Noise and ISO, One more thread

David1961 said:

Chato said:

jrtrent said:

Perhaps each camera is different. With mine, images shot any higher than ISO 400 are going to be unacceptably noisy to me, and images shot at ISO 200 (my camera's lowest setting) have more clipped highlights than those shot at ISO 400, so my camera stays at ISO 400 all the time. In light too dim for handheld exposures at that ISO, I either mount a flash or put the camera away. Following this procedure, noise is never an issue, and I never have to think about ISO.

Click to expand...

I'm sure they are.

But what was beaten into my head was that raising the ISO does NOT Raise the sensitivity of the sensor. What it does is increase the amplification of the photons that the sensor receives.

Click to expand...

That's not quite correct but you are on the right track.

Raising the ISO increases the amplification of the voltage the photons generate on each sensor pixel. Once the shutter closes, the number of photons doesn't change.

There are 2 sources of noise in an image:

1. Shot noise - noise inherent in the light coming from the scene.

2. Read noise -

a) Front End Read Noise - noise introduced by the camera itself; heat, camera electronics prior to the voltage being amplified.

b) Back End Read Noise - additional noise some cameras introduce when the voltage is converted to a digital value.

Shot noise is the result of low levels of light hitting the sensor.

Both the effects of shot noise and read noise are included in the pixel's final digital value.

Now, when you raise the ISO in camera only the shot noise and Front End Read Noise are amplified when the voltage is amplified. The Back End Read Noise, if any, is introduced after the voltage has been amplified and during its conversion to a digital value.

If you raise the "exposure" in a raw converter you are essentially just multiplying the final digital value for each pixel which will include Shot Noise, Front End Read Noise and any Back End Read Noise.

Unless you are fully aware of the camera's noise properties (whether it introduces back end read noise or not), to minimise noise in the final image you are better off raising the ISO in camera (assuming you cannot widen the aperture and/or slow the shutter because of DOF, motion blur constraints) than shooting with a lower ISO in camera and increasing the image lightness in post because raising the ISO results in only the Shot Noise and Front End read Noise being amplified. In PP, raising the "exposure" results in the Shot Noise, Front End Read Noise AND Back End Read Noise being amplified.

Click to expand...

That last sentence is much too long. I had to read it twice.

What you say makes sense. But how are photographers to find out how a particular camera behaves ? The DPR reviews don't test or measure it. Maybe Jim Kasson does, or Bill Claff.

Does the amount of analog gain increase with every step of the "ISO" numbers, or are there only two or three changes across the whole "ISO" range ? Or no changes at all ?

FingerPainter said:

PhotoFactor said:

Great Bustard said:

Xasan said:

FingerPainter said:

You are attributing a misconception that does not exist

Click to expand...

You will need Houdini-level skills to explain the following differently from the misconception Bob describes (taken from https://www.colesclassroom.com/understanding-iso )

The job of ISO is merely to amplify the light signal that the camera receives. In the process, it can produce a similar effect as opening the aperture or using a slow shutter speed.

Click to expand...

Not even remotely true. Tell me, aside from lightness, what do the following photos have in common, and what is different, if photos are taken of the same scene with the same camera and lens:

1. f/2.8 1/200 ISO 100
2. f/5.6 1/200 ISO 400
3. f/2.8 1/800 ISO 400
4. f/5.6 1/800 ISO 1600

In what way is the ISO setting "amplifying the light"? In what way does it "produce a similar effect as opening the aperture or using a slow shutter speed"? Aside, of course, from resulting in photos with the same lightness.

Click to expand...

It say's "amplifying the light signal", not "amplifying the light.

Big difference.

Click to expand...

Also a big difference between a "light signal": and an "electrical signal".

A signal can be sent through any of several different media. A light signal is a signal sent via light. The job of the sensor is to convert a light signal to an electrical signal. The light signal is nor amplified by the application of a higher ISO setting. The electrical signal may be amplified.

Click to expand...

It is perfectly possible to amplify a light signal, for instance with a photomultiplier. But this is never done in cameras.

(It may be done in drum scanners.)

FingerPainter said:

fferreres said:

FingerPainter said:

FingerPainter said:

The job of ISO is merely to amplify the light signal that the camera receives.

Click to expand...

No. The job of the ISO setting is to control how large a digital number is stored in the camera's JPEG output for a given number of photons captured by a sensor element.

Click to expand...

Funny, I move the ISO setting in A-Mode and I see that right after changing the ISO, the shutter also changes.

Click to expand...

That's the autoexposure system at work, not the ISO on its own.

Click to expand...

Setting a bias on the auto exposure system is one of the main effects of raising the ISO number above the usual 100.

You can get the same effect either by adjusting the Exposure Compensation, or by changing the ISO number. The difference is that the "ISO" setting also does other things, not stated in the camera's manual.

FingerPainter said:

FingerPainter said:

I guess In need to go back and take classes with you, should you accept me as your pupil, which is highly improbable.

