ISO RANGE

[SrMi]

It's true that higher ISO settings tend to produce less noise,

Actually, the reverse is true.

He means higher ISO settings with the same shutter speed and aperture. These might have less read noise. I doubt if the difference would be apparent, and this technique requires you to set speed and aperture manually -- by which time the bird has flown.

I think your method is the best.

Don

I am not in a position to discuss ISO at the level discussed here. I know what it does to my pictures and that is sufficient level of knowledge to me.

Besides, it's just a hobby for me

I like hobbies that challenge me and where I can learn new things. However, the old adage rings true: the more you know, the more you know you don't know.

It was simpler when I thought that ISO changes sensor sensitivity . On the other hand, knowledge feels good.

 

[Quarkcharmed]

How does higher ISO limit the exposure? Are you thinking about metering?

In Auto ISO mode, ISO is not really a 'setting', it's an exposure index. So it doesn't limit the exposure. In manual ISO mode, the ISO setting does limit the max exposure you can use without highlight clipping.

I see, but that limiting behavior limits the maximum possible exposure, which would mean that (because of clipping), you will have more noise with high ISO, not less, as you wrote. Correct?

We will have more visible noise (random grain pattern). Also we will have less shot noise because the exposure is lower, and decreased SNR.

What does "has" mean? When I go through my image library, low ISO shots have less noise than high ISO shots.

Here you suggest that low noise depends on ISO, not exposure, which is not true.

No, it's just an observation anyone can make. 'Depends' is a bit tricky here, because, as above, the noise may depend on the ISO setting, if it's manual; however the low noise is caused by higher exposure.

Your observation that low ISO has less noise is caused by the metering that automatically raises ISO when exposure is low.

No, I rarely shoot with auto ISO. I normally do ETTR, and sometimes I use 'ISO invariance' which also implies a fixed ISO setting.

But my observation will be true for anyone's raw photo library regardless of the metering mode and techniques used. So it definitely establishes a correlation between high ISO and high visible noise (before noise reduction).

If lower ISO by itself had less noise, we would often shoot with base ISO to get the best possible IQ, even if it meant pain during the image review and processing.

Let's unchain ISO from exposure by not allowing metering to interfere in the discussion.

As above, of we unchain it, there's still the correlation.

 

[Quarkcharmed]

Exposure does not change, ISO does. We can see very clearly that as the ISO increases, noise decreases.

Here are the same ISOs different exposure



(unfortunately we've added another variable here, EFC, but its effect is minimal). Now compare each patch with the same ISO patch above. It shows that the patch with more exposure is the noisier. So we have two independent variables. ISO, which has quite a small effect on noise, and its noise decreases as ISO increases; and exposure, which has a larger effect, and its noise decreases as exposure increases. So, the rule is, for lowest noise maximise exposure. Set the ISO as high as you can for that exposure.

Why use a Sony example ? Sony apply noise reduction to raw files. Pick a camera that can save true raw files.

Afaik Sony used to apply hot pixels removal, which caused the 'start eater' effect, but not noise reduction.

 

[SrMi]

How does higher ISO limit the exposure? Are you thinking about metering?

In Auto ISO mode, ISO is not really a 'setting', it's an exposure index. So it doesn't limit the exposure. In manual ISO mode, the ISO setting does limit the max exposure you can use without highlight clipping.

I see, but that limiting behavior limits the maximum possible exposure, which would mean that (because of clipping), you will have more noise with high ISO, not less, as you wrote. Correct?

We will have more visible noise (random grain pattern). Also we will have less shot noise because the exposure is lower, and decreased SNR.

What does "has" mean? When I go through my image library, low ISO shots have less noise than high ISO shots.

Here you suggest that low noise depends on ISO, not exposure, which is not true.

No, it's just an observation anyone can make. 'Depends' is a bit tricky here, because, as above, the noise may depend on the ISO setting, if it's manual; however the low noise is caused by higher exposure.

Low noise is always caused by higher exposure (shutter speed and aperture), regardless of mode. When using the camera's automatic modes, the camera uses high ISO to "signal" low exposure and higher noise. The higher the noise, the higher the ISO. So, you could say it the other way around: ISO setting depends on the noise.

Of course, one would not describe it that way to a beginner .

