Reading mode: Light Dark

Increasing the ISO is a contributory cause of increased noise

Go to last post
T

Tom Axford

Veteran Member
Messages
11,595
Solutions
57
Reaction score
13,488
Location
Midlands, UK
There is an argument raging in another thread (in the Beginners Questions forum!) about whether or not increasing the ISO causes noise to increase.

As I see it, the disagreement essentially revolves around the meaning of the word "causes".

Some people appear to think that a cause must be a sufficient cause for it to called a cause. Others (including me) think that a contributory cause may also be called a cause.

Philosophers of causality distinguish between necessary causes, sufficient causes and contributory (or partial) causes. A contributory cause is neither necessary nor sufficient. It is sufficient only if other conditions hold as well.

Increasing the ISO is a contributory cause of increased noise because if various other conditions hold (such as the camera being in an auto-exposure mode), then increasing the ISO causes a decrease in exposure which in turn causes fewer photons to be collected by the sensor, which in turn causes an increase in the noise (relative to the signal).

What is your opinion on this matter?
 
Tom Axford said:
There is an argument raging in another thread (in the Beginners Questions forum!) about whether or not increasing the ISO causes noise to increase.

As I see it, the disagreement essentially revolves around the meaning of the word "causes".

Some people appear to think that a cause must be a sufficient cause for it to called a cause. Others (including me) think that a contributory cause may also be called a cause.

Philosophers of causality distinguish between necessary causes, sufficient causes and contributory (or partial) causes. A contributory cause is neither necessary nor sufficient. It is sufficient only if other conditions hold as well.

Increasing the ISO is a contributory cause of increased noise because if various other conditions hold (such as the camera being in an auto-exposure mode), then increasing the ISO causes a decrease in exposure which in turn causes fewer photons to be collected by the sensor, which in turn causes an increase in the noise (relative to the signal).

What is your opinion on this matter?
Click to expand...
I might argue that since read noise usually decreases with increasing ISO setting, increasing ISO decreases noise. This holds when other settings are kept constant until you clip highlights.

And now we got a contradiction since increasing ISO apparently causes increase or decrease of noise.
 
Last edited:
My opinion is the same as yours.

Those who only ever work in the fully manual mode tend to insist that changing the ISO number cannot ever increase noise (because for them, it doesn't change exposure). They may add that it actually decreases noise, because the JPG engine will apply stronger noise reduction at higher ISO settings.

I normally work in A or S mode, and save raw files, so my experience is different from theirs.

I must read up on causation.

Don Cox
 
MCLV said:
Tom Axford said:
There is an argument raging in another thread (in the Beginners Questions forum!) about whether or not increasing the ISO causes noise to increase.

As I see it, the disagreement essentially revolves around the meaning of the word "causes".

Some people appear to think that a cause must be a sufficient cause for it to called a cause. Others (including me) think that a contributory cause may also be called a cause.

Philosophers of causality distinguish between necessary causes, sufficient causes and contributory (or partial) causes. A contributory cause is neither necessary nor sufficient. It is sufficient only if other conditions hold as well.

Increasing the ISO is a contributory cause of increased noise because if various other conditions hold (such as the camera being in an auto-exposure mode), then increasing the ISO causes a decrease in exposure which in turn causes fewer photons to be collected by the sensor, which in turn causes an increase in the noise (relative to the signal).

What is your opinion on this matter?
Click to expand...
I might argue that since read noise usually decreases with increasing ISO setting, increasing ISO decreases noise. This holds when other settings are kept constant until you clip highlights.

And now we got a contradiction since increasing ISO apparently causes increase or decrease of noise.
Click to expand...
No contradiction as logically it is possible for a contributory cause to work in a different way if the associated conditions change. Of course, for the knowledge to be most useful the associated conditions should be stated.
 
