ISO Invariance

[HDNitehawk]

here, but smaller process generation allows for the construction of smaller, more efficient (less noise) systems on the chip level. It also seems that the smaller substructure may be important for moving the ADC onto the chip.

Larger pixel = Less Noise due to the S/N ratio.
Smaller more efficient structure would allow for better handling but you end up with an unfavorable S/N ratio to deal with as pixels get smaller.

Since you posted this I have been doing some reading, I think to develop a good understanding a line diagram would be helpful. There are multiple steps to remove various types of noise during the process.

This is wrong. ISO should not be part of the "exposure triangle." Actually, there is no triangle. Rather, you have the light that hits the sensor controlled by aperture and shutter speed.

I do not agree with this statement. You have to have ISO because in many cases the required shutters speed and aperture required is outside of what is possible. You have to have the third element to compensate.

ISO is the boost in gain of the signal for a digital camera. Much of what you are talking about is how it is handled in camera and in computer.


I did a quick google search. This miss-leading explanation popped up, this is a dumbed down statement for the masses. Most people this explanation will suffice even if it is not accurate. You really do not need a higher understanding to take great pictures.

"In Digital Photography ISO measures the sensitivity of the image sensor. The same principles apply as in film photography – the lower the number the less sensitive your camera is to light and the finer the grain. ... By choosing a higher ISO you can use a faster shutter speed to freeze the movement."

If any part isn't real it is "ISO Invariance", it is nothing more than your ability to adjust the gain at a different point other than inside the body of the camera. Why would you even want to do this and underexpose pictures, so you can chimp your black pictures? What would really be the functional use for this "knowledge", other than you now know you can adjust your ISO in post without penalty?

[Kayaker72]

Larger pixel = Less Noise due to the S/N ratio.
Smaller more efficient structure would allow for better handling but you end up with an unfavorable S/N ratio to deal with as pixels get smaller.

Since you posted this I have been doing some reading, I think to develop a good understanding a line diagram would be helpful. There are multiple steps to remove various types of noise during the process.




I do not agree with this statement. You have to have ISO because in many cases the required shutters speed and aperture required is outside of what is possible. You have to have the third element to compensate.

ISO is the boost in gain of the signal for a digital camera. Much of what you are talking about is how it is handled in camera and in computer.


I did a quick google search. This miss-leading explanation popped up, this is a dumbed down statement for the masses. Most people this explanation will suffice even if it is not accurate. You really do not need a higher understanding to take great pictures.

"In Digital Photography ISO measures the sensitivity of the image sensor. The same principles apply as in film photography – the lower the number the less sensitive your camera is to light and the finer the grain. ... By choosing a higher ISO you can use a faster shutter speed to freeze the movement."

If any part isn't real it is "ISO Invariance", it is nothing more than your ability to adjust the gain at a different point other than inside the body of the camera. Why would you even want to do this and underexpose pictures, so you can chimp your black pictures? What would really be the functional use for this "knowledge", other than you now know you can adjust your ISO in post without penalty?

Hi Rick...No, this is not an attempt to say never use ISO. In terms of practice, I have every intention to continue to use ISO as it makes little sense to run around shooting black images and add gain in post. But, in some instances, I might use ISO differently and hopefully, better.

And how to best articulate this is still evolving, but I think we can agree that ISO is something that occurs after the image capture (electrons captured-initial analog signal generated) . Yes, in actual photography when you want to control exposure, ISO will influence your selection of shutterspeed and aperture. So, ISO's influence on exposure is indirect as in it does not directly influence the sensor itself.

It is, of course, a good question of how to use this type of knowledge. First, ISO is not a lossless process. Adding ISO, or gain, at the analog stage causes you to lose information as ISO raise the floor, but the ceiling does not change. So, if your goal is to maximize dynamic range and do care about shutterspeed or aperture, you should shoot at the lowest ISO where your camera becomes ISO invariant. Then, in post, you can control whites, highlights, midtones, shadows, and blacks separately. In short, shooting at the lowest ISO where your camera becomes ISO-invariant maximizes the information you have to work with in post. I've had a rudimentary appreciation for this, but working through this thread is highlighting this approach (this is what astro photographers are doing).

But really, ISO invariance is not a cause, it is a result of a lower noise floor, better handling of the ADC, and overall cleaner post image capture process. There are lots of uses of that knowledge and understanding (granted, each is camera specific). In my rudimentary understanding, I was already shooting a bit differently for high contrast scenes. For example, I will shoot a loon with more light on it than before, knowing I can expose for the white breast and bring up shadows in post. I have seen references to landscape photographers shooting with few exposures in an exposure bracket to capture a scene. Then, as mentioned, I am seeing astro- or nightscape photographers use ISO invariance to optimize their settings. I am still digging into this, but part of this has made me recall some videographer discussions I saw a year plus ago that talked about controlling dynamic range above and below "proper exposure" (18% gray) and maximizing dynamic range at different ISOs. As video is baked in jpeg equivalents, this makes sense.

So, I am still working through some of this, but thinking of ISO as gain that is applied to the analog signal rather than something that inherently and directly impacting the image on the sensor can change a few things. In the future, I will be thinking more about when it is best to add gain, in camera (often) or in post (sometimes).