Tvdb changes to a subscribe model

My issue TMDB is I don’t think it’s as accurate and I don’t like that you can’t set individual runtimes for episodes. So stats will remain wildly off. Not such a big deal for regular TV. But I watch a lot of independent news and those episodes don’t fit in a nice time lot of 21 or 41 minutes.

This gonna be a gigantic mess.

2 Likes

@LiaraT Shows like https://trakt.tv/shows/the-haunting would be split into separate shows to match TMDB. This requires some manual moving behind the scenes that we’ll handle as part of our support process.

We also think contributors will start adding to TMDB along with TVDB, or move to TMDB completely. I think many smaller projects will be moving to TMDB since the TVDB business model won’t make sense for those type of apps in my opinion.

@pauls67 we’ve seen a lot of improvement when it comes to currently airing shows. Those mostly match up with TVDB and stay up to date. Older releases and less popular is where I envision the most missing data.

@exuberantShadow we use TVMaze as a fall back for the show air day and time, but that is all currently. I don’t anticipate using TVMaze as the primary, but it might make sense as a secondary source.

@bastardsheep I assume Plex will come out with a rebuilt TV scraper, like how they did for movies. I’m sure they’ll use several sources of data to build their ultimate Plex source. I agree that specials are going to be an issue, those seem to drastically differ, even if the other seasons and episodes match up.

@Parzival we’re also thinking of ways that Trakt can be helpful for TMDB, maybe by providing some TMDB <> TVDB mappings, things like that. From what I know, TMDB’s moderation process will be able to handle this and they have a good team of mods.

@anubis81 That’s a good question, I’m not sure how long people spend editing the data. From what I’ve seen, most people have a handful of their favorite shows and really focus on getting those all up to date and accurate.

@JBaby_9783 we don’t currently have runtime from TVDB with their v3 API. The v4 is supposed to include it, but as of now Trakt doesn’t get the benefit so TMDB won’t be any different in that regard. I do agree this is a must have feature though, and according to Trello it is “started”. With some additional comments, maybe TMDB will get it implemented sooner then later :slight_smile:

10 Likes

There is one issue which I was facing few weeks ago. In your conditions (to fetch data from TMDB) is that there have to be the English names/translation available. I did that for one local TV show and some TMDB admin (superguru) deleted whole my work :frowning: With the explanation, that this TV show was never aired in the USA/UK etc. I should stop with putting confusing English names/labels for the show/sessions/epizodes and keep just and only those in local language which correspond with the TV show. With the reference to some their conditions.

So, either there will be some alignment between trakt.tv and TMDB or you will allow to import whatever TMDB show regardless the availability in english language or the track.tv will finish as world-wide service and will be limited just and only to English users :frowning:

1 Like

What is the TMDB link for the show? I can do some testing and see how it imports. We handle translations a bit differently with TMDB.

1 Like

No kidding. :grimacing:

Also TMDB has their incorrect content reporting system working way better that TVDB ever did.
I have always added to both TVDB and TMDB and while TMDB may not have as much cataloged, I believe it does have better systems/rules in place.

Specials have always caused issues when you are trying to add correct data anyway because the difference causes conflicts so standardizing this will probably fix some episodes at the same time. :crossed_fingers:Watched status of those episodes may be complicated to figure out though. :thinking:

With a better implementation imo.
Quote: “Episode runtimes would default to the show runtime but could be overridden for special episodes.”
Whereas you have to manually set it for each episode on TVDB.

1 Like

Oh and btw TMDB handles images/videos/actors waaaaaaaay better!

3 Likes

My two pennorth.

Always been incredibly resistant to this idea but the “safe” approach that Justin has outlined for the transition sounds great. It will be interesting to see how this would work in practice. I would certainly be more than glad to get involved in getting TMDb up to scratch if this would help.

@justin my only concern would be the loss of historic watch data where the two sources disagree about ordering. Is there any way we can “preserve” historic watched data so it aligns with the TVDb database so nothing is lost in these cases? Or would this be a massive headache technically??

5 Likes

Yes, we could do that if needed. We’d just need a way to “lock” a show which basically stops it from getting any updates. The downside is even old or ended shows are still incomplete in some capacity (translations, actors, guest stars, etc) so we’d lose that ability. But, I do get your point and if it’s an extreme case or TMDB just can’t be adapted to match then we might need to “lock” it.

1 Like

Sorry but tmdb is not in any way a replacement for tvdb

I wish it wasn’t this way but there is a reason everyone uses tvdb despite the issues

Also this is going to cause havoc with scrobbling with plex no matter how you end up doing it

I say this as someone who has been subbed to vip for 7 years and have been very happy with the decisions that have been made up to this point

4 Likes

I know Trakt doesn’t have runtimes. But TVDB is going to implement them. The problem is not every show fits nicely in one time period. This is a great example. I’m sorry Justin but I’m not convinced this is going to be ok. In fact the more I think about it, the more I think this is all going to go horribly wrong. I know TVDB support is hot garbage, but the database is just so much better than TMDB. I apologize for being a Debbie Downer.

