Obtaining videos from social media for analysis

I use yt-dlp. I use it on Mac. It's python and should work on Windows the same, but I have a strong preference for doing Windows CLI things in Windows Subsystem for Linux. It receives regular updates to custom handlers for popular sites (e.g. x, youtube, reddit) and a generic handler that often works. Apparently there's a simple GUI frontend called yt-dlg, but I've never used it.

For what it doesn't work for, I use 'Video Download Helper' https://v10.downloadhelper.net/ which is good at grabbing pretty much anything from within an open browser tab. It got in some controversy a couple years ago for restricting the free version to push people to use the paid version, but I don't mind a little one-time payment for lifetime license.

Youtube has had some recent(?) updates that makes some videos require additional steps in order to pass in valid request cookies/headers but yt-dlp has implemented support for that. It can read your google cookies from a browser like Chrome that you have installed. It may be a little concerning to have a tool read your authenticated browser cookies but it's open source and I try not to worry too much.
 
I have zero experience with these tools. Do the downloads preserve any of the original video's meta-data?
 
I have zero experience with these tools. Do the downloads preserve any of the original video's meta-data?
No the metadata is stripped on upload on the vast majority of (probably all) social media sites. So its not even there to download in the 1st place, it's possible some sites might read it, store it elsewhere then strip for presentation, but the law around this might be strict in some places.
 
No the metadata is stripped on upload on the vast majority of (probably all) social media sites. So its not even there to download in the 1st place, it's possible some sites might read it, store it elsewhere then strip for presentation, but the law around this might be strict in some places.
Unfortunate as that would be one of the fastest ways to deconflict multiple reports as being/not being all of the same incident.
 
Unfortunate as that would be one of the fastest ways to deconflict multiple reports as being/not being all of the same incident.
Phone videos by default contain data like GPS co-ordinates and sometimes other metadata that could easily id someone so it does need to to be stripped on social media.
 
OTOH that same data would actually help validate your claim as genuine and not part of a hoax. Believers want to have it both ways.
 
Both Google Photos and iCloud Photos let you generate shareable links to original files backed up to cloud storage, which can provide the original metadata. As do Google Drive and iCloud Drive, for photo/video files not taken with an Android or iPhone device, but which can still be backed up to those generic file storage platforms.

Of course anyone sufficiently motivated to fake the metadata could still do so. But this at least is a way to make that metadata available in a more reliable and specific way than simply saying "I took this in X town around 7PM", which if correct is sometimes enough, but for things like flight paths or identifying specific non-ISS satellites, is often too vague.

EDIT: I do know that Google Drive will throttle downloads if too many people download the same file in a short period of time, which can sometimes happen when someone posts a Google Drive link in a popular Reddit thread.
 
Back
Top