The digital sensor chip converts feeble levels of light energy to electrical charges which are weak and thus need amplificon. Each photosite has an amplifier and the expectation is, all will work at the same efficiency. This is not the case. Different efficiencies induce what we call fixed pattern noise. Image noise is the counterpart of static in an audio circuit. In other words, there is always present some bad signal intermixed with the good. The ratio of good to bad is made worse when we up the ISO so we can image under feeble light conditions. We are forced to up the gain of the image sensor by turning up the ISO. Now the ratio of bad signal increases -- thus the introduction of unwanted image artifacts.
 
D Cox said:
My opinion is the same as yours.

Those who only ever work in the fully manual mode tend to insist that changing the ISO number cannot ever increase noise (because for them, it doesn't change exposure). They may add that it actually decreases noise, because the JPG engine will apply stronger noise reduction at higher ISO settings.

I normally work in A or S mode, and save raw files, so my experience is different from theirs.

I must read up on causation.

Don Cox
Click to expand...
Let me propose a car analogy. Does steering wheel cause the car to turn left or right? Or does car turn left or right because of the direction where are the wheels pointed? Steering wheel affects the direction only when the link between it and wheels is present.

In a similar way, ISO causes a change in noise (and now I'm neglecting read noise) when it is interpreted as a command to change exposure time and/or aperture setting.
 
MCLV said:
Tom Axford said:
D Cox said:
I'm keeping out of that thread. It's like the La Brea Tar Pits.

Don
Click to expand...
Very wise!

I pulled out when I found the same things being repeated time and again.
Click to expand...
And creating a second tar pit to drag more people in is a masterpiece ;)
Click to expand...
Well, the hope is to start a sensible discussion about what people mean by "cause". I hope that it doesn't become another tar pit!!!
 
Tom Axford said:
There is an argument raging in another thread (in the Beginners Questions forum!) about whether or not increasing the ISO causes noise to increase.

As I see it, the disagreement essentially revolves around the meaning of the word "causes".

Some people appear to think that a cause must be a sufficient cause for it to called a cause. Others (including me) think that a contributory cause may also be called a cause.

Philosophers of causality distinguish between necessary causes, sufficient causes and contributory (or partial) causes. A contributory cause is neither necessary nor sufficient. It is sufficient only if other conditions hold as well.

Increasing the ISO is a contributory cause of increased noise because if various other conditions hold (such as the camera being in an auto-exposure mode), then increasing the ISO causes a decrease in exposure which in turn causes fewer photons to be collected by the sensor, which in turn causes an increase in the noise (relative to the signal).

What is your opinion on this matter?
Click to expand...
Actually, it causes noise to decrease. It indirectly causes the ratio of noise to signal to increase (or SNR to decrease).

Even this is only true under the condition that image lightness is maintained and integration time is not long enough to create dark noise.
 
57even said:
Tom Axford said:
There is an argument raging in another thread (in the Beginners Questions forum!) about whether or not increasing the ISO causes noise to increase.

As I see it, the disagreement essentially revolves around the meaning of the word "causes".

Some people appear to think that a cause must be a sufficient cause for it to called a cause. Others (including me) think that a contributory cause may also be called a cause.

Philosophers of causality distinguish between necessary causes, sufficient causes and contributory (or partial) causes. A contributory cause is neither necessary nor sufficient. It is sufficient only if other conditions hold as well.

Increasing the ISO is a contributory cause of increased noise because if various other conditions hold (such as the camera being in an auto-exposure mode), then increasing the ISO causes a decrease in exposure which in turn causes fewer photons to be collected by the sensor, which in turn causes an increase in the noise (relative to the signal).

What is your opinion on this matter?
Click to expand...
Actually, it causes noise to decrease. It indirectly causes the ratio of noise to signal to increase (or SNR to decrease).

Even this is only true under the condition that image lightness is maintained and integration time is not long enough to create dark noise.
Click to expand...
Actually, in this case, the (standard deviation of the) noise increases with the ISO, assuming that the higher ISO forces us to lower the exposure.
 
Aw, come on, Tom, don't get into that. We all know the answers here.

1. Increasing the ISO setting usually decreases the read noise , here compared with actual physical units. (It's a mistake to compare read noise in DN units because the numbers are scaled by a different factor at every ISO setting! Furthermore, they may be scaled differently for every camera.)

