Imageres setting on server

Hello.

Had a mishap this weekend where a load of my artwork went AWOL after rescanning a few sources. I’m treating it as an opportunity to get things in order and make sure all TV season art is complete and matches, etc, etc.

This might be a stupid question but can’t think my way through it so here goes…

On a system like mine with a shared MySQL library, would there be any point in specifying imageres 9999, fanartres 9999 and imagequalityjpeg 0 in the advancedsettings.xml on just the server device? In the hope that this would maximise the quality of scraped images before the various client devices rescale them to 1080 as set in their own xml.

The server device isn’t used to watch videos on, so any performance hit due to larger image size isn’t really an issue.

Also, do the client devices take their artwork from the pre-cached file on the server, or direct from the originally scraped file?

Thanks for any help.

It doesn’t really work the way your thinking. Each Kodi profile maintains a thumbnail cache and this cache is indexed by a local database that is seperate from the one that your sharing via MySQL and also cannot be shared. Because there is no way to actually keep the cached artwork in sync with the databases on each client even a shared MySQL database this means that one can’t maintain any kind of server level artwork settings or files. The source of the artwork is local files that are placed with the media (and named correctly) and if those are not present they are downloaded from whatever site your scraper is set to.

That setting your referring to for the artwork is the size the cached files are stored in. You likely don’t want to set it like that. See this thread…

The process you seem to be talking about in practice would simply be storing full size artwork with the files. This is fine but will have no difference to the end result over Kodi grabbing them online when it scrapes. Having local files will make it scrape faster but grabbing full resolution artwork online and full resolution local will still start with full resolution artwork that will be cached on each profile using whatever settings set for that profile.

If you want to have a more resilient setup on how you have your files on your server you might consider a media manager such as Tiny Media Manager that will make all local metadata that Kodi would then use when scraping. It also allows for easy customizations such as changing movie sets or custom titles.

Thanks for the (again) detailed response.

Now that I know that artwork is handled in the way you describe, the behaviours I’ve been experiencing make more sense.

One question though. When I change a poster for a particular film on a client, the change is (usually) relatively quickly updated on all devices. What’s going on there?

I did read the thread you linked when it was posted but hadn’t reviewed it again after encountering my issues so will revisit it.

On the topic of TMM. I’ve dabbled with it a little before but always felt it was another hoop to jump through in order to try and keep things tidy. However am I correct in saying that I would create all artwork and nfos via TMM, save them in relevant folders alongside the media , then set the scrapers on all devices to local info only, so they would all use the exact same files to then add to their own cache?

The main database stores the file path to the artwork. If that file path changes then Kodi will update the artwork. Note that if you are using local artwork and you swap it out (ie the art changed but the file name was the same) this may not force an update, or an update may require 24 hrs since it changed it last before it will recache.

Although there are some people who scrape local metadata only you certainly don’t need to use it this way. If you use a regular scraper it will only grab metadata from the internet if local files are not present. Thus, you can selectively add local metadata just for what you want to and let Kodi do it regular thing for everything else. Personally I use TMM for all the movies I add and usually for shows that are already finished. For TV shows that are new/current I normally will use TMM to create metadata for whatever I already have but let Kodi scrape online for anything that gets added later.

My rationale for the way I do this is that I want my movies and shows to have artwork I have chosen. I also want to make sure that I am getting the correct metadata as it really isn’t uncommon with a big enough library to have items with the same title. But once you scrape a show it adds in the online scraping source so you don’t have to worry about additional episodes getting scraped from the wrong show. As for the older shows thing an example I just dealt with is adding American Gothic to my library. I already had the newer version of the show and adding in the older one came with some issues. In this case I use TMDB scraper in Kodi but that site has the unaired episodes listed as specials. I didn’t want this but with TMM I could just scrape it there using TVDB which has them listed in the intended order all as one season and also rename both shows to add into the title the year they premiered so It is rather neat, tidy, and clean in Kodi’s listing. If I ever have to re-scrape again I don’t need to muck about at all with either show. It was an extra step but much quicker and easier to do with TMM than inside Kodi and a Kodi export (to get Kodi to output the nfo and artwork) would require some special considerations to get full size artwork saved to my server.

NOTE: With the older (free to use) version of TMM I’m using when I scrape a show it adds in id tags for TMDB, TVDB, and IMDB. If a TVDB id is in a show then it saves its url as the data source in the tvshow.nfo file. If your using TMDB scraper in Kodi this will prevent it from scraping any episodes that don’t have their own nfo file. To fix this issue you just need to open the window for the show itself and delete the TVDB id and that will make TMM set the TMDB url in the nfo file. With older shows that you make complete local metadata for in TMM there is no need as you wouldn’t ever be adding in new info to scrape.

Thanks again, will give this some thought. What ver. of TMM are you using? :wink:

3.1.18 which is the last version before they switched to the limited/pro paid version scheme.