Your observation that low ISO has less noise is caused by the metering that automatically raises ISO when exposure is low.

No, I rarely shoot with auto ISO. I normally do ETTR, and sometimes I use 'ISO invariance' which also implies a fixed ISO setting.

Do you use (full) M-mode? It works best with stationary subjects.

But my observation will be true for anyone's raw photo library regardless of the metering mode and techniques used. So it definitely establishes a correlation between high ISO and high visible noise (before noise reduction).

If lower ISO by itself had less noise, we would often shoot with base ISO to get the best possible IQ, even if it meant pain during the image review and processing.

Let's unchain ISO from exposure by not allowing metering to interfere in the discussion.

As above, of we unchain it, there's still the correlation.

I would love to hear more about that correlation argument.

Assume ISO and exposure are not tied and set completely independently. We would use higher ISO to brighten the image and reduce noise.

BTW, the idea that higher ISO reduces noise is also used in the ETTR approach for ISOs above native ISO: set the best exposure and increase ISO as far as possible. IMO, higher ISOs reduce noise but not enough to warrant the danger of clipping highlights.

 

[Marek M]

It's true that higher ISO settings tend to produce less noise,

Actually, the reverse is true.

He means higher ISO settings with the same shutter speed and aperture. These might have less read noise. I doubt if the difference would be apparent, and this technique requires you to set speed and aperture manually -- by which time the bird has flown.

I think your method is the best.

Don

I am not in a position to discuss ISO at the level discussed here. I know what it does to my pictures and that is sufficient level of knowledge to me.

Besides, it's just a hobby for me

I like hobbies that challenge me and where I can learn new things. However, the old adage rings true: the more you know, the more you know you don't know.

It was simpler when I thought that ISO changes sensor sensitivity . On the other hand, knowledge feels good.

For all intents and purposes, that is enough for me to know, that ISO is like film sensitivity.

Technical background muddles this, not giving me any practical knowledge. If that is too simple for others, I am perfectly fine with it.

Knowledge is good, avalanche of unnecessary information, not so much.

 

[D Cox]

Neither do I. Pictures lack consistency.

That's because you aren't using it properly. If you have a method of exposure setting with manual ISO, then there is an analogue for auto ISO that will achieve exactly the same results, but you need to think it through. In general though, it pairs most easily with a raw workflow.

I think I'll stay where I am. And I do RAW only.

Well, of course it's your affair how you work, but it would be interesting to discuss the reasoning behind it.

I think you overestimate me. I use simple processes, some of them are simply force of habit. For example, I always shoot RAW, I never use Auto ISO, I never use Spot Metering, 99% of time it is Aperture priority, simple things like that. I would not even attempt discussing them, that's the way I like them. You may call me limited, I am perfectly fine with it.

I haven't called you limited, but if it's just 'that's my preference because it suits me' there's not much anyone else can take from that. Their preferences could be different and their's no reason to think that they might be better or worse. So coming back to the thing being discussed here, whether or not auto ISO is a good idea, there's no reason to think that what you choose illuminates the discussion.

Pictures taken at different ISOs are not the same, if there were, there would be no upper limit to ISO.

Maybe we could discuss the ways that you think pictures taken at different ISOs are not the same.

Not sure what there is to discuss. That they are not different? Or, why are they different?

I was thinking about how they are different, which might lead to understanding whether they are significantly different.

I think you may need a better partner for your discussion. I am satisfied with simple: They are different.

The real point I think is that for you 'it's my personal preference' is a satisfactory reason for doing things, but as above it's not something anyone else would be interested in.

It's true that higher ISO settings tend to produce less noise,

Actually, the reverse is true.

Nope. Higher ISO settings in general produce less noise. That's a demonstrable fact. It's why cameras are engineered as they are.

It's demonstrable with manual settings and JPG output. It's not demonstrable with A, S or P modes and raw files. When higher ISO settings lead to less exposure (because you want a faster shutter speed), then shot noise will increase. The only thing a camera can do to reduce shot noise is to apply (crude) noise reduction to JPG files.

(I hear rumours that one company applies noise reduction to "raw" files. If so, they are not raw files.)

It is demonstrable with raw files: shoot images with the same exposure and varying ISOs.