2.Increasing ISO setting causes the exposure to decrease in automatic exposure modes, and causes the meter to recommends reduced exposure in manual mode.

3. Decreasing exposure decreases shot noise, but causes a noisier image (lower S/N), because of an unfortunate quirk in terminology.
 
Last edited:
ThrillaMozilla said:
Aw, come on, Tom, don't get into that. We all know the answers here.

1. Increasing the ISO setting usually decreases the read noise , here compared with actual physical units. (It's a mistake to compare read noise in DN units because the numbers are scaled by a different factor at every ISO setting! Furthermore, they may be scaled differently for every camera.)

2.Increasing ISO setting causes the exposure to decrease in automatic exposure modes, and causes the meter to recommends reduced exposure in manual mode.

3. Decreasing exposure decreases shot noise, but causes a noisier image (lower S/N), because of an unfortunate quirk in terminology.
Click to expand...
Yes, I agree, people here do know those answers.

I am sorry I wasn't clearer, but my intention was to start a discussion about when it is ok to say that A causes B, about the language we use.

However, almost all the replies seem to ignore that question and want to return to the discussion about ISO and noise that we have had many times before. I am not going to join in to that discussion.
 
Tom Axford said:
ThrillaMozilla said:
Aw, come on, Tom, don't get into that. We all know the answers here.

1. Increasing the ISO setting usually decreases the read noise , here compared with actual physical units. (It's a mistake to compare read noise in DN units because the numbers are scaled by a different factor at every ISO setting! Furthermore, they may be scaled differently for every camera.)

2.Increasing ISO setting causes the exposure to decrease in automatic exposure modes, and causes the meter to recommends reduced exposure in manual mode.

3. Decreasing exposure decreases shot noise, but causes a noisier image (lower S/N), because of an unfortunate quirk in terminology.
Click to expand...
Yes, I agree, people here do know those answers.

I am sorry I wasn't clearer, but my intention was to start a discussion about when it is ok to say that A causes B, about the language we use.

However, almost all the replies seem to ignore that question and want to return to the discussion about ISO and noise that we have had many times before. I am not going to join in to that discussion.
Click to expand...
It seems your question would be better suited for Philosophy and Linguistics Forum. But wait, we don't have that one here :-)
 
MCLV said:
Tom Axford said:
ThrillaMozilla said:
Aw, come on, Tom, don't get into that. We all know the answers here.

1. Increasing the ISO setting usually decreases the read noise , here compared with actual physical units. (It's a mistake to compare read noise in DN units because the numbers are scaled by a different factor at every ISO setting! Furthermore, they may be scaled differently for every camera.)

2.Increasing ISO setting causes the exposure to decrease in automatic exposure modes, and causes the meter to recommends reduced exposure in manual mode.

3. Decreasing exposure decreases shot noise, but causes a noisier image (lower S/N), because of an unfortunate quirk in terminology.
Click to expand...
Yes, I agree, people here do know those answers.

I am sorry I wasn't clearer, but my intention was to start a discussion about when it is ok to say that A causes B, about the language we use.

However, almost all the replies seem to ignore that question and want to return to the discussion about ISO and noise that we have had many times before. I am not going to join in to that discussion.
Click to expand...
It seems your question would be better suited for Philosophy and Linguistics Forum. But wait, we don't have that one here :-)
Click to expand...
Yes, it would!

However, we all use language in describing how things work and the language we use seems to be a major contributory factor in many of the misunderstandings and arguments that occur in these forums.
 
  • Like
Reactions: tko
Tom Axford said:
MCLV said:
Tom Axford said:
ThrillaMozilla said:
Aw, come on, Tom, don't get into that. We all know the answers here.

1. Increasing the ISO setting usually decreases the read noise , here compared with actual physical units. (It's a mistake to compare read noise in DN units because the numbers are scaled by a different factor at every ISO setting! Furthermore, they may be scaled differently for every camera.)

2.Increasing ISO setting causes the exposure to decrease in automatic exposure modes, and causes the meter to recommends reduced exposure in manual mode.

