First of all, is the technique called "resampling" ?
Does youtube make live resampling for every user ? Isn't it too cpu costly ? So far what i know about codecs is that it has to go through many encoding steps ( compression, motion prediction etc etc ). All this things happens live ?
Or it encode and save different video files for different bitrate, so that when a user changes bitrate it servers that file.
Isn't that a disk space destroyer too ?