Calextone has some polyrhythm going on, but the disruptions are localised—they resync after a couple of bars, and the metres are displaced by a beat or a third of a beat, which makes for some very pleasant syncopation. Calextone also has some interrupted half bars, but blink and you'll miss 'em: there's only a couple. So Calextone is a little complex in notation, and sounds recognisable to us.
Things with Corps feminin are different. The top voice is passionate and melismatic—and again, mostly well-behaved metrically. The one beat interruptions from Calextone show up again in the cantus, and syncopate it forward as well. But there are also a lot more interrupting half bars, in all voices; so the disruption of the beat in different voices is greater, and lasts longer. Just three bars in, in fact, you get a double interruption in the countertenor, vs. a single interruption in the tenor. Let me illustrate:
The cantus is going along with five melismatic bars of 6/8. The tenor does half a bar of 6/8—then switches to three bars of 3/4, and then finishes its initial bar of 6/8. I've inserted "1…" and "…2" to indicate these half bars; Christina had a bracket notation, but that was clashing with the bars used as ligatures, and the point of this notation is to make the original more accessible, not less.
The countertenor, like the tenor, does half a bar of 6/8—"1…", then switches to 3/4. After just one beat of 3/4—"(1…", it gets bored of that, and switches back to 6/8 for three bars. What undermines the double switch to our ears is, his three bars of 6/8 𝄾 𝅘𝅥 𝅘𝅥 𝅘𝅥𝅮 sound like syncopated 3/4 to us now; but switching metres was a way of doing syncopation back then anyway. After that, the countertenor has to sync back up with the other two voices. That means first finishing off the other two beats of his 3/4 bar—"…2 …3)", and then finishing off his original 6/8 bar—"…2"—one measure after the tenor finished his.
I don't know if graphics will make this any clearer, but:
Why on earth is Solage doing all this? Because he can, of course. And also because the notation made it easy. If you wanted to change from 6/8 to 3/4, all you had to do was change ink colour, from black to red. If it was that easy, and if you were in an experimental phase already, then of course you'd switch ink after one note, just to see what would happen.
And what does it sound like? Like I said, the cantus is impassioned, and its syncopations work. The other voices sound—well, random. They're too displaced to sound syncopated, so they just sound like they're from somewhere else. Not unpleasant, at times effective—but untethered.
The Taruskin Challenge bloggers have a great post on why music historians view the Ars Subtilior innovations with distaste, though in general they approve of innovation. Their take: historians are unsettled by old art that comes too close to contemporary art, and doesn't stay in its historical box. Historians don't like stuff that resists fitting into historical boxes to begin with. There is a lot to this: we can't make sense of music that sounds mediaeval, but is also more adventuresome than Richard Strauss—it disrupts our notion of lineal progress. We're frustrated because we don't see the progress leading somewhere; but where we expect 14th century progress to lead to is madrigals (or, as the Taruskin Challenge say more knowledgeably, John Dunstable before we get to the madrigals); not Stockhausen or Babbitt.
To be fair to ourselves, though, we're also frustrated by the things Solage does, because we can't hear them make any difference. What performance of Corps feminin will bring out the fact that the Countertenor has 6/8 interrupted by 3/4 interrupted by 6/8, and make sure the catch-up half-bars sound like the completions they are? Given that it's just an accompanying voice, what performance *should* bring it out? The syncopation will come through alright; but our ears are ears molded through the path that led away from Solage: can we hear the bar resumption business at all? For that matter, could Solage's audience hear it, as opposed to seeing it? Or to use a fairer example: the first bar of the cantus is in 6/8, the second in 3/4. There's supposed to be a world of difference between Dum-dee-dee Dum-dee-dee and Dum-dee Dum-dee Dum-dee. Can you hear it in the recordings that have ended up in YouTube?
Perhaps, but it's work to bring out the subtleties, even more work to hear them; and these guys are obscure. As the Taruskin Challenge puts it, "a new breed of composers—none of whom are known today outside of the academy". (And these are musicologists to whom Machaut and Dunstable are old hat.)
It's work to transcribe these guys too. I got quite lost in the line above, and that was not as bad as it got; by the second page, I was obligated to break a couple of lines mid-bar. Lilypond was valiantly avoiding it, but it was making its staves a soup of dots; the alternative of course has made its staves a rather thinner broth. The MIDI this time is a prog rock combo: guitar, flute and bass. The default instruments in GarageBand are embarrassing enough that I've forked out the money for the Symphony Jam Pack. So the arrangements from now on will still sound embarrassing, but embarrassing in different ways.
The PDF is available for download at my site, and also at IMSLP.