3. Decreasing exposure decreases shot noise, but causes a noisier image (lower S/N), because of an unfortunate quirk in terminology.
Click to expand...
Yes, I agree, people here do know those answers.

I am sorry I wasn't clearer, but my intention was to start a discussion about when it is ok to say that A causes B, about the language we use.

However, almost all the replies seem to ignore that question and want to return to the discussion about ISO and noise that we have had many times before. I am not going to join in to that discussion.
Click to expand...
It seems your question would be better suited for Philosophy and Linguistics Forum. But wait, we don't have that one here :-)
Click to expand...
Yes, it would!

However, we all use language in describing how things work and the language we use seems to be a major contributory factor in many of the misunderstandings and arguments that occur in these forums.
Click to expand...
Based on my experience from work, it is really hard to formulate something in way that is clear, understandable and unambiguous. It takes a lot of effort and time to do so. And it's not so uncommon that even people who have to do it as a part of their job and are payed for it produce something that has one of following faults:
  • It's not understandable by other people
  • It's quite understandable but everyone understands it differently
  • It's understandable and unambiguous but it says something different than author intended
Therefore, I wouldn't have high expectations regarding discussions on an internet forum.
 
Tom Axford said:
ThrillaMozilla said:
Aw, come on, Tom, don't get into that. We all know the answers here.

1. Increasing the ISO setting usually decreases the read noise , here compared with actual physical units. (It's a mistake to compare read noise in DN units because the numbers are scaled by a different factor at every ISO setting! Furthermore, they may be scaled differently for every camera.)

2.Increasing ISO setting causes the exposure to decrease in automatic exposure modes, and causes the meter to recommends reduced exposure in manual mode.

3. Decreasing exposure decreases shot noise, but causes a noisier image (lower S/N), because of an unfortunate quirk in terminology.
Click to expand...
Yes, I agree, people here do know those answers.

I am sorry I wasn't clearer, but my intention was to start a discussion about when it is ok to say that A causes B, about the language we use.

However, almost all the replies seem to ignore that question and want to return to the discussion about ISO and noise that we have had many times before. I am not going to join in to that discussion.
Click to expand...
I think it has to be clear and succinct--even more so than what I wrote. But this is not easy, because many people have little comprehension of mathematics or numbers beyond elementary school. I'm not entirely certain that everyone even on this forum completely understands it.

For most purposes it may be sufficient to say that decreased exposure causes noisier images, and that in automatic exposure modes, if you increase the ISO setting, the camera will decrease the exposure accordingly.

That doesn't teach anyone very much because it doesn't encourage them to explore the limits of that generalization. If more information is called for, it gets a little trickier. I think that over time, many good explanations have been offered.
 
Last edited:
Tom Axford said:
MCLV said:
Tom Axford said:
ThrillaMozilla said:
Aw, come on, Tom, don't get into that. We all know the answers here.

1. Increasing the ISO setting usually decreases the read noise , here compared with actual physical units. (It's a mistake to compare read noise in DN units because the numbers are scaled by a different factor at every ISO setting! Furthermore, they may be scaled differently for every camera.)

2.Increasing ISO setting causes the exposure to decrease in automatic exposure modes, and causes the meter to recommends reduced exposure in manual mode.

3. Decreasing exposure decreases shot noise, but causes a noisier image (lower S/N), because of an unfortunate quirk in terminology.
Click to expand...
Yes, I agree, people here do know those answers.

I am sorry I wasn't clearer, but my intention was to start a discussion about when it is ok to say that A causes B, about the language we use.

However, almost all the replies seem to ignore that question and want to return to the discussion about ISO and noise that we have had many times before. I am not going to join in to that discussion.
Click to expand...
It seems your question would be better suited for Philosophy and Linguistics Forum. But wait, we don't have that one here :-)
Click to expand...
Yes, it would!

However, we all use language in describing how things work and the language we use seems to be a major contributory factor in many of the misunderstandings and arguments that occur in these forums.
Click to expand...
I do not think that this is a linguistic problem. The question is not clear/well posed. As such, it allows for different implicit assumptions leading to different answers.
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top