The camera will, at the end of the day, always arrive at ONE reading/estimate of exposure.
There will always remain areas over and under exposed. It is the camera man to decide whether the areas "correctly" exposed is "correct" for the picture's purposes.
And the most straightforward way to do that is spot metering - you tell the camera exactly where you want it exposed "correctly" and you can ignore all the algorithms and databases and blackboxes the dumb camera uses to decide on exposure, if in the first place these are unsatisfactory to you, the cameraman.
But if there is more than one area that are differently lighted and that you want correctly exposed, then it cannot be done in the camera. No camera ever could.
But with digital photograph, you can do local processing off the camera.
Then the question is a different one now, namely what is the best compromise exposure, derived in some ways from the to-be-correctly exposed areas, that best facilitate post processing?
I do not know what this answer is.
The best exposure for me is one that allows maximal data capture with my sensor, the rationale being that I have also maximal flexibility in data manipulation - no data no manipulation - to make that correct picture, which may include multiple correctly and differently exposed areas.
There will always remain areas over and under exposed. It is the camera man to decide whether the areas "correctly" exposed is "correct" for the picture's purposes.
And the most straightforward way to do that is spot metering - you tell the camera exactly where you want it exposed "correctly" and you can ignore all the algorithms and databases and blackboxes the dumb camera uses to decide on exposure, if in the first place these are unsatisfactory to you, the cameraman.
But if there is more than one area that are differently lighted and that you want correctly exposed, then it cannot be done in the camera. No camera ever could.
But with digital photograph, you can do local processing off the camera.
Then the question is a different one now, namely what is the best compromise exposure, derived in some ways from the to-be-correctly exposed areas, that best facilitate post processing?
I do not know what this answer is.
The best exposure for me is one that allows maximal data capture with my sensor, the rationale being that I have also maximal flexibility in data manipulation - no data no manipulation - to make that correct picture, which may include multiple correctly and differently exposed areas.