found this very interesting on the subject of the lowly MP3 {snip}
Yes, that is a well known post on YouTube. You should go to the original article where they have the files for you to listen to, just as the woman in the video does.
https://www.npr.org/sections/therecord/2015/06/02/411473508/how-well-can-you-hear-audio-quality
Not that it matters, but note the change in title (how well can you hear ... to audiophool?). It does indicate an introduction of expectation bias, which should not exist in a true blind test. Just sayin'
A better test, with music you are familiar with and actually like, would be to select a song from your stored music, create versions with different lossy compression as well as the lossless original ... if you have a HiRez file that would be even better, you could also create a downsampled 16/44.1 version and throw both into the mix. You will have to name them uniquely, one that is descriptive is OK since you aren't going to see them when you take the self-test.
If you're worried about volume matching, just normalize every track, for pop music try -14dB RMS. You might have to fire up your audio editor for that, Audacity for some of you, others will have their own favourites (I would use Amadeus). It won't matter if normalizing causes a resample of your original. You don't need bit-perfect copies of the original, for the purpose of the test they will be close enough.
If you're going to use Classical Music, you should normalize to -20dB RMS. Not everyone likes the stuff, but just as an FYI there are characteristics of Classical that do bring out differences in lossy compression more than Pop music.
If you want to use more than one song, it's better to create multiple playlists, one for each track. It will make it simpler when it comes time to listen to different versions of the same track.
Then create a playlist with just the versions you've selected, have your player software choose them at random, and play them through your main system, take notes, and see how you do. Don't be afraid to turn it up to a nice listening level, and try to choose a time and setting when you are relaxed. Performance Anxiety (some people have problems with tests, period, regardless of how well they know the material) will give poorer results, so really try to be comfortable if you can.
Maybe set your playlist with a few more tracks than actual formats you've created, so that a few appear more than once.
Don't look at the playlist until you are done testing yourself. Maybe get someone to actually operate your computer so that you can't peek, although with care you can pull it off (change the window size so you can't read the playlist, for example, or tape some paper on the screen, randomize a few times without peeking, and so on).
If I've left out a few details as to how to pull this off, well, everybody's bright around here, you'll figure it out.
