Hi,
Technically speaking nothing of that is true.
The pixel count really decides how large you can print with good detail. Present displays cannot reproduce more than 8MP (for 4K). So anything you see on screen is downsized, with downsizing adding a lot of artifacting of it's own. Or you can look at a picture at actual pixels which results in obscene magnifications.
Regarding the number of different tones a system can reproduce, that is limited by what is known as full well capacity. At full well capacity any sensor clips. That applies to S2, Nikn D810 or essentially every camera ever made with the exception of some early dual pixel Fuji models.
So all this talk about highlight roll off is pure nonsense, technically speaking. On the other hand, a camera system may underexpose and create headroom and different camera profiles may give different compression of highlights.
Check the image below (or this link for full size:
https://4.img-dpreview.com/files/p/E~forums/58984278/2ae5eacd49d34b248c71368a4a313844 )
The image shows a print white text over a grey gradient going from Lab (100, 0, 0) to Lab(80, 100, 100). The mage here is a print, so paper white limits white to around Lab (95, 0, 0). The text begins with Lorem ipsum and the first separable letter in the original image is the 'm' in Lorem. It is also the first letter separable in both images. Top one is Sony A7rII and close relative of the Sony Exmoor sensor used in the Nikon D810, the one at the bottom is from a P45+ back using a Kodak CCD based sensor very similar to the the one used in the S2.
So what this illustrates is that old or new sensors, the highlight handling is virtually identical. Now, raw processors can add highlight roll off, but that is not related to a camera. The profiles are supplied with the raw converter (*).
The same posting contains an extremely high contrast image covering a luminance rage > 13EV, with some controlled and contained higlight clipping. Here is the highlight part:
(Here is the link to full size:
https://2.img-dpreview.com/files/p/E~forums/58984278/7ca61b80d0494593a586be910e162a8c )
The crops are from P45+ (Kodak sensor similar to the one used in the S2), Sony A7rII (present generation Sony Exmoor) and Sony A900 (using first generation Sony Exmoor, same age as the P45+). Not a lot of difference.
Now, what has happened in the last ten years is that pixels went smaller while Full Well Capacity was maintained. But with modern CMOS the readout noise has been much reduced. So we had a hike in resolution while highlight capacity was maintained, but darks got much cleaner. Lets look at the dark side of the previous image:
https://4.img-dpreview.com/files/p/E~forums/58991988/9b2c3bf485cc41e69d5c5c448c7287daHere you can see that the Sony A900 and P45+ images have real problems while the Sony A7rII still has clean darks. That means that the Sony A7rII is significantly superior to the older cameras in simultaneous handling highlights and deeps darks, it has more dynamic range.
Smaller pixels reproduce better detail. Large pixels cannot resolve fine detail, so the fine detail the pixels cannot handle is converted into artefacts.
Check the images below. They are shot from the same position, both using the same Hasselblad Planar 100/3.5 lens. The image on the left is small pixels (4.5 micron pitch) while the one on the right is larger pitch (6.8 microns). The small pixel sensor resolves the image much better.
(
https://3.img-dpreview.com/files/p/E~forums/58656097/6bf08a08d72e470faf3ea8993ab7d0ec )
Downressing the 4.5 micron pitch image to lower resolution of the 6.8 micron sensor still results in a much cleaner image:
(
https://3.img-dpreview.com/files/p/E~forums/58656097/313cbbac87a442679b634149c5e99a1e)
A high resolution lens combined with a low resolution sensor is not a great combination, as it will produce a lot of fake detail, a phenomenon called aliasing. The fake detail is an alias of the real detail that is not resolved.
So, your friend is wrong, technically speaking, and I think Leica needs to reconsider their pixel sizes. Just to say, bad technique can mask the disadvantages of large pixels. Shoot handheld, stop down to f/16 and don't focus properly and there will be no fake detail, as a matter of fact there will be barely any fine detail at all…
Here are the links to the original postings:
https://www.dpreview.com/forums/post/58984278https://www.dpreview.com/forums/post/58656097(*) The Leica S2 uses DNG as file format. So the image can have an embedded DCP profile that will yield Leica's preferred rendition. The few Leica S2 images I have do have an embedded profile.
Best regards
Erik
An old friend who is a Leica chauvinist claims that the Leica S2 makes far better images than the Nikon d800 because despite the same pixel count Leica images have greater and more subtle gradations of tones.
This point generalizes to tonal quality versus resolution.
I wonder what the experts here think of this controversy and if it is valid which cameras solver the best tonal continuum.