The author of the article is technically correct
No, he is not
(the best type of correct!). All the people calling him "misinformed" should take a while to think before committing this sort of language to the permanence of the internet.
That being said, you may not agree with his practical conclusions.
The size of the sensor is irrelevant to the quality, DR, etc. It's the quality (and indirectly, size) of the photosites that determine that. A sensor is simply a collection of individual photosites, and the whole doesn't affect the individual photosite.
But what is really relevant here is the DR and quality of the
image, and the image is always made up of an aggregation of photosites (images made from single photosites would be a bit dull), which mean you will always need to look at the performance of the
aggregation of photosites used. And as long as we use teh entire sensor of both the D800 and the D7000, the D800 indeed will produce better DR and quality by virtue of gathering more light (from its larger area).
The D7000 from a practical standpoint (though not exactly), is just a D800 sensor with part of it chopped off.
Correct.
The d800 sensor does get hit by more light,
Correct, but not accepted by the author of the article
but that light is spread out over more pixels,
No, that is not how it works, which is the core of the misunderstanding.
To achieve that spreading out effect you are thinking about you would need to add a teleconverter, which is btw why you lose light from using teleconverters. And which is why the Metabones adapters who do the opposite thing of a teleconverter actually increases the light hitting the surface of smaller sensors. They 'gather' light to a smaller area as opposed to spreading it out as with a teleconverter.
If your reasoning was correct, for example the Metabones adapter just would not have the effect it has, and believe me it does
so the light gathering ability of any one photosite is the same on both sensors.
Yes, which contradict the very point you were making moments before.There is no 'spreading out' of the light with a bigger sensor, just more light in relation to the bigger area.
And since there is room for 36 million of those photosites in a D800 sensor and a mere 16 million on the smaller D7000 sensor, the image created from the D800 sensor will be made up by 36/16=2.25 times more light. Which is what the author of the article don't accept. Which is why we say he is wrong
Of course, there are considerations about DOF, quality of lenses, etc - but that's not in scope for the article.
In any case, the analysis is correct.. What it is saying is this:
A D800 image at 100% will be of the same quality as a D7000 image at 100%. It just so happens that the D800 image will be larger.
That is a backwards way of comparing images.
From a practical standpoint, I don't quite a agree with the author, as I find that most D800 photos are downsampled when printed/displayed.
Because what really is relevant is to look at the same output size from both.
What this means is that even though from a "pure" standpoint, the D7000 will ahve about the same image quality, from a practical standpont, the D800 will outperform because it's being downsampled. It's almost like pixel binning.
The D800 will outperform the D7000 simply by virtue of gathering more light to create the same image. More data to work with = better end result.
Even if the D800 had a 16 megapixel sensor, the outcome would be the same: 2.25 times more light gathered, better image. But would not gain resolution as we do now as a side effect.
The benchmark tests "normalize" by assuming a common print size - so if you want to compare with same-sized prints, then your conclusion should be that the d800 outperforms, and that you don't agree with the conclusion of the author.
If you want to compare pure quality at 100%, and assume that D800 shots will be printed larger (because you can), then you conclusion would be that the the quality will be about the same, and that the D800 does not outperform on a final output quality.
But that is like comparing two cars, one with a top speed of 70 mph with one with a top speed of 140 mph, drive both at full speed, and complain the faster care is noisier and has a less comfortable ride. It is not a fair comparison. To determine which of the cars who has the most comfortable ride, you need to drive them at the same speed (normalize them).
To compare image quality from any two cameras, the most fair way is to print an image with the same size from both and look at it from the same distance.
For myself, I crop a lot, and I like to print as large, so I can safely assume that the D800 isn't getting me better quality prints, it's just getting me larger prints.
How could the be larger if they did not contain more information?
--
-----------------------------------------------------------
I don't suffer from insanity, I enjoy every moment of it!
By the way, film is not dead.
It just smells funny