A photographer that wants the best IQ should not be at the mercy of the camera's automatic behavior. That is why Auto-ISO and M mode can often yield the best IQ: you control the exposure with the M setting, and Auto ISO + EC is used to protect highlights or reduce noise by lowering or raising the ISO.

Don

So why be at the mercy of Auto ISO ?

I agree with your instinctive distrust of Auto anything. However, experience of a particular camera tells us what is safe in practice.

Don

 

[D Cox]

It's true that higher ISO settings tend to produce less noise,

Actually, the reverse is true.

He means higher ISO settings with the same shutter speed and aperture. These might have less read noise. I doubt if the difference would be apparent, and this technique requires you to set speed and aperture manually -- by which time the bird has flown.

I think your method is the best.

Don

I am not in a position to discuss ISO at the level discussed here. I know what it does to my pictures and that is sufficient level of knowledge to me.

Besides, it's just a hobby for me

I like hobbies that challenge me and where I can learn new things. However, the old adage rings true: the more you know, the more you know you don't know.

It was simpler when I thought that ISO changes sensor sensitivity . On the other hand, knowledge feels good.

Only if you can show it off in a forum. :-D

Don

 

[SrMi]

It's true that higher ISO settings tend to produce less noise,

Actually, the reverse is true.

He means higher ISO settings with the same shutter speed and aperture. These might have less read noise. I doubt if the difference would be apparent, and this technique requires you to set speed and aperture manually -- by which time the bird has flown.

I think your method is the best.

Don

I am not in a position to discuss ISO at the level discussed here. I know what it does to my pictures and that is sufficient level of knowledge to me.

Besides, it's just a hobby for me

I like hobbies that challenge me and where I can learn new things. However, the old adage rings true: the more you know, the more you know you don't know.

It was simpler when I thought that ISO changes sensor sensitivity . On the other hand, knowledge feels good.

For all intents and purposes, that is enough for me to know, that ISO is like film sensitivity.

Technical background muddles this, not giving me any practical knowledge. If that is too simple for others, I am perfectly fine with it.

Knowledge is good, avalanche of unnecessary information, not so much.

That knowledge allows me to create technically much better images than before. I wish I had known it when I started photographing seriously with digital cameras.

 

[Quarkcharmed]

How does higher ISO limit the exposure? Are you thinking about metering?

In Auto ISO mode, ISO is not really a 'setting', it's an exposure index. So it doesn't limit the exposure. In manual ISO mode, the ISO setting does limit the max exposure you can use without highlight clipping.

I see, but that limiting behavior limits the maximum possible exposure, which would mean that (because of clipping), you will have more noise with high ISO, not less, as you wrote. Correct?

We will have more visible noise (random grain pattern). Also we will have less shot noise because the exposure is lower, and decreased SNR.

What does "has" mean? When I go through my image library, low ISO shots have less noise than high ISO shots.

Here you suggest that low noise depends on ISO, not exposure, which is not true.

No, it's just an observation anyone can make. 'Depends' is a bit tricky here, because, as above, the noise may depend on the ISO setting, if it's manual; however the low noise is caused by higher exposure.

Low noise is always caused by higher exposure (shutter speed and aperture), regardless of mode.

Low visible noise I should have said.

When using the camera's automatic modes, the camera uses high ISO to "signal" low exposure and higher noise. The higher the noise, the higher the ISO. So, you could say it the other way around: ISO setting depends on the noise.

It depends on the exposure, if ISO is on auto.

Of course, one would not describe it that way to a beginner .

Your observation that low ISO has less noise is caused by the metering that automatically raises ISO when exposure is low.

No, I rarely shoot with auto ISO. I normally do ETTR, and sometimes I use 'ISO invariance' which also implies a fixed ISO setting.

Do you use (full) M-mode? It works best with stationary subjects.

Not necessarily stationary, but ETTR is awkward when you shoot action.

But my observation will be true for anyone's raw photo library regardless of the metering mode and techniques used. So it definitely establishes a correlation between high ISO and high visible noise (before noise reduction).

If lower ISO by itself had less noise, we would often shoot with base ISO to get the best possible IQ, even if it meant pain during the image review and processing.

Let's unchain ISO from exposure by not allowing metering to interfere in the discussion.

