I'm screening a large media library (20TB) wherein some files got corrupted when I did a transfer via filezilla (by my guess ~10%). The corrupted files display with a green "filter" over every frame (when played via plex and a number of local video players playing the file directly).

I'd like to screen the library, and want to write a script to get an average color reading.

Are there any libraries that would let me return a value AND specify how many frames I want it to take the average of? Because of how consistent and defined the issue is, it's really not necessary to average the whole file.

It would also be great if it automatically skipped non-video files, but I imagine a simple "try/except" would be fine.

My skill level here is best described as "high level hobbyist". I'm familiar with what I need to do iterating over the folder etc, but would prefer not to learn how to pull specific frames from a video container unless I have to.

Thanks for any help!

you are viewing a single comment's thread
view the rest of the comments

[–] batmaniam@lemmy.world 3 points 3 months ago

Thank you for the tips! It turns out the files weren't corrupted at all, it was a transcoding issue somewhere in the hardware acceleration stack. I've disabled it which will cause some bottlenecks, but there's more than enough processing power in the rig for the limited amount of people that use it simultaneously.

Thank you for the reply and the links though! My original plan was NOT to fix the files, just to figure out which had an issue and re-download them. Like I said it turns out that was unnecessary.

Thanks all the same for pointing me to that thread. I do want to start screwing around with some imagine manipulation at some point, and this has already been really educational in learning the top level flow of this kind of thing. Again, I'm more used to passing data for sensors and motors.