Now we see through a glass, darkly...
- Apr 5, 2016
There are noise comparison tools and if you for example compare the 1D X and the 1DX Mark III, there are eight years between them and you see some noise improvements, but not a lot. It is definitely much less than one stop in eight years. I would have expected at least two stops of improvement in that time frame. In the eight years before that 1D X, we easily had over two stops of improvement.
Here is a noise comparison tool to play around with:
Everyone had more improvement, noise wise, between 2004 and 2012 than they have between 2012 and 2020.
We've been approaching the theoretical limits of the random nature of photon distribution in a light field of given intensity over time (Poisson distribution noise, which is a property of light, not of cameras and sensors) for several years.