Does changing shutter speed and aperture create equalized exposure?

[Ysarex]

You're playing word games. Have a nice day.

So you put words into my mouth and claim I said something I did not.

And now I'm playing words games.

Here's what you claim I said; "You also said that increasing ISO will increase noise depending."

OK, how about you list the post number where I said that and quote me.

In the meantime here's what I said back on page 2 post #24. "Implemented via amplification of the analog sensor signal ISO suppresses read noise. The shot noise in the signal is not increased or decreased by the action of ISO but it is revealed."

 

[Ysarex]

The other problem with raising ISO is some cameras have maximum native ISO and some have Enhanced. With enhanced the camera sensor will purposefully underexpose the image at the highest base ISO setting, and then through processing, bump up the exposure to simulate what it would look like at your chosen extended ISO. The result is more shot noise as Joe explained above.

No argument here about the process you're describing but we should pay attention to how we name things and describe the process in light of our existing international standard. The ISO standard does not specify implementation. From the standard's perspective there is no difference between the camera achieving the brightening result from amplification of the analog sensor signal or digitally scaling applied in the ADC or simply applying a different tone curve and input profile in the image processor.

The most common way to implement an ISO brightening change is to amplify the analog sensor signal. Over time lots of folks have latched onto that method as defining ISO such that if the camera implements ISO with analog amplification that's "real" ISO and if the camera uses a different method that's enhanced or extended or even fake ISO. It's fine for us to recognize these implementation differences and understand how they produce different results. And we do need names to describe them like extended ISO. Just want to be careful that as far as the standard is concerned they are all legitimate implementations of ISO.

The method of implementing ISO via analog amplification has been preferred because it provided the added benefit of suppressing read noise. As the engineers succeed in removing read noise from the hardware that added benefit becomes less important and we're seeing an increase in use of digital scaling to implement ISO.

 

[smoke665]

One at ISO 200 (base ISO of my camera) and the other at ISO 3200 -- both the same exposure: 1/15 sec, f/5.6. In both I processed the raw files.

I can understand @AlanKlein 's confusion after completing the above experiment on my own. I used ISO 200 and 3200, f/5.6 but went to 1/20 shutter to get a nearly balanced exposure on the left. The only processing done to the raw files was to crop and place side by side, as your example states "both the same exposure aperture and shutter".

So to put the confusion to rest, @Ysarex a little explanation on how your example from your post shows almost identical image brightness, and why mine above doesn't, if in fact you were at the same exposures. Was your first one grossly underexposed and then raised post? I sort of picked up on that idea in a statement you made earlier in the same post, "Raising ISO without lowering exposure won't over brighten the output image if it's too dark to begin with" and that coincides with your premise on shot noise. If that's the case then it's something you didn't make clear in your post.

 

[Ysarex]

I can understand @AlanKlein 's confusion after completing the above experiment on my own. I used ISO 200 and 3200, f/5.6 but went to 1/20 shutter to get a nearly balanced exposure on the left. The only processing done to the raw files was to crop and place side by side, as your example states "both the same exposure aperture and shutter".

So to put the confusion to rest, @Ysarex a little explanation on how your example from your post shows almost identical image brightness, and why mine above doesn't, if in fact you were at the same exposures. Was your first one grossly underexposed and then raised post?

In both of the two raw files that I saved the sensor in the camera was equally underexposed. If the exposure I used for the ISO 200 photo was an underexposure of the sensor then that same exposure used for the ISO 3200 photo was the same underexposure of the sensor. That's why the shot noise is the same in both. The two JPEGs the camera created are below. The difference in brightness of the two JPEGs is the result of ISO brightening applied or not after the exposure. What I did was make sure the exposure I used would not ISO-clip the highlights in the ISO 3200 photo and then I processed the two raw files to the same output brightness. The two camera JPEGs from the same identical underexposure are below.

The correlation between ISO and noise/image brightness in the camera generated output image (JPEG) is strong. The correlation between ISO and noise/image brightness in raw files is weak.

