Note: I'm not asking for ISO advice advice like "what should I do to get less noise? what settings should I use?".. I'm past all that. This is just to improve my understanding of how ISO works.
Many articles that claim to tell you how it works, don't... they boil down to "here's what settings you should use" and "here's what happens if you don't". They gloss over how it actually works.
Example: "The reasons for potential decreases in quality when only increasing exposure in post-processing have to do with digital amplification (post-processing) versus analog amplification (changing ISO), but they are beyond the scope of this article."
It seems like post-process brightening cannot be better than brightening with ISO, for some physics reason. At best it can be roughly the same. But why is that?
Maybe the answer lies in this pretty detailed answer, but I think anyone who understands a subject well enough to can give an "ELI5" summary.
When I look at examples of supposedly "isoless" cameras, they still have a very gentle upward curve on photonstophotos.net, indicating they aren't quite isoless, it's still slightly better to shoot with proper ISO. Except one example, I found, the Fuji XT1, which has this odd dip, which then resumes its near-horizontal plot on the graph:
So what can this analog process do, apparently post-capture, that can't be replicated by a digital process?
To forestall some anticipated responses:
• I already know that real light on the sensor is better than boosting iso... whether that's through slower shutter speed, larger aperture, or more external lighting. More signal = less noise.
• I'm aware that increasing ISO isn't actually increasing 'exposure', because the amount of light hitting the sensor doesn't change. So changing ISO alone, without touching shutter speed or aperture, is not an effective way to ETTR.
• My understanding is that ISO happens after the exposure is captured, that the photons create an electric charge, which are fed into some analog-to-digital converter, and now the charge becomes the 1's and 0's needed to make a brightness number (typically 8 or 10 bit).
• I know that some cameras are fairly ISO invariant, so that you can pretty much shoot at base ISO all the time and brighten in post, and it's almost the same as just using ISO during the capture.
That's the part that baffles me: the "almost"
If ISO boils down to an analog way to post-process the shot, why is that analog way always superior to any digital post-process brightening, even by just 1%?
Many articles that claim to tell you how it works, don't... they boil down to "here's what settings you should use" and "here's what happens if you don't". They gloss over how it actually works.
Example: "The reasons for potential decreases in quality when only increasing exposure in post-processing have to do with digital amplification (post-processing) versus analog amplification (changing ISO), but they are beyond the scope of this article."
It seems like post-process brightening cannot be better than brightening with ISO, for some physics reason. At best it can be roughly the same. But why is that?
Maybe the answer lies in this pretty detailed answer, but I think anyone who understands a subject well enough to can give an "ELI5" summary.
When I look at examples of supposedly "isoless" cameras, they still have a very gentle upward curve on photonstophotos.net, indicating they aren't quite isoless, it's still slightly better to shoot with proper ISO. Except one example, I found, the Fuji XT1, which has this odd dip, which then resumes its near-horizontal plot on the graph:
So what can this analog process do, apparently post-capture, that can't be replicated by a digital process?
To forestall some anticipated responses:
• I already know that real light on the sensor is better than boosting iso... whether that's through slower shutter speed, larger aperture, or more external lighting. More signal = less noise.
• I'm aware that increasing ISO isn't actually increasing 'exposure', because the amount of light hitting the sensor doesn't change. So changing ISO alone, without touching shutter speed or aperture, is not an effective way to ETTR.
• My understanding is that ISO happens after the exposure is captured, that the photons create an electric charge, which are fed into some analog-to-digital converter, and now the charge becomes the 1's and 0's needed to make a brightness number (typically 8 or 10 bit).
• I know that some cameras are fairly ISO invariant, so that you can pretty much shoot at base ISO all the time and brighten in post, and it's almost the same as just using ISO during the capture.
That's the part that baffles me: the "almost"
If ISO boils down to an analog way to post-process the shot, why is that analog way always superior to any digital post-process brightening, even by just 1%?
Last edited:
