Usually - IF optics + sensor + image processing trio is producing a good quality video stream without limitations/artifacts (like with low bandwidth video compression codec) at a given resolution like 4K, THEN it makes sense to use higher resolution even when target resolution for final production video will be lower, like 1080p.
In this case video stream is resized in post production using some good software.
Final result will be (if all the IFs are properly handled) better then native 1080p recording coming from the device.
However if you need to make compromise when shooting in larger resolution, like codec limitation, image sensor sampling limitation, artifacts, etc, then this theory is broken. As the higher resolution recording already makes trade-offs in possible maximum quality or framerate or storage space/memory card writing speed.
The sweet point is different case by case.
Did some tests myself:
GoPro Hero7 Black Action camera - I wanted 60fps that was the baseline. After that maximum video quality possible. As I have only 1080p displays at this time I don’t need larger resolution. Targeted video audience also using 4K displays rarely.
4K@60FPS disabled image stabilization in the camera so despite good baseline video quality (no trade off in recording bandwidth or sensor sampling) it was disregarded.
2.7K@60Fps and 1080p@60FPS were identical In everything except resolution.
Post production (software) downscaled 2.7K -> 1080p video shows significantly better details all around, then 1080p video directly from the camera. This is logical as the post production processing doesn’t need to be realtime, you have more time and more processing power to use more sophisticated processing algorythms.
Same excercise on my
DJI Mavic Pro resulted shooting in 1080p only. This device has a 65Mbps codec bandwidth which introduces artifacts in all modes above 1080p (due to compression) and also image processing and coding engine make tricks at 2.7K / 4K to be able to process video coding in real time with onboard limits of SOC resources provided by Ambarella A9. This is a fairly weak SOC also used in GoPro 3.
So this is always case by case what and how the manufacturer implemented using which hardware and what is the desired expectation of the user.
If properly implemented, despite using the Sony Exmor 583 sensor (and its limitations) an
EVO2 8K post production downscaled video can have better quality than a 4K native video from device. Or if the post processing zoom possibility is interesting, than it gives 8K some edge, the way OriginaldoBo used this on the helicopter scene.
TBH I'm really interested in the
EVO2 6K model and if it will be
capable shooting videos without those tricks what DJI Mavic 2 Pro has to use to process 4K videos onboard in real time.