Thank you. Not exactly an easy thing to find or get in some parts of the world. Ordered a cable from StarTech that has the VMM1700 chip in its specs.
Edit: Bad cable. Has the right chip but not the capability; eats DDC.

Meanwhile, I visited the colleague and almost immediatly noticed one of his two identical monitors shows banding, while the other does not. That, we found in system_profiler, is due to one screen getting a full HBR3 (8.3 Gbit/s) link, with all color modes available including 10-bit RGB, while the other only gets HBR (2.7 Gbit/s), for which BetterDisplay only lists 8-bit YCbCr 4:4:4 available.

Do you have a trick for that, too? I was thinking of reducing the 144 Hz to 120 Hz and therefore free bandwidth so both screens can be used at 8-bit RGB. Obviously, selecting this mode after the DP link is already established does not free up bandwidth. I tried crafting an EDID with YCbCr support and the 144 Hz refresh rate removed, but applied with BetterDisplay, it changed nothing. Is that something an EDID emulator dongle can help with?

Just bought a BD License, whether I'll need BetterDisplay in the end or not.
You're doing outstanding work here, and you/your tools helped me understand a lot of things.

Thank you. 😊

Regarding the connection issues, I don't think limiting one display will free up bandwidth for an other display if the displays are directly connected - there might be some different reasons why this works like this. Using both screens with 8-bit RGB should not help with banding, quite the opposite - banding is not because of YCbCr vs RGB, but 8-bit vs 10-bit (so if both screens are 8-bit, both will have banding). But normally the framebuffer depth is 10-bit for a non-HDR desktop in macOS and macOS uses temporal dithering (if not disabled in BetterDisplay) which should help with banding and make the result similar to a 10-bit connection and a 8-bit+FRC panel.

The colleage had complained about eye fatigue since upgrading Monterey to Sonoma. I've recently read a thread on hackernews where people complained about the same thing and blamed it on dithering, so he disabled dithering with a single-purpose tool for that. Can't recall its name.

What I clearly saw, though, was 8-bit without dithering looking fine in RGB, but having horrible irregular banding in YCbCr. I can only explain this if MacOS lies to BD about the actual color depth, or if his screens do something like FRC upsampling, but only in RGB 🤷‍♂️.

Yes. BD also has a toggle to turn dithering on/off on a per display level - my experience is that this tends to actually affect YCbCr connections only (?) in a substantial way. At least on my monitor. :) But I am not entirely sure and I understand that turning dithering on/off helps a lot of people in a psychological level so I don't want to take that assurance away (therefore it might not be wise to share this insight with the colleague if he feels better with dithering off), so it is entirely possible that I am wrong about this. The GPU dithering should be mostly there to gap the bridge between a 10-bit framebuffer and a 8-bit connection and it should have an obvious visible effect with the appropriate content. MacOS can also use other kinds of dithers (non-temporal) - which should not cause any eye-strain, so maybe that's the explanation...