As above, of we unchain it, there's still the correlation.

I would love to hear more about that correlation argument.

You have two variables: visible noise and ISO setting. They definitely correlate, go through your photo library and check the images before noise reduction.

Assume ISO and exposure are not tied and set completely independently. We would use higher ISO to brighten the image and reduce noise.

With visible noise and manual ISO control, as you increase the ISO, you brighten the image and magnify the visible noise.

BTW, the idea that higher ISO reduces noise is also used in the ETTR approach for ISOs above native ISO: set the best exposure and increase ISO as far as possible.

It's a good technique although I wouldn't call it ETTR, it's rather ITTR (ISO to the right).

IMO, higher ISOs reduce noise but not enough to warrant the danger of clipping highlights.

Higher ISO improves SNR in raw data, not visible noise. It happens because the read noise in absolute values stays the same while the signal magnifies. When you increase ISO and it doesn't improve SNR, you're in the ISO-invariant range and further increase doesn't make sense of you shoot raw. To save the highlights it's best to keep the ISO as low as possible (but adhering to other constraints).

--
https://www.instagram.com/quarkcharmed/
https://500px.com/quarkcharmed

 

[SrMi]

Neither do I. Pictures lack consistency.

That's because you aren't using it properly. If you have a method of exposure setting with manual ISO, then there is an analogue for auto ISO that will achieve exactly the same results, but you need to think it through. In general though, it pairs most easily with a raw workflow.

I think I'll stay where I am. And I do RAW only.

Well, of course it's your affair how you work, but it would be interesting to discuss the reasoning behind it.

I think you overestimate me. I use simple processes, some of them are simply force of habit. For example, I always shoot RAW, I never use Auto ISO, I never use Spot Metering, 99% of time it is Aperture priority, simple things like that. I would not even attempt discussing them, that's the way I like them. You may call me limited, I am perfectly fine with it.

I haven't called you limited, but if it's just 'that's my preference because it suits me' there's not much anyone else can take from that. Their preferences could be different and their's no reason to think that they might be better or worse. So coming back to the thing being discussed here, whether or not auto ISO is a good idea, there's no reason to think that what you choose illuminates the discussion.

Pictures taken at different ISOs are not the same, if there were, there would be no upper limit to ISO.

Maybe we could discuss the ways that you think pictures taken at different ISOs are not the same.

Not sure what there is to discuss. That they are not different? Or, why are they different?

I was thinking about how they are different, which might lead to understanding whether they are significantly different.

I think you may need a better partner for your discussion. I am satisfied with simple: They are different.

The real point I think is that for you 'it's my personal preference' is a satisfactory reason for doing things, but as above it's not something anyone else would be interested in.

It's true that higher ISO settings tend to produce less noise,

Actually, the reverse is true.

Nope. Higher ISO settings in general produce less noise. That's a demonstrable fact. It's why cameras are engineered as they are.

It's demonstrable with manual settings and JPG output. It's not demonstrable with A, S or P modes and raw files. When higher ISO settings lead to less exposure (because you want a faster shutter speed), then shot noise will increase. The only thing a camera can do to reduce shot noise is to apply (crude) noise reduction to JPG files.

(I hear rumours that one company applies noise reduction to "raw" files. If so, they are not raw files.)

It is demonstrable with raw files: shoot images with the same exposure and varying ISOs.

A photographer that wants the best IQ should not be at the mercy of the camera's automatic behavior. That is why Auto-ISO and M mode can often yield the best IQ: you control the exposure with the M setting, and Auto ISO + EC is used to protect highlights or reduce noise by lowering or raising the ISO.

Don

So why be at the mercy of Auto ISO ?

I agree with your instinctive distrust of Auto anything. However, experience of a particular camera tells us what is safe in practice.

Don

You are much more in control of your camera in the Misomatic mode than in A, S, or P mode. With Misomatic mode (M-mode with Auto-ISO), you control the two most significant settings: aperture and shutter speed. Auto-ISO helps keep the image bright enough in the viewfinder and image review; it is less relevant. In A or S mode, you are at the mercy of the camera's metering system to determine either aperture or shutter speed. You "fight" with the camera in those modes by manually changing the ISO.

Of course, M-mode gives you the most control, but it is not always practical.