Login at Kodi Home

Crissov · 2021-06-01, 16:46

Sometimes, guessing the movie or TV show from the file name (and path) works quite well, especially if it conforms to the conventions established by release groups for pirated content. It still often fails for me, though.

Like most people do with ID3 tags for MP3 audio files (and similar in M4A/MP4 etc.), I also keep my M4V/MP4 video files properly tagged with metadata.
Fortunately, Kodi v18 and later support reading these embedded tags. Unfortunately, this is now disabled by default, because apparently many preexisting files have bad data stored in their tags and their users did not notice before.

I realize that centralized metadata at sites like IMDB.com, TheMovieDB.com and TVDB.com will _eventually_ be more complete and more accurate than what I put into the file metadata originally. However, it often takes some time after the release to be added there, and sometimes (especially for non-US content) only one or two of these sources already have any info at all upon release.

I would like to be able to combine the best of both worlds, offline and online. I think this would be best achieved with an ID of some sort embedded into the video file, which could then be used by online scrapers to unambiguously identify the correct resource. URLs are one possible type of ID. That is, much of how Parsing NFO files are used. Therein, the IMDB URL is preferred, because other sites often also store it for reference.

For Matroska video files, Kodi basically first probes for an embedded classic NFO file, then looks for a tag named `IMDBURL` first and `IMDBURL` second to use these for scraping, before it tries `TITLE` in which case it uses this data.

For MP4 video files, Kodi assumes that the embedded metadata should be preferred (and is complete) if either the description or synopsis field is present (FFmpeg terms, in MP4: `desc` and `ldes`). There ios no equivalent for IMDB and TMDB URLs.

I wish I could provide the IMDB title ID or URL within MP4 tags so Kodi can use it to scrape further data from web sources.

mYnDstrEAm · 2021-06-01, 18:10

That's good suggestion imo. I think it would be best if at some point basically all media on the net has a unique identifier which is contained in or can be added to its metadata.
I suggested this "identifier-mechanism" here for the file-explorer "Dolphin" as it's especially useful for files whose filenames change – for example for reading tags also in the metadata. The best way to store the hash and retrieve information is Wikidata, or maybe a fork for it dedicated to media in specific.

Kodi could also read and write tags to files such as the flags for movies that have been watched.
It's also useful for fast and simple exports and imports of libraries and ambiguous filenames.

I don't think it's an urgent issue though but/and that a main issue is getting more developers to develop Kodi further. That would be helped by making Kodi a lot more popular and competitive to other popular media centers. Hence I propose prioritizing issues which make Kodi a lot more useful, versatile, popular and easy to use (i.e. making it simple to use, improving gaming-center-support, properly enabling documentaries), rather than working on improvements only small numbers of people would benefit (much) from. Nevertheless, it would be pioneering implementation of what would help enable a large change to the Internet, so even if it's not that useful at first, Kodi could play an important early role in facilitating that 'upgrade'.