>>3408245>obvious>>3408153>Sure think mr. olympus is cheating at iso...Okay, so, imagine you have an image sensor.
Two companies use that image sensor. One of them thinks it's a bit more important to have extra shadow detail you can recover. The other thinks it's more important to have extra highlights you can recover.
So when you set ISO 100 on both of them, one of them ups the sensitivity a little more which gives you a slightly noisier image but with more shadow detail, and the other uses a lower sensitivity which gives you a slightly less noisy image with more highlight detail. Both of them output JPEGs that are exposed the same.
Now, is one of them cheating? No, they're just prioritizing different things. A site like dxo has to normalize these so they can be compared, so they have to use a different method to calculate the ISO on the raw image. It doesn't mean that one manufacturer's ISO rating is "wrong"--DxO even says so in their explanation of ISO--it just means that one company prioritized a different part of the histogram than the other.
The key thing to understand is that there's no definitive mapping from a given number in a raw data stream to a particular shade in the output image. Everyone can agree that all zeros is gonna be solid black and all ones is gonna be solid white, but every value in between those is determined by how the software processing the raw decides to do the curves. In fact there *can't* be a definitive mapping because some sensors have more dynamic range than others, and some sensors have more bits than others, etc.
Of course, no one is going to consider this argument because they're too invested in the idea that some companies are BAD and their company is GOOD so if their company's ISO matches more closely the arbitrary measured ISO value that dxo uses, that means their team is winning and the other is cheating rather than a perfectly valid difference in design philosophy.
GO WILDCATS.
