Quote Originally Posted by Karsaa View Post
Then there is one thing that myself would love to see on tests. Low light vs night time photography. Nobody seems to do good tests with cameras on night time, like shooting astro etc. Those low light tests are quite useless on my point of view. Why nobody goes and testes taking real low light images, milky way etc.
As you know, Clarkvision is still a great resource, but has not kept up with current cameras. Then I tend to use photonstophotos.net for read noise, as that should cover the key difference between a well lit subject and a poorly lit subject. The first link were to cameras I've owned, this link is more relevant to you as I think you have shot the 6D and now R6. As I recall, you provided a link awhile ago where the thought was as you drop below 1 Log DN unit, Read Noise would not be much of a factor even for low light photography. You can see the current sensors dancing around that value.

And I agree. There has been a shift, most noise tests are well lit. Two to three sensor generations ago, these tests could see a difference in camera bodies even with well lit subjects because the read noise was higher than today. Now, read noise is much lower, pretty much for all cameras, so these well lit tests are mostly observing shot noise at increasing ISO, which is the same for any image. So, it is no surprise that when you normalize all factors, you tend to see very similar results. This tells you most cameras today are very capable for most types of photography, but as you get into specific types of photography that diverge from their test conditions, it can matter. For example, a very low signal to noise ratio, you can still see the effects of read noise.