Every store I know that sells MP3s gives you either 320k CBR, or a high quality VBR encoding. Neither is perceptively different (by humans) from a CD, even with the best equipment.
And most stores are now offering lossless downloads anyway.
If you're making your own MP3s, don't make 128k MP3s.
Later codecs (AAC, Vorbis, Opus, etc.) all allow you to use fewer bits for the same quality.
Time to delete them has come though, funnily it's much quicker to <arrgh me hearties> CDs I already own than rip them myself.
I think there is way more individual differences here than what some people assume. I have a few friends that often perceive way more details in images and image quality than me, to the point that some compressions that are fine for me will bother them. I think the same is true with sound: some people will benefit from CD quality, and some others won't.
That may also depend on the music that people listen to. If the tests are not made with the music you listen to, is the result the same? Is there a way to say "there's no difference for song X, so there is also no difference for song Y"?
There are certain trouble files that can't be encoded well with MP3 even at 320k. I once successfully ABXY tested one such file 20 times in a row.
That particular disc seems to have been made at the cross roads time during which seamless natural-sounding digital effects were possible in music, but were still blocky enough in implementation to cause some type of Moire effect with digital encoders.