So, one could compare a subject crop from a high MP capture to the same subject captured with a telephoto on the same camera, same distance, na-ni-na. Or, more conveniently, a zoom at min and max FL. The final images with the one image having more pixels that the other based on angle of view (steradians?).
If the subject were a slanted edge, MTF could be the metric that says which is ābetterā ā¦
ā¦ almost tempting ā¦
By taking pictures of a test chart and comparing the resolution targets.
Or one could look at an MTF chart for a lens and calculate the angular resolution at the center for some acceptable cutoff (say, 0.8 or 0.9, depending on the lens line), and convert that to pixel size for each sensor.
I am pretty sure that manufacturers of course calculate a combined MTF for lens + sensor as part of their design process, but I donāt think they make those public. In any case I would guess that current sensor resolutions are within a factor of 2 of what their pro lens lines can resolve for each manufacturer, and thatās why R&D is not spent on increasing sensor pixel density because it does not buy you that much; just makes your image files bigger, increasing camera and desktop processing time, etc.
Note that DXOMark has a neat summary score called perceptual megapixels, which is a neat way of summarizing lens performance. Eg if you look at the scores for the 15mm f/1.7 PanaLeica, which is a very sharp and well-designed prime lens, you see that 10 megapixels is more or less enough for the information you get from this lens. Most lenses are probably worse. Sensor sizes for micro 4/3 are already double than this in recent cameras, but that may just be marketing pressure, not something driven by lens quality. I guess the situation is similar for most mounts.
Time to glue 4 247MP cameras together and get nearly 1000MP of resolution, now this is the way forward 
Good! Thanks.
An interesting approach - thanks!
Long ago, I bought QuickMTF Light which produces a system MTF chart from an image containing a slanted edge. Now Iām playing with some slant-edge shots to somehow simulate cropped versus tele shots. Will post if that gets anywhere ā¦
I like angular measure. One could also consider an object of angular size āxā radians where the crop versus the lens FL produce the same angular size (frame) for a given shooting distance.
Thinking more about a crop versus a long lens, one could consider extremes, i.e. at some known distance, equal-framing crops of some distant object with largely different focal-length lenses. Like a crater on the moon shot with a 50 mm versus a 1000 mm, Certainly a huge difference in pixel count for any sensor but what metric would tell the tale of relative image quality between the two captures?
Iām thinking that DxOās blur units might work but need to understand them better.
[edit] initial reading suggests that the āblur unitsā metric removes the cameraās sensor MP count from the equation, relatively speaking.
There is a widget that appears to measure blur units for a lens at various settings. I looked at a Sigma 18-200mm zoom:
https://www.imaging-resource.com/lenses/sigma/18-200mm-f3.5-6.3-dc-os-hsm/blur/d200-(2010)/
It seems that a crop from the middle at 18mm FL would have about 1 blur unit but full frame at 200mm would have about 2 blur units all over.
That could be taken to mean that, for that lens on the same camera, the crop shot would be ābetterā unless Iāve missed something - I did find the DxO article to be way above my 84-yr old head.
All those low pixel count cameras are just childās toys.
THIS is a proper camera sensor:
So just a lot of 16 MP toys, eh?
An hypothetical 16 MP camera with 6um pixel pitch and a 600mm lens subtends 2.06 arc-sec ā¦ so must be quite a big equivalent focal length in the Observatory, eh?
My day job is in automated semiconductor microscopy. One of our machines takes full-wafer scans, which routinely amount to dozens of gigapixels per wafer. Those are stunningly detailed images.
I need a new house with huge walls to hang the images.
Thatās playing āWhere is Waldo?ā on a really high level and the aim is to have no Waldos. 
Cool stuff.
Weāre looking for single-pixel defects. Itās interesting work!
Iām currently trying to convince management that we need a full-resolution print of a (part of a) wafer. Just to get a sense for the scale of it.
I switched tracks to image processing from audio signal processing. Fun fact, it wouldnāt have been possible to do that without the many discussions about photo processing algorithms on this here forum!
I did get the 70 - 300, 1.4 TC, and X-T5 all recentlyā¦ the X-T5 was a few weeks after the other two. Yeah, thatās a lot of new gear. Whoops! 
Anyway, I noticed that the 70 - 300 + 1.4 wasnāt so hot on the X-T2, but itās much, much better on the X-T5. Different days, of course, so different lighting and allā¦ but I think the difference was that the focusing of the X-T5 was able to nail the shot, whereas the X-T2 couldnāt quite get it perfectly right. And then if you have bird or animal mode turned on and are shooting a bird or other animal, then the X-T5 really shines.
From my own ātestsā, the 1.4x is still āworth itā over cropping in some cases (distant birds in trees on a sunny day). But itās a hassle otherwise, and being able to crop 40 MP down is excellent anyway. Iāve done both on both days.
But this is all anecdotal and comparisons of real birds in the middle of the lens and even stopped down a little bit, not a test chart. Itās still almost close enough to not be absolutely sure, however, and the hit to the focus speed (which is supposedly minor with the 1.4x compared to the 2x, I hear) and the loss of the amount of light still definitely makes the 1.4x TC not worth it under many circumstances.
(In other words, I mainly agree that itās not entirely worth it, except possibly under certain circumstances, like enough light and (probably) stopping down a notch or two.)
But, anyway, the big takeaway: You really need to have top-notch lenses and also focus them perfectly well to really take advantage of larger sensors. This probably holds true for any size format of these larger lenses. Basically, the larger the resolution of the sensor, the more important lens quality and perfect focusing becomes. (Thankfully, changing the aperture helps with getting things in focus, if thereās enough light.)
Pardon my confusion but the degree of focus exists in the image plane, so the resolution of the sensor simply affects how well that degree of focus is resolved?
Iāll just wait until AI is able to seemlessly upscale my mere 16MP images to 16,000MP.
Just think how many easter eggs you could hide in thereā¦
I just posted in another thread but maybe it was more relevant here.
The high mexapixels are a big part of this technology, but also the upscaling/sharpening technology. I canāt help feeling a little sad at how āeasyā it is becoming to achieve results that hobbyists and professionals have spent a lot of money and time to achieve.
Why do you feel this way though? That is how technology democratizes things.
Itās a good question and Iām not sure I have a good answer. I guess itās to do with the fact that us hobbyists put money and effort into owning/learning something, and then there is pride in the unique results you can achieve. It garners interest when you can show your work and the average person canāt do it because they havenāt got the gear or the skills. I donāt think itās selfish pride and I donāt actually need admiration from others to enjoy the hobby. But I think the reward means something to me, and that reward comes from the challenge (and the money invested).
It lessens the reward when anyone can just whip out their phone and get similar results. Obviously certain skills still need to be developed, but you can still get lucky by taking a photo of a golden eagle on a branch and it looking like it was shot on a 600mm telephoto lens.
Itās not gatekeeping because I actively encourage anyone to get into photography, but I admit I have less interest in things when they become mass market and/or low effort. I guess I can also draw a parallel with my job (translator/editor) where my skills are becoming devalued by technology. Itās hard to take when youāve been doing something for decades and it suddenly becomes something everyone has access to. Iām willing to have my mind completely changed, though, if anyone can convince me to be happy about all this.