Two observations from me:
1) On the results... The degradation at 5 passes was about what I, as a 6 year 828mkII owner, would have expected. But then it wasn't like I had a super low opinion of the unit. I bought it (like others) on the features and price point (FW multichannels were still relatively few, particularly under $1000) and the generally decent rep of the company's products. I was coming from a stereo Echo Mia card and, frankly, my impression was not that I had stepped up in sound quality. Actually, I was afraid that I liked the Mia better. (The Mia was then stranded in an old-underpowered desktop when I moved to a then-relatively speedy laptop as primary in 2004 [laptop is still doing fine but it doesn't seem so speedy... that said, it's long overdue for a clean system reinstall]. I keep meaning to do a shootout. Maybe sometime before another 6 years is over... but the MOTU is pretty engrained in my rig. I even use the CueMix dsp mixer for other, non-critical signal routing like everyday audio, video sound and such.)
(Now, there is little question, as with so much else in the trendy, faddish world of record making, that familiarity breeds contempt. Sometimes that is just faddishness but, of course, there's another side: as production runs go on, profits are milked and, sometimes, corners are cut or fundamental flaws that were initially unnoticed and later, never fixed, can quite legitimately undercut a product's rep. But mere ubiquity can be a real factor, as well.)
2) On the test material... To really get a get a handle on high frequency handling, I tend to like to focus on the decay of material with delicate HF content -- room ambience, reverb of various types and sources, and cymbal decay. A good, well-recorded cymbal, in fact, is a great way of focusing on HF handling since with poor HF handling, you can still have what at first sounds like relatively similar HF levels but, on inspection, the lesser devices or processes often remove detail and character and one gets a glassy/glossy sound that may well be pleasing, but doesn't actually reflect the character of the initial sound. I'm with those above that feel like the initial sounds used for this test might have been more valuable had they had more of the subtle but important HF material (and that often comes from excellent recordings of real instruments). When I heard the original here, I was already 'disappointed' in the sense that I knew the choice of sounds would make it difficult for me to make the sort of evaluation one would like in a case like this.