It's rock solid - there's only one print head, and cutter is separated - so the carriage is much lighter and doesn't shake the machine while printig. It's very silent indeed (sorry for the FB):
https://www.facebook.com/MarcinCzornyjKaluza/videos/1109039909166037/
Screening is invisible, and looks smooth even under loupe inspection. The transitions look perfect, tonality is at least as good as in case of K4 PIXMA Pro-1.
It reaches L*15 on best matte media, and the matte prints are very 3D looking and have lot of depth impression. Ink layer is smoother in touch than in case of x400, so the matte prints are less fragile.
The dmax on baryta media is as good as it gets (L*1), black is black hole like, so there's no need to improve anything. The density of black is rather a matter of pigment layer structure uniformity than ink amount, so I guess that the ink amount is just sufficient and there won't be any gain of dmax if we put more ink on paper. If you print black again and again in the same place it won't get darker, it can even get worse 
I had tried to apply CO to matte prints, but it didn't make any noticable difference.
Thanks Marcin! I don't mind the FB link. Wow. What an amazing printer from Canon! The carriage looks much smaller than before indeed. Did you notice if it decelerates before stopping and moving the opposite direction? Or is the more silent operation just because its mass is less? I feel it's more of the latter, judging from just your video.
Sounds amazing about the dot pattern and visual tonality. Would you say it's even smoother than Epson's screening now? L*15 is not bad, although L*14 is a quite visible difference and properly solid. The P7/9000 Epsons let you lay down more ink with the color density slider, and the dmax goes from L*16 down to L*14 with heavier ink loads, at least on some matte media. Maybe the Canon could do better, maybe not? It sounds like you already optimised the inking in your tests for highest dmax.
When I mentioned CO on matte prints, I am thinking of the CO sealing the hydrophillic nature of the ink receptive coating layer of these aqueous inkjet papers, so they will be less susceptible to harmful fumes in the air. Some folks spray their prints with Print Shield for that protective effect. Mark McCormick-Goodhart does that. I'm adverse to spraying anything like Print Shield (risk of encephalopathy among other downsides), so letting the printer do the coating would be nice. You say it doesn't change the visual appearance, so that is actually a good thing! May I know how did you fool the printer into applying a CO pass? Did you double print and cheated the driver by saying its a glossy media?
Oh, did you make a close up picture of the pinch rollers that feed the paper? I'm curious as to how Canon re-designed the feed mechanism. How is it different from the iPF83/400 printers? My and other 8400 printers mark softer papers like Canson Platine and Rag Photographique. I wonder if the new rollers are any kinder to less structurally rigid media, or if we could even adjust the amount of clamping force applied.
The old design uses a spring in the center of a lever style arm, with the rollers on the other end. The Epson P9000 uses a lever spring that sits between two of three rollers, on the roller axle, which deforms slightly and results in uneven clamping force across the rollers. In addition, the clamping force is a bit high. A friend who has the P9000 had a look at the P20000 and informed me it is re-engineered - it uses two springs of about half the strength, one in each of the two roller gaps and the plastic chassis/bracket/assembly has been approximately doubled in rigidity. The peak roller force seems to be about half what it was, but the minimum force (at the outer ends) is still much lower than the force near the spring ends. Better but not ideal. I'm very much of the opinion that the Canon design is overall better, as it distributes the clamping force more evenly, it just is a bit too much clamping force for delicate papers.
Canon iPF8400 part:
