I don’t believe you about the GPS harming the battery life that much. Unless you’re talking about the logging.
It was probably a coincidence.

A solid-on satellite icon does not mean it will record the geolocation into photos taken. It should, I agree, but I’ve seen [rare] cases where it does not. Possibly just an outright bug.

The battery life is clearly dominated by GPS recording. When it’s off I can get at least a whole day of use without any concerns. When it’s on, only half a day.

Thanks for the tip on GPS logging. I still haven’t really played with it. I vaguely recall there was some reason I’m not using it, but I can’t recall right now.

Re: Battery life… man I am getting a lot more battery life than you..I got about 20 hrs and 1,600 shots, it was split into four 4-6hr sessions
BUT.. I turn the GPS off when I am no longer using the camera… and yes, I am looking for a “quick” way (i menu or button… but nothing yet) to do that rather than going through the Setup -> Location data menu.

FYI: If you GPS enable logging, it will not save it until you go into “Location Data” and turn off the “Create Log”… I turned off the “Record location data” twice and it totally looses any logs… does just fine if you toggle the Create… BUT the logs are NOT compatible .gpx logs (ARG)

I figured it might be, but it turns out it’s not really. When I used the low-resolution versions from this post, the Z9 basically didn’t recognise the head or eyes at all, ever. At any focal length on the Sigma 150-600 C, at any distance I could fit within my office. The Z9 did better in real life (not great, as I noted in the body of this post, but it at least sometimes caught the eye, or at least the nostril).

But, when I showed the same images via Lightroom – i.e. the full resolution versions – the Z9 was able to start finding the eye. And interestingly it never thought the nostril was an eye. But overall the subject-detection performance was still impractically poor – the AF box flittered rapidly and unpredictably around the screen (in 3D Tracking or Auto-area) including on the beak, the overblown sky, randomly next to but not on the eye, etc. But most frequently on the neck feathers (mimicking the affinity for fluffy necks that I saw in reality).

I did find that significantly (1-2 stops) under-exposing the image made the eye detection work marginally better, against the iMac screen. Matrix metering by default over-exposed a bit (quite explicably because of the black pillar-boxing & iMac screen borders).

The behaviour was essentially the same with the 80-400G.

What does that tell us? Not much, of course. There’s so many ways in which these two situations varied, that could impact AF behaviour. I just thought it was an interesting thing to try.

Mike Dale apparently got quite different results, but with a different lens and likely other factors different as well (distance to iMac, screen brightness, lens, etc). Maybe it suggests that a native Z-mount lens like the 100-400 would work better. I wouldn’t be surprised – I’ve definitely seen that the native Z-mount lenses have much better autofocus accuracy and consistency than even Nikon F-mount lenses, let-alone 3rd party F-mount lenses. Unfortunately Nikon don’t make any lens – Z- or F-mount – equivalent to the 150-600 (though a 200-600 is on the Z roadmap, which I have my fingers crossed for).

In any case, the behaviour I’ve described in this post isn’t specific to that one situation or bird. Generally I find the Z9 does a decent job in finding bird eyes, just not always. The Condor was just an example of what happens when subject detection doesn’t work well.