I sort of picked up on that idea in a statement you made earlier in the same post, "Raising ISO without lowering exposure won't over brighten the output image if it's too dark to begin with"

This is exactly the mistake I would catch my students making. Because they incorrectly thought ISO caused noise they would stop raising ISO at their "too-much-noise-limit" and allow the camera JPEG to be too dark. At least they were working with raw files, so they intended to brighten the photo in post and mistakenly concluded they were gaining an advantage. But a dozen years ago out of say 18 students in a class more than half of them came to class with some version of a Canon Rebel T5. T6, etc. and those cameras generated much more read noise than todays cameras and so my students were losing where they thought they were winning because amplification of the sensor's analog signal would suppress that read noise. Raising ISO further would make their photos less noisy. They didn't understand that because they had been misinformed.

Sorry I didn't make that clear enough.

and that coincides with your premise on shot noise. If that's the case then it's something you didn't make clear in your post.

 

[smoke665]

make sure the exposure I used would not ISO-clip the highlights in the ISO 3200 photo and then I processed the two raw files to the same output brightness. The two camera JPEGs from the same identical underexposure are below.

Yup this wasn't real clear in the post, but makes sense once explained. Possibly another source of confusion is the term "exposure", the general photography population will look at the combination of aperture, shutter, and ISO as "exposure". Not arguing the right or wrong of that, just saying that's the way it usually is.

 

[Ysarex]

Yup this wasn't real clear in the post, but makes sense once explained. Possibly another source of confusion is the term "exposure", the general photography population will look at the combination of aperture, shutter, and ISO as "exposure". Not arguing the right or wrong of that, just saying that's the way it usually is.

That came up earlier in the thread talking about the Exposure Triangle. Back in post #18 I said this:
"The exposure triangle as a model makes it's presenters want to balance it and that leads them to misinform.

For example, someone infected with triangulitis will identify the two setting combinations below as producing the same exposure:

A) 1/250th sec, f/5.6, ISO 200
B) 1/250th sec, f/8, ISO 400

In fact the B) settings deliver 1 stop less exposure to the film/sensor than the A) settings."

It's not hard. In fact one reason I really like digital is because it's so easy! It gets confusing when we mix up cause and effect. In one of your earlier posts you referenced Richard Butler's article on noise at DPReview. The first sentence of that article; "How would you react if I told you that the aperture and shutter speed you choose make more difference to image noise than the ISO setting?" There it is.

 

[smoke665]

It's not hard. In fact one reason I really like digital is because it's so easy! It gets confusing when we mix up cause and effect. I

Any ISO noise is mostly a mute point for me as the majority of my shots are in studio with controlled lighting at Base ISO. In field, experience, knowledge of my equipment, knowledge of my abilities to recover post and the end use of the shot, tell me the maximum I can go up on ISO. As you say "cause and effect" may limit me to ISO 400 or lower, on another I may spin the dial. My avatar is an extreme crop from an image shot at ISO 25600. At it's current resolution the noise is not really noticeable, blow it up to an 8x10 and it would be horrible.

 

[Ysarex]

Any ISO noise is mostly a mute point for me as the majority of my shots are in studio with controlled lighting at Base ISO. In field, experience, knowledge of my equipment, knowledge of my abilities to recover post and the end use of the shot, tell me the maximum I can go up on ISO. As you say "cause and effect" may limit me to ISO 400 or lower, on another I may spin the dial. My avatar is an extreme crop from an image shot at ISO 25600. At it's current resolution the noise is not really noticeable, blow it up to an 8x10 and it would be horrible.

Blow this up as much as you want. With an APS sensor shot at ISO 25600 (and without high-end AI noise filtering) the noise is minor. I kept the noise down by addressing the cause of noise -- exposure. ISO-25K.jpg

It's admittedly a parlor trick -- I took it for my students as a demonstration and I exposed as much as the camera would permit without clipping the diffuse highlights in the raw file. Obviously when forced to use an ISO like 25K we're typically not in a position to expose as much as I did. If we could we'd drop the ISO. But as a demo it makes the cause/effect point. The noise level is fixed by the exposure.