Im an editor by trade. so the way it works
you shoot video, every camera shoots different codecs and this usually affects the ingest of the media and how its used before you edit.
now, when you start editing you import the clips to the project window and from there is where it gets fun.
This is non-destructive editing-
the NLE (non linear editor) sees the clips as references and when you "Cut" you are simply creating a route to the information you want to be used for the video.
most projects can use more than 50-100 clips for a smallish project with 1 camera and exponentially doubles as you add multi camera shoots. so 50 can become 100, 100 can be 200 etc.
when you render the video (export) what ever you want to call it, you are literally creating a new interpretation of 0's and 1's you created pathways too.
from here you have so many codec options (compress / decompress....co-dec) for youtube you more than likely export in the h.264 format as it is what youtube prefers and limits the processing on their end and can represent the video you cut in its truest form.
If their editors like most editors are they will export a full rez file and an online version.
kinda got lost where I was going with this and wrote it on my phone while on luncg, any more questions, feel free to ask.