Bill, when I look at the Northrop video I see that the lens is a constant and the variable is the type of sensor of each camera attached to that lens - which appears to be the exact opposite of what I understand you are seeing. If the lens was varied then the experiment would be totally invalid, but it wasn't.
Part of this, as I have commented in other materials, is one of nomenclature. I look at the image that a lens projects onto a sensor, which is essentially a cone of light, and see that as delivering an Angle of View. Angles are measured in units of arc, not mm of x by y. What the sensor captures from that projected cone of light is what is actually recorded - which is what counts in considering the end result, so the sensor must be considered - and is measured in mm x mm. All consumer cameras use some kind of rectangular image within that projected cone. Given that convention has identified a sensor of 24mm x 36mm as 'Full Frame' (and I agree with others that this is a rather ad hoc thing, but it is generally accepted).
As to lenses... I think we all agree that any lens of a given focal length and aperture range has those characteristics within its physical makeup no matter what camera body or sensor it is associated with. Where I think we differ is that I consider that the resultant image is the result of the angle of projection of the lens together with what is captured by the sensor. This, to me is not an angle of view, it is a Field of View or Field of Capture. It is the terminology and lack of differentiation between lens output and sensor capture that starts to undermine the discussion. It soon progresses into talking 'chalk and cheese'.
I won't go back over the ground that Norton, who by the way is only ONE reference I provided, including some seriously respected publications and authorities in optics. So far, from what I see the contrary argument is based on personal belief or experience.
I will suggest as regards looking at the impact of the combination of lens + sensor to simply go through the various Depth of Field calculators available on the market. For ONE example, I will give a shot and link to that explained and used on the website Cambridgeincolour - a well-respected website: the link is HERE . One can see how to get the same effect for perspective and DoF one must use different values for different sensors. Here a FF camera with 50mm at f/8 is compared to an APS-C camera and to get the same result it requires substantially different characteristics for a lens.
The phenomena of reduced FoV and DoF exist, and there is lots of documentary and practical proof to that effect. The question is, is that of any practical impact on a photographer? I see that Ernie does not see that it does, and is rather dismissive of those who may have ventured into this territory in his classes. I don't dismiss it or those that ask, I engage with them. In the majority of my classes - especially those for non-technical novices, I don't bring it up unless asked - but it DOES come up where people are engaging with gear between sensor platforms.
In one case I had a chap in a class who was getting into real estate photography. He had a friend who was already doing this, using a FF Canon camera and an EF 17-40L lens. He tried that lens out on his friend's camera and based on its performance, he bought the same unit. He was upset that he didn't get anything like the coverage on his own camera. Why? Because he was using an APS-C camera and the Field of Capture was significantly reduced to an effective 27-64mm FoV. The lens he should have bought would have been in the region of 10-20mm - like the EF-S lens of that range that would have rendered a FoV of 16-32mm. The issue was he was ignorant of the phenomenon and worked off the numbers on the lens alone, without considering the differences in sensor.
As to DoF. I agree that this is an issue to a much smaller constituency: I would suggest predominantly those engaged in portrait photography where DoF and the ability to limit that with wide apertures is of significance. That is what I saw demonstrated by Northrop. The phenomenon of DoF variance when trying to maintain a level of perspective and depth of field is measured in the calculators of which the one I liked to is an example.
Finally, as to the relevance of considering how sensor and lens designs will interact is demonstrated by posts here and in the documentation for the R-series FF bodies. I have seen people wanting to use EF-S and RF-S lenses on their FF R-series bodies. Now, it can be done for the former via the EF-RF adapter, or directly for the RF-S lenses. It should be noted that the 15-45mm and 18-135 lenses now available are meant to be used on the R7 and R10 bodies but CAN be physically connected to the FF bodies - this was not possible on the DSLR platform, so was not an issue. Because those EF-S and RF-S lenses are designed for a smaller crop sensor, they do not project onto the whole area of a FF sensor, and this results in the effective resolution being reduced dramatically - by a factor of 2.56 in fact. This is not obvious to those not aware of the context within which these two platforms engage, but is it buried in the manuals, for example this one for the R6 body that documents how the resolution is reduced from 20MP to 7.7MP:
My point is that these are traps for the unwary and they are of practical significance, not just theory. So consideration of these issues is not unreasonable.
