Equivalence of ISO Noise on different Sensor Sizes

Thanks to DPReview, I can demonstrate the equivalence of ISO noise on different sensor sizes. In various DPR reviews, you can find a comparison tool that helps estimate image quality. You can even download the JPG and the RAW files.

In the setup above, I compared three cameras at ISO6400. Although a much higher ISO may be needed for some photography, even this ISO setting clearly shows the limits of the small sensors. As explained in much detail in this blog, those limits are often no real limits, because you can find equivalent settings in most everyday situations.

We study the differences, not to bash small sensors,
but to learn how to handle them.

I am only interested in the details. So, here is a magnification.

The cameras shown here are the following, all with comparable pixel counts.

• The "full frame" Nikon Z5(factor 1.0)
• The APS-C (aka DX) Nikon 7500 (factor 1.5).
• The Micro Four Third (MFT) Olympus OM-5 (factor 2.0).
• The Sony Cybershot DSC-RX100 VII (factor 2.7).

Clearly, the noise gets worse in each step down the sensor size. The Nikon Z5 no longer produces a clean image, of course, but a usable one. I allow it to go up to ISO1600, which is sufficient for my photography. The photo at the start of this posting was taken at ISO 1600.

However, we can also use the tool to compare equivalent ISO settings. For this, we decrease the ISO approximately by the square of the sensor factor. 

Now, the images look pretty similar. Any differences may be due to the lenses DPR has used. As far as I know, they used the best prime lenses for these shots.

We learn that there are equivalent ISO settings
for different sensor sizes.

As explained in other postings, you can compensate for these settings with different apertures to maintain the same DOF and background blur. In many situations, you get equivalent images. 

Here is another spot.

If you are shooting out-of-camera (OOC) in JPEG format, the results depend on your camera's noise reduction.

The Nikon Z5 was set to a very aggressive noise reduction. I do not like that result because many details are lost. 

Of course, any noise reduction applied in post-processing to the RAW file is superior to in-camera processing. There is no magical internal way to process RAW files in these cameras. Note that some cameras (including the Sony RX100) offer a way to combine multiple shots into a single, higher-quality image or an HDR image. This is another story.

We need to be aware that the equivalent ISO settings require a different f-stop or a different shutter speed. In wildlife photography, the shutter speed must be high enough and is thus set accordingly. So, you need a wider aperture to get the same image quality. But that yields higher costs and a larger lens, which erodes the advantage of the smaller camera. So, a larger sensor is an advantage in low light. I have discussed this with examples already in another posting.

Choosing a sensor is not easy. All sensors provide advantages in some situations, and disadvantages in others. E.g., the Sony RX100 is a pocketable camera. This is hard to beat, even with MFT. In fact, the OM cameras are rather huge for MFT. With comparable lenses, I see no size advantage over the Nikon Z5 (see the image below). 

There are small MFT cameras and small telephoto lenses. For hiking birders, those might be worth consideration.

However, one needs to add that a lens for MFT must resolve twice as well as a comparable lens on full frame to get the same resolution. The Nikon Z lenses have an incredibly high resolution in the center, even wide open, but not on the edges of the frame. I don't think MFT lenses can be twice as good. 

In real life, people produce excellent images on all sensor sizes. A pocketable camera like the Sony RX100 may be the best solution if you would otherwise have no camera with you. You need to be less critical of the dynamic range in some situations. And due to its zoom lens and rich features, it still surpasses any smartphone.