vorbis VS vgmstream

This a hard to answer question as it depends on the compression algorithm and even then its not a simple answer. Usually you see the stuff above 20 khz removed as well as the hard to hear detail in the high range. They are based on psychoacoustics so they try to only remove unheard detail. Another example is a quiet tone close to a loud tone in the frequency spectrum will be masked and harder to hear. The real answer is the code. Here is the reference code for ogg which Spotify uses. https://github.com/xiph/vorbis

Using vgmstream plugin with combination of media players that supports it. But for this, you will need to create a supplementary TXTH files since these VB files won't be recognized properly by the plugin, and also you will need to copy the VB files off the disc to somewhere to create this file because this file need to be put on the same directory as where you'd copied the VB files. You can read more about TXTH here