???
You must not have ever done CCTV systems 'professionally' when you were installing. Or gotten into photography professionally or as a hobby.

Picture quality is a function of the imaging sensor, lens, and processing.
Imaging sensor: The smaller the sensor gets, the worse low light and overall detail gets. They have steadily been making larger sensors to put into smartphones but there is serious limitations to work with. It is the primary reason why smart phone cameras suck and always will (compared to a dedicated camera).
Lens: Glass matters. Not only the material it is made from, the coatings used, and the layers but also the size in relation to the aperture and F stops. Again, small lens suck and degrade picture quality.
Processing: The DSP has become very important today in getting the most out of a less than ideal camera setup. It is the primary reason for the advancements we do have in smart phone camera quality. But more advanced (and power hungry) DSPs require space and heat sinks. And cost money. You are not going to get a Pixel3 grade camera 'system' at a budget cost. And like mentioned, it still is going to look like crap in the dark with such a small lens and sensor.

So in conclusion; your design criteria for the camera in a doorbell, is unrealistic and defies the laws of physics. You WILL have to make compromises on 'your' personal aesthetic choice, if you want even a 1/2 arse quality image from your doorbell.