What's the official position on scraping metadata from MP4 files using the data embedded into the files itself (I think in the moov atom) and not in separate .nfo files?
My MP4 movie files show their embedded cover art within XBMC but not stuff like Title, Date, Description, Show etc.
I'm just wondering seeing as I didn't expect my cover art to show up in XBMC with MP4's.
You can do it, but XBMC won't read it as yet. Patches welcome. I've said a bunch of times how to do it, but noone cares enough to get it done.
I don't have any experience in writing code whatsoever, so even if I wanted to I couldn't implement it.
I don't want to sound imposing but would opening a feature request ticket on Trac help?
I've also noticed in Gotham (AFAIK it's new) that cover art is extracted from MP4 instead of using a frame capture. Any hint on where in the code that is being done (I guess I'm being lazy, but I find searching the code on github a challenge)?
scott s.
.
Its gets extracted I bet from online since you already been told that any embedded anything is not supported. Unless the online one doesnt have a frame uploaded to match then it reads the actual file for a thumb not anything embedded.
(2014-04-21, 23:13)uNiversal Wrote: [ -> ]Its gets extracted I bet from online since you already been told that any embedded anything is not supported. Unless the online one doesnt have a frame uploaded to match then it reads the actual file for a thumb not anything embedded.
Nope, XBMC is indeed extracting it from the file itself. No one seems to know why, either. Maybe something from ffmpeg caused it to be exposed to XBMC, who knows.
I have music videos with local nfo for every file. Never scrape online. Very easy to see, just add or remove the cover art from the mp4 file and reload.
scott s.
.
Is Spiff Ironic_Monkey?
Also, how difficult would it be to add his/her support to mainline builds?
If you build yourself it's as easy as cherry-picking the commit and building.
If you don't, you're reliant on someone taking the time to do the above, do any cleanup required and do a pull request against mainline, where it will be reviewed for inclusion.
As suggested in the post above, what is there isn't perfect, but it's a start.