I completely agree. This is a mess. I’m sick just thinking about it. I maintain 5 shows. They’re news programs so they air 5 days a week. Going back to input all of these episodes is a task I’m not certain I want to take on. The more I look at TMDB the more I think it’s seriously lacking. I’m not saying TVDB is perfect. It’s not and the support there is generally trash. But the database is way ahead of TMDB.

3 Likes

Overall there is an easy answer to all that: if Trakt sticks with tvdb it’s not going to be a free service to everyone anymore, simple as that.

So, since Trakt relies on a third party api and the devs probably want the service to stay free for all, changes are inevitable. Right now, from an end-users point of view, tmdb is the obvious and best choice. It’s running stable, they’re already working with it, it’s easy for everyone to collaborate.

Of course there will be hiccups and problems in the transition, and I’m not looking forward to them as well but from a business perspective I totally get it and the sooner one starts working on it the better.

6 Likes

@mcrommert It’s hard to say what Plex is planning to do. They might be moving away from TVDB for all we know. It’s worth noting that Kodi has used TMDB for their default TV data for a few years now. With community help, we think TMDB can become more accurate for Trakt and a bunch of other apps who can’t or won’t pay the TVDB licensing fees.

@JBaby_9783 Your concern is totally valid. With TVDB’s proposed licensing model, we just don’t see how that will work moving forward with them. If Trakt pays an application based license fee, it’s going to be a huge expense that we feel is spent better elsewhere (i.e. mobile development, dedicated support person, etc). Having user’s pay the fee is a non starter since they are essentially asking all Trakt user’s to pay TVDB for access. That leaves us with option 3: find a new database like TMDB. We still have some time before TVDB shuts off their v3 API in March 2021 and we’re building some comparison tools to get a sense very soon what data differences we’re dealing with.

@Parzival yeah, that’s what we are thinking too. TVDB’s licensing model makes it tough to use without passing on more costs to the Trakt user.

4 Likes

I’m working on some tools to compare TMDB and TVDB data. Here’s what it looks like so far with comparisons for seasons and episodes between the 2. This isn’t a perfect way to match, but should be a good starting point. All seasons and episodes are linked to their respective databases for quick access.

Please let me know if there is specific info that would help with the comparison page. Thanks!

9 Likes

Cool, Is it possible to bulk move users watch status/ratings/lists to the correct episodes from TMDB? That way no one’s history/ratings/lists are messed up. Or do you have a way to handle this already? I ask because specials have a lot of episodes that are in a different order.

1 Like

I understand. I just don’t like it. I realize I’m in the minority. I get that most of us are geeks. So we can handle this. But I’m thinking of my Muggle mom and sister. This is gonna be a sh*tshow for them.

If I could throw more money at this I would.

This is awesome. And I bet for regular shows this will be handled. But I’m thinking about all the obscure shows. The once’s that don’t have tons of fans who will help make corrections. I’m a crazy news junkie who Trakt’d every episode for the last 7 years.

1 Like

Specials will for sure be the most annoying episodes to deal with. I’m not sure of the process yet, but right now it’s something we manually handle and move episodes to the correct place.

I’m hoping to release this update next week so we can start gathering stats and get a sense of how many shows match already, match except for specials, and ones that don’t match up otherwise.

4 Likes

Is this also going to be available at the episode level?
Need to know if episodes have a description and also if the aired date/time is correct (based on country of origin), or as good as it is currently
Thanks

2 Likes

I wish someone could just pull a dump of thetvdb and go from there

1 Like

I have been lurking this topic for a while and haven’t chimed in due to other obligations but here’s my take:
TVDB is currently a far more superior database for TV.

However, I understand where @justin is coming from and Trakt would not exist with the upcoming changes to TVDB’s licensing model.
Though I got to admit, it would be great if all shows and movies are from one database which will make editing much easier in the future. Now, I edit both TVDB and TMDB so eliminating one would save me a lot of time.

I was worried about the sudden changes to current shows (like @deemur) but the compare tool looks awesome. I am not opposed to locking shows to ensure the data stays correct, even if that means that some minor data is missing.
@justin Will the transition tool also compare individual episodes or just the series on a global level? There could a lot of differences there and we should be aware of those changes. Like @deemur, I would be more than happy to assist in the transition to TMDB.

However, I was really excited about getting accurate runtimes with the v4 TVDB API (along with @JBaby_9783). It’s a good thing that TMDB has this feature on the roadmap. Since the rise of streaming, episodes have such a varying runtime (one episode could be 25 minutes, the other 34 minutes for example). Hopefully, Trakt will implement this feature as soon as it is available.
@arcreactor I like TVDB’s runtime solution as with streaming every episode has a different runtime. I can’t find your quote in the Trello board but I hope that we can freely edit the episodes runtimes. For broadcast and cable shows, TMDB’s approach might work but not for shows on streaming services.

4 Likes

This is my worry too. I don’t watch a lot of mainstream news. Online shows don’t have to stick to TV rules. I have episodes that are 15 or less. Others are 75 minutes or more. As of now I don’t at any attention to Trakt stats because there’s no point. And that kills me because I’m a stats junkie. I love tracking all the things. You should see how obsessive I am about tracking my reading.

3 Likes