I feel like the issue here is that, basically, your earlier posts translated in my mind as saying that the a9 will give better results for the OP than the a7rIV. But, that doesn't seem to be supported.
You quoted Lloyd Chambers, but I think you left out a key part of what he said. You said:
But, what Lloyd actually said is (emphasis added):
He's talking about the same PER-PIXEL DoF and you're just talking about absolute or overall (apparently) DoF.
With the a7rIV having more pixels (in the same image/composition), won't the overall DoF work out to be the same between the a7rIV and the a9? Isn't overall DoF really only dependent on focal length, aperture, and overall sensor size?
To circle it back to the OP's (and my) question: You're in your "spot". Your subject is where it is. You have the same lens. Which camera will give you a better result? It seems to me that (assuming correct focus and exposure), the a7rIV would be the clear winner.
If the final composition you want is cropped so much that an a7rIV source produces a 5MP file, that tells me that the a9 version of that same original is not even going to be usable. I.e. when cropped to the same composition, it will be way less than the 5MP you need.
In other words, if the subject at the limit of the reach of an a7rIV, it's out reach of the a9, so the a7rIV is the obvious winner.
If the subject is in reach of the a9, and we suppose that the final cropped image produces a 5MP image from the original a9 source, then that same photo shot with an a7rIV is going to have considerably more (2.5X?) pixels worth of detail. And, while the per-pixel DoF might be less, the overall DoF that the eye actually sees will be the same, won't it?
In this scenario, you MIGHT argue that the extra detail in the a7rIV image brings no benefit - it's a wash between the 2 cameras (for image quality). You got the 5MP image you needed, either way - one, achieved only by cropping, the other by cropping and then reducing the image size (down-scaling the resolution). But, I am still unclear on how you would support the notion that the a9 result would be actually better?
Stuart -- Let me address these items 1 by 1. They are good points.
Note: Below when talking about DoF, I use "sensor size" to mean the MP assuming a fixed physical size. In your post you said sensor size in terms of physical size. I have tried to fix my wording below.
0) High MP overall.
I did mention in my post a couple times that I went to the d850 so I would have a lot more MP and said that I would likely choose the A7RIV over the A9 because I tend to need to crop a lot. The caveat is that one needs to know that more MP has some downsides, such as a smaller effective f-stop range.
In my opinion, the main issues with the A7RIV are (a) small 68 exposure buffer @ 10 FPS jpeg, (b) need to stay around f/5.6 maybe f/8 max to optimize DoF vs diffraction, (c) shake and vibration will need better IS or faster shutter speeds (assuming you want the same per-pixel sharpness as 24MP sensor). (b) and (c) are only significant issues if you are concerned about big enlargements or heavy crops or high micro contrast. If you have pixels to spare, you can likely sharpen your way around them.
If you can work with those limitations, the A7RIV is a great camera and will give stunning results. And you can crop it 1.6x more for the same MP output & IQ output assuming you control for the (b) and (c).
For me (a) is not too much of a problem as I rarely shoot long bursts. What I have run into with the A7RIV when I borrowed one is that with an older sd card shooting BIF, it would take a long time to write out the buffer when I was shooting raw. So, even though I was shooting maybe 10-shot bursts, I was shooting them frequently enough that I still backlogged the buffer. I assume that a modern super fast sd card would handle this a lot better. One just needs to budget for a few of those cards with the camera.
To reiterate, I would likely choose the A7RIV over the A9 for my shooting style. The reasons to choose the A9 are faster fps, deeper buffer, better high ISO performance, better performance at higher f-stops (assuming you are pushing the pixel resolution), maybe better AF.
1) Composition vs. maximum crop at a given MP.
The OP specifically asked about heavily cropping photos, so I was trying to answer his question in terms of what is the maximum crop you could get, assuming the same MP output at the end.
Your aspect of having a specific composition and comparing the two cameras is different. In that case, the A7RIV will always have 2.5x more MP and assuming (b) and (c) are not issues, will almost always give better results. My reading of the OP's question, however, is that he wanted to use those extra pixels for more cropping.
I agree with your example of the A7RIV with a 5MP composition being far superior to the A9 with the same composition. In terms of my post, the A7RIV is in the 1736mm - 2766mm effective focal length range (at 5MP), whereas the A9 has topped out at 1736mm. This is the point I was trying to make with the max equivalent focal lengths.
Really the view of which is better in a given composition vs which has a longer effective focal length are two sides of the same coin. If you are within the effective focal lengths of each camera, the A7RIV will have 2.5x more MP for the same composition. If you are outside the A9's effective focal length, the A7RIV will become increasingly and increasingly better than the A9.
I like to think in terms of effective focal length, as I can then immediately have an idea of the distance to subject size that I can achieve with a given rig. So that is why I focus on the 1.6x more focal length than the 2.5x more pixels. But they are really expressing the same thing: 1.6 = sqrt(2.5).
2) DoF. Depth of field is caused by how much a point of light is blurred over a circle by the lens. The size of the circle is independent of the sensor. So, the A9 and A7RIV, for the same lens and f-stop will have the same circle of confusion. Because the A7RIV has smaller sensor sites, the circle covers more pixels that it does on the A9. The circles themselves are the same size on the sensor. So, if you have a big enlargement or do a lot of cropping, the A7RIV will show worse DoF (more blurring) when you enlarge it, on a pixel-by-pixel basis. Of course, with the A7RIV you get to enlarge more or crop more to get to that point.
If you are not heavily cropping, then one likely would not notice the DoF different or the diffraction difference. But when you want to crop to the limits of the sensor, you will start to notice the per-pixel effects. The 1.6x more cropping you can do on the A7RIV will only be as good as the DoF and diffraction allow. If you are not cropping down that much you won't notice it.
If you take the approach that I want the same composition (not maximum cropping) for a scene, the A7RIV will most often give the better results. I think the only times it would not is when you don't have enough light and need to either increase the f-stop or ISO in to the ranges where the A9 would have a clear advantage.
The OP also specifically asks about a 400mm f/2.8 with a 2x TC for 800mm optical focal length. That is another reason I harped on the DoF issue. It gets really small at that optical focal length unless you stop down, but stopping down will start to hit your diffraction limit sooner on the A7RIV than the A9 -- assuming you want to crop heavily, which the OP does. The DoF I listed predict that at f/8 the A7RIV at 800mm at 40m is 0.28m and the A9 is 0.47m. That might be really significant for you if shooting a large bird if flight and you want the whole wing on your side to be sharp.
If you are shooting at shorter optical focal lengths (say 400 - 500), it will not be such an issue and the A9 will lose that advantage (or at least it becomes less significant for bird-sized things).
3) The lloyd chambers quote.
As I tried to explain in #2, DoF is a multi-pixel effect. A single pixel does not have DoF. If you are looking at a full-res image at a modest enlargement (screen or print), you will not notice the DoF difference. But if you enlarge greatly or heavily crop, you will start to see those pixel-level blurring. If you are not pushing the detail to see the benefits of all those pixels, then the DoF different will not matter much, but if you want to exploit all those pixels it does matter.
For example, some people use super-high resolution prints, so even at modest enlargements they get very high detail and micro contrast. It adds a lot of pop and liveness to the print. In that case, you need to be aware of DoF and diffraction as they will rob you of that fine detail at the pixel level. If you are doing a 300 dpi print to an 8x10, it will not matter.
Marc