(Disclaimer: I haven't yet been able to do more than quickly handle an A9 Mark III, and that's probably all I'll be able to do until it becomes generally available later this year. However, for the purposes of this discussion, I don't need to have an extended session with the camera.)
Now that dpreview has given credence to all the measurebators' fears (lower dynamic range) and provoked even more misinformed discussion, it's time to adjust people's views.
Simply put, the Sony A9 Mark III buyer would not typically be all that worried about maximum dynamic range. Never was, never will be.
The primary technology change in the A9 Mark III is the global shutter. Global shutters have basically three benefits at the expense of two drawbacks, all else equal:
#1 Benefit: No rolling shutter, LED banding, or display line mismatch effects.
#2 Benefit: Potentially higher bandwidth (higher frame rate).
#3 Benefit: Flash sync at any shutter speed.
#1 Drawback: The extra electronics involve decrease well capacity, which triggers a gain adjustment that increases base ISO.
#2 Drawback: The same changes basically lower dynamic range compared to a non-global shutter version, due to noise.
That's really it. For general purpose photography, the two drawbacks would make it unlikely that you'd buy an A9 Mark III. For particular use cases, particularly in sports and wildlife photography, the benefits might make it so that you do want an A9 Mark III.
Here's the thing: if I were a full-time working pro sports photographer running from venue to venue, that #1 benefit would easily outweigh the two drawbacks. That I'd also get that #2 benefit might come into play every once in awhile. The #1 drawback isn't really a drawback, as I'm rarely at base ISO for this type of work. The #2 drawback might come into play in a few situations (ironically, including when LED displays are backlighting the action on the field, so instead of a scrambled display I'd get a blown out display).
Putting on my hat as a full time wildlife photographer, benefit #2 starts to become more interesting, though with small birds in flight I could see some benefit to #1 when it comes to wing tips. Unfortunately, the #2 drawback starts to come into play, particularly as a lot of my work these days happens in very low light (and no direct light) but high contrast conditions.
I'm still trying to figure out just how much of a benefit #3 actually would be to anyone. Granted, flash sync at 1/200, as it is on the Nikon Z9, is too slow. I encounter plenty of situations where I need that to be higher. However, even being to sync at 1/500 would be enough, and I'm sure that's coming soon in a non-global shutter camera. (Actually, coming again. We had 1/500 flash sync during the early DSLR era with CCD image sensors.)
Dpreview's test results aren't surprising. As I wrote when the A9 Mark III was announced, this was definitely a technology play by Sony, and one that seemed timed and presented in a way that suggested more than it was. The A9 Mark III is going to have customers, for sure, but it's a fairly small subset of the market. Nikon's Z9 followed by the Z8 has taken quite a bit of wind out of Sony's sails (and sales), but I'm not sure the A9 Mark III really gets them to the mark in first place.
I wrote that we need to adjust some folk's views. That's very true with all that I see from the comments that are proliferating.
For example, that dynamic range thing. Basically, we measure that from some "noise floor" to saturation. I've already noted that the extra electronics at the pixel level reduce the maximum saturation, so the noise floor starts to become very important. Almost everyone (including dpreview) that I see commenting about A9 Mark III noise tendencies seems to be doing it from "what I see." Have you tried measuring their samples? In particular, use a largish area on one of the ColorChecker patches and observe the standard deviation. Curiously, the Sony A6600 has a slightly lower standard deviation in every same ISO sample I took. But then again, that 24mp APS-C sensor in the A6600 was essentially Exmor state-of-the-art, so I don't find that surprising.
The question is whether the A9 Mark III has enough dynamic range for its intended use(s). My answer in looking at others' results so far is, "yes if the image is properly exposed." All sports venues are lit, for instance. Sure, some of that lighting is low in value and has a fair amount of fall-off from the central point, but it's direct light. If you're accounting for the difference in images taken, say, in the far corners of a poorly lit soccer field versus the center field with your exposure, you should be fine with the A9 Mark III. About as fine as you'd be with the A9 Mark II, which I don't hear any of my pro friends complaining about. Moreover, there's the issue of frequency with some lighting, and the global shutter is going to deal with that, too.
The fact that a lot of pro sports output is JPEG means I'd be looking at how Sony tweaked BIONZ for the A9 Mark III, too. Judging from dpreview's examples, that seems to be fairly strong JPEG noise reduction with only modest edge softness at ISO 12800. Indeed, with those settings, the standard deviation is now lower than the A6600, and the A9 Mark III seems to be retaining edges a little better. I also look at the A9 Mark II results against those from the Mark III, and see similar things, so I'm just not seeing that the intended audience for this camera is likely going to be disappointed in it.
