Update at bottom.
There's a great deal of energy on the Interwebs right now as to whether or not the Z9 firmware introduced a bug or not into the notion of hybrid button autofocus.
Let's start with a definition of that term.
Hybrid Button Focus (for the Z9) first got suggested via some YouTube videos early in the camera's introduction process. Basically, these videos advocated not implementing Back Button Focus (Custom Setting #A6 not set to AF-ON only), and assigning AF-ON+AF-area mode/3D-tracking to the AF-ON button. You'd start your focus via a half-press of the shutter release, have the camera find a subject automatically, then switch to the AF-ON button to track that specific subject once it was identified, even if it went out of the Large-area AF box.
The concept here is called "focus method handoff." This method worked with the original firmware, but only if Custom Setting #A7 was set to Auto (Focus point persistence).
My problem with that particular Hybrid Button Focus method was simple: it kills one of the key benefits of Back Button Focus, which is that you can let go of buttons and not have the camera focus when you finally press the shutter release. I believe it was actually a bug in the original firmware that allowed you to still have the semblance of Back Button Focus.
In my Z9 book I advocate a different Hybrid Button Focus method (set #A6 to AF-ON only, assign AF-ON+AF-area mode/3D-tracking to one of the L-Fn buttons, or perhaps an Fn# button, though all the lenses I use in situations where Hybrid Button Focus is useful have L-Fn buttons right where my left thumb is positioned). Again, CSM#7 has to be set to Auto. The reason for my suggestion is that it completely preserves the notion of Back Button Focus: I can release buttons and focus won't move again. Indeed, that's how I accurately tweak focus manually at times. It's the same method I recommended with DSLRs (D500, D850, D5, D6).
The C3.00 firmware update has created a situation with the YouTube Hybrid Button Focus method that seems to be causing some people issues. In particular, if you're in an AF-area mode that uses Subject detection and that is On, when a subject isn't detected, then pressing your alternate focus button with 3D-tracking on it produces random results. I'm not 100% convinced that this is completely true, as I can't reliably reproduce this; there may be yet another setting that is coming into play, or even the timing by which people are transitioning between buttons. (3D-tracking should pick the center of the focus area box if there's no subject detected, and this is what I saw when I tested the new firmware.)
No doubt something did seem to change between the earlier firmware and C3.00 if you use the YouTube-recommended Hybrid Button Focus method (but not the method I advocate). I suspect that Nikon simply removed or changed something they hadn't fully considered in the first place. Many of those using the YouTube Hybrid Button Focus method that I've observed are pressing two buttons simultaneously (shutter release half-press and AF-ON button), while others are backing off the shutter release until they press it to take the photo, which brings into play what does the camera think you want it to do (e.g. change focus to the shutter release again?).
I'm forwarding this information to Nikon and asking them to keep me in the loop on it, because I'd like to know what their real thinking is. Obviously, they were trying to bring the DSLR AF-ON+AF-area mode construct over to the Z9 as I and others have been constantly asking them to, but the Z autofocus system works differently than the DSLR one in terms of a lot of things, including the internal timing of the focus system. The Z's operate off the image sensor stream, while the D's operate off a separate stream with a unique (and very fast on the D5/D6) dedicated sensor.
Update: I've now come up with a very specific group of camera settings and three observed (and different) results that can happen with them in regards to this, and have forwarded that along with commentary to Nikon.
This does bring up a very good question, though. I'm not 100% sure what is supposed to happen in my setting scenario. The specific issue has to do with 3D-tracking taking over focus on a button press. Exactly what should 3D-tracking track? I'd say that with no subject detection it should probably be the center of the currently active focus box. With subject detection active it should be the detected subject, but that can be bigger than the 3D-tracking area, so again, it should be the center of the subject detected box. However, if that's what the camera should do, this still can result in the 3D-tracking box moving from where you think it should be. There's more going on here than first meets the eye, particularly because of subject detection, which can have four states (none detected, subject detected, head/cockpit/front detected, eye detected).
It's starting to become clear to me that many of us had been taking advantage of a bug that created a feature (;~). Something in the firmware update changed the characteristics of the bug, which changes the feature. It will be interesting to see what Nikon does with this, and what their eventual "solution" is.
