Character Poster Sets

What counts as a set

A set is at least two uniform posters from one advertising campaign, with no maximum. The largest I saw was 32, for Avengers: Endgame. A normal poster is unique in its layout, and a character poster shares its design with its siblings. These were the rules for the catalogue:

• English first, UK and US, with exceptions where a mix of languages is intended.
• Same resolution and aspect ratio within a set. Sets may differ from each other.
• No ultrawide and no square posters, because many skins stretch posters and these look bad.
• A deliberate order. If the campaign gives none, the TMDB cast order is the fallback.

How it works in Kodi

Technically the sets are numbered posters, poster# in the Kodi wiki, where they have a section under Artwork types as a semi-official type: several posters per film, a fixed order, and more than one set per film possible.

This only works with local artwork, the same way ArtworkBeef's multi-poster does. The naming follows your main poster:

• Main poster poster.jpg: use poster1.jpg, poster2.jpg and so on.
• Main poster Moviefilename-poster.jpg: use Moviefilename-poster1.jpg, Moviefilename-poster2.jpg.
• Films not in their own folders must use the prefixed form.
• Choose one scheme and do not mix them.

Getting the files into the Kodi database took ArtworkBeef on Kodi 17 and 18, Artwork Dump on Kodi 19, or an advancedsettings.xml entry from Leia on. Recent Kodi versions need hardly any of this: place the files correctly, and a supporting skin shows them.

The skin has to take part. It calls ListItem.Art(poster1) and onward, usually in a multi-image control, and offers a toggle. Random mode is the simple one. A plain multi-image sorts correctly only up to poster9, after which you get 1, 10, 11, 2. A true order therefore needs a list with a single image control and a toggle between the two modes.

Skins that took it up include AEON TAJO and AEON MQ 8 TAJO by manfeed, as far as I know the deepest implementation, wall views included; AEON MQ 7 Krypton Mod, Leia Mod and MQ 8 Mod by latts9923; and Aeon Nox: SiLVO by mikeSiLVO. bsoriano, who maintains Amber, asked how to add support, and that question produced my skinning guide.

The plan, and the people who made it real

This is pre-AI work: handmade community work. I first tried the idea in my own skin for a few months as a case study. Then I opened a read-only catalogue thread with one post per set, naming the film, the set number, the type and the size, and a second thread for discussion.

The plan was a chain: content, database, API, artwork provider, skin. I could do the first link alone, up to about 250 to 300 films, and said so. After that I needed the community.

rschiks built the second and third links, character poster support in KyraDB with its website and API, and imported the sets. My base pack ended at about 240 films, 373 sets and over 2,000 posters. rmrector's ArtworkBeef was the planned fourth link. It gained KyraDB support, but bulk use of the sets through it stayed uncertain.

None of the poster art is mine or Konon's in the original sense: it comes from the film studios' campaigns. My part was the concept, the conventions, the order logic, the sorting, the first catalogue, the skinning guide and both threads.

Passing it on

Then Konon arrived, first with sets that were missing from KyraDB, some of them better than what it held. He had a great eye for content, and when I stepped back I asked to hand him the project.

Forum rules made that harder than it sounds. A thread cannot change owner, and a first post cannot be replaced by a later member's post. The moderators Karellen and DaVu explained the rule, Konon proposed his own list, and that is why three threads exist: my catalogue, my discussion, and Konon's list, to which mine now links. My terms were simple: "as long you are interested in i will not take it back". My last advice was to keep originals original and to give uniform sizes only to polished versions.

He has been at it ever since, with one long pause. Konon introduced the split into three list types, which I praised:

• Original: the posters as released.
• Retouched: 2:3 ratio, with billing blocks and dates removed, and some foreign posters translated. Every retouched entry I read from the later years is 2,000 by 3,000 pixels.
• Keyart: 2:3 ratio.

He works by hand in Photoshop and Illustrator, without AI. Counted from his list at the time of writing, it holds 401 films, 612 sets and about 3,624 posters, and it is still growing. Every post in the catalogue thread since my last two additions is his.

Where the chain broke

Manual work. People liked the sets but found copying them film by film too much. That was the most common feedback. Users without XML skills could not use the project at all until their skin supported it.

The last link never closed. The sets reached the API, but a one-click import of everything never came. Then KyraDB was reported down, and the content links ran through it. API and database support are still missing, and manual copying stayed the way in.

Hosting. After KyraDB, no platform really fit. Konon and I went through DeviantArt, MediaFire, Imgur, Mega and Google. The forum's image host stops at 1 MB, DeviantArt hides downloads behind a login, which I found out late because I always saved by right click, and cloud storage costs money or runs out. I also tested heavy compression at equal resolution and saw clear loss in the lettering, so uncompressed archives stayed.

Skin engine quirks. A multi-image with a visible condition flickers on every list move. My workaround keeps the control off screen and slides it in after a delay, tuned to the slideshow interval. Together with the sort limit after poster9, this was the hardest part on the skin side.

Mixed shapes. Konon showed that originals in other ratios distort when they share a slot with 2:3 posters. That is a compromise you have to live with, and the retouched versions are the better answer.

The push of a button stayed a goal

The idea itself, the conventions and the ordering all held. The naming needed no change in Kodi, it made it into the wiki, real skins adopted it, and the catalogue has survived every change of platform. Konon's DeviantArt is the ready solution today: one archive per set, already named poster1.jpg onward, so with one folder per film you just unpack it, and with all films in one folder you add the film name as prefix. An account is needed, and the forum copies stay compressed to 1 MB.

The push of a button never came. The whole chain from content to skin did not come together in one piece, and the original goal of one action stayed a goal. An experimental variant with actor label matching stayed an alpha, and the manual actor folder route that nordicguy and I tried gave good results but meant a lot of manual work. Konon's keyart publication is also paused until new storage exists.

A project like this stands on its content and on one person who cares for it, more than on its pipeline. A catalogue can outlive a database, and a good successor is worth more than a good plan. Konon's list is the proof of that. The same posters and the same names now live on in my Jellyfin plugin ArtworkPlus, as Extraposter and Extrakeyart, and its Sequential order setting exists because of sets like these.

Connected

• ArtworkPlus
• Red Carpet
• Character Art for Movies
• Movie Director Thumbs