at large you should always prefer testing files you converted yourself. as you have foobar, the ABX component makes that real practical and honest(so long as you yourself wish to be^_^).
I've seen so many fake tests over the years with the signal doctored ways a file conversion never would
. the very Tidal mentioned in the link had such a BS online test when they started. I don't remember the details(they're in a topic on Headfi for sure) but I think they had simply EQed the files "I'm not a crook!"...
just a few days ago another such test was linked by a member about cables this time as "evidence". the tracks weren't aligned, the levels weren't matched and the track was shorter with the same music, like one track was playing faster, starting later ending about the same time...
so instead of always wondering what dishonest crap someone is trying to pull on you, convert your own tracks from the same single original(so you don't have to wonder about different mastering).
on this test the files seem fine at first glance(I only downloaded 2*3 songs, and the operations and variations seem similar). the wave file(no flac is involved here
wave, 320mp3 and 128mp3) starts just a tiny little bit sooner, but I wouldn't worry about such a small delay, the browser handling different formats is just as likely to create more than that as the file loads when we press play. volume level is close enough to be considered just a side effect of the conversion(well below 0.1dB). so aside from how silly it is to have a test letting people guess once(statistical significance is necessary to rule out lucky guess), I'd say they offered that test in all honesty. which is nice for a change.
I'd still suggest to do your own test instead if you're really interested in the sound from your files and your audio player.