a new season. At least we can assume some things like this will be improved over time, if you ever implement something with theTVDB.

[JimH]

I want to ask that everyone stick with the substance and avoid remarks about style.  Life is short.  Be nice.

[MrHaugen]

So, how this going? I've been very eager to check builds that indicates work on this field lately, but have seen nothing so far. It's still in the planing stage, or are you stuck on certain things? I'd be more than happy to brainstorm with you if it's needed at all.

[JimH]

We have our hands full at the moment, trying to build a work-around for the IE9 problem in case MS doesn't fix it.

[MrHaugen]

Yes, there have been quite a few changes there, and I understand the priority.

[jmone]

Any more on this?
Thanks
Nathan

[Matt]

We added support for TVDB backdrops to Theater View recently.

Metadata lookup is on the list, but a ways down.

[fitbrit]

We added support for TVDB backdrops to Theater View recently.

Metadata lookup is on the list, but a ways down.

I love the backdrop feature. I'm looking forward to the metadata as well... copy-pasting text for each episode is soooo tedious. 

[leezer3]

We added support for TVDB backdrops to Theater View recently.

Metadata lookup is on the list, but a ways down.

Any chance of caching these please?
This was requested in the main forum, and this is pretty logical step for those with bad / no internet connections.

-Leezer-

[newsposter]

As long as it's a long-term deal.  Longer than 2-3 years.

Nothing worse than delivering a new feature dependent on an external data provider and then have that provider pull the rug out from underneath everyone.  Unless the external provider starts charging for the data.

I've seen both scenarios happen over the years with providers of TV (and radio) programming meta-data.

[Matt]

Any chance of caching these please?
This was requested in the main forum, and this is pretty logical step for those with bad / no internet connections.

-Leezer-

Coming next build:
NEW: Theater View theme images that are downloaded are cached between runs (previously it used a runtime only cache).

[leezer3]

Coming next build:
NEW: Theater View theme images that are downloaded are cached between runs (previously it used a runtime only cache).

Neat
Are these part of the thumbnail database, or are they cached in the appdir somewhere? I already copy the library & the thumbnail DB between machines, and these are a logical addition to the copy process.

-Leezer-

[JimH]

bump

[JustinChase]

Okay JimH, you got me!

I saw this was transferred over from the beta board, and I came over to see what the new news was, and kept thinking 'doesn't MC already do this?'

I thought I was losing my mind when I saw rick.ca talking about pulling in data for shows not on disk, and thought, 'this idea again?'

Then wondered when he came back to the boards, then finally saw this was all from 2011.

Doh!!

[RoderickGI]

I thought JRiver may be going to integrate lookup to TheTVDB during the EPG load, or immediately after to enhance the EPG . . .

But alas, no. I saw the dates straight off. History, and a little forum tidy up.