Click to expand...

Improbable for geographic reasons, perhaps.

Click to expand...

bobn2 said:

fferreres said:

PhotoFactor said:

Great Bustard said:

Xasan said:

bobn2 said:

You are attributing a misconception that does not exist

Click to expand...

You will need Houdini-level skills to explain the following differently from the misconception Bob describes (taken from https://www.colesclassroom.com/understanding-iso )

The job of ISO is merely to amplify the light signal that the camera receives. In the process, it can produce a similar effect as opening the aperture or using a slow shutter speed.

Click to expand...

Not even remotely true. Tell me, aside from lightness, what do the following photos have in common, and what is different, if photos are taken of the same scene with the same camera and lens:

1. f/2.8 1/200 ISO 100
2. f/5.6 1/200 ISO 400
3. f/2.8 1/800 ISO 400
4. f/5.6 1/800 ISO 1600

In what way is the ISO setting "amplifying the light"? In what way does it "produce a similar effect as opening the aperture or using a slow shutter speed"? Aside, of course, from resulting in photos with the same lightness.

Click to expand...

It say's "amplifying the light signal", not "amplifying the light.

Big difference.

Click to expand...

99.9999% (or 99,9999% if not American) of cases the discussion arises as a discussion of what happens where.

In G.B. the only thing common to all photos is the average light intensity we see in the final recorded (or observed) photo. The all seem to have the same amount of light intensity or "lightness" in terms of Church of No Triangle credo.

Click to expand...

'Light intensity' and 'lightness' are completely different things. Lightness is a perceptual measure, it locates a tone on a grey scale running from perceptual 'black' to perceptual 'white'. You can describe paint in terms of lightness, its very hard to do so with light intensity. The misidentification of lightness with light intensity is the heart of the 'light in-light out' fallacy, which Porky believes doesn't exist. The fact that that fallacy is embedded right in the middle of the 'triangle' is one of the many reasons it's a very poor teaching aid, and its use in general results in learners adopting multiple misconceptions. Not their fault - bad teaching.

Click to expand...

Tom Cornsweet's book "Seeing" explains all this, with diagrams.

David1961 said:

Chato said:

jrtrent said:

Perhaps each camera is different. With mine, images shot any higher than ISO 400 are going to be unacceptably noisy to me, and images shot at ISO 200 (my camera's lowest setting) have more clipped highlights than those shot at ISO 400, so my camera stays at ISO 400 all the time. In light too dim for handheld exposures at that ISO, I either mount a flash or put the camera away. Following this procedure, noise is never an issue, and I never have to think about ISO.

Click to expand...

I'm sure they are.

But what was beaten into my head was that raising the ISO does NOT Raise the sensitivity of the sensor. What it does is increase the amplification of the photons that the sensor receives.

Click to expand...

That's not quite correct but you are on the right track.

Raising the ISO increases the amplification of the voltage the photons generate on each sensor pixel. Once the shutter closes, the number of photons doesn't change

Click to expand...

It's important to separate the actual definition of ISO from the underlying implementation in any specific medium. After all, ISO can be used with film, digital sensors, liquid emulsions painted on fabric, even sun acting on human skin.

One way to explain ISO is to consider the difference between human perception and a how a camera records a scene. A camera makes a record of the light falling on the photosensitive medium. It's an absolute amount ... more light, a larger effect. The amount of light recorded on the medium is measured in lux-seconds.

Humans are different, they perceive relative amounts of light. A grey cat on a brown background, for example, looks the same whether in broad daylight or a dim restaurant despite the fact that amount of light is different by a 1000x or more.

The point is that cameras and human perception are different. To go from data recorded by the camera to an image representing what a person perceives, one has to translate from the absolute quantities measured in lux-seconds to the relative scale of human perception.

That's what ISO does. An ISO of 100, for example, means that 0.1 lux-seconds of light recorded on a medium should be interpreted as middle grey. An ISO of 500 means that a record of 0.02 lux-seconds should be interpreted as middle grey. The details of how this is done, whether by analog, digital, or chemical processing, is immaterial to the definition of ISO.

There are many reasons to want to adjust ISO. You might want a larger depth of field so need to use a narrow lens aperture, or you may need a high shutter speed to avoid motion blur, or you may be working in low light. Or you may want to capture more light to avoid shot noise on a digital sensor, or to use a finer grain film. But whatever the reason, ISO is how you relate the absolute amount of light recorded on the medium to the relative scale of human perception.

In short, ISO is a photographic principle that applies to all media. What you're describing above are details of a how a camera performs the operations needed to implement ISO ... a very different thing.

David1961 said:

There are 2 sources of noise in an image:

1. Shot noise - noise inherent in the light coming from the scene.

2. Read noise -

a) Front End Read Noise - noise introduced by the camera itself; heat, camera electronics prior to the voltage being amplified.

b) Back End Read Noise - additional noise some cameras introduce when the voltage is converted to a digital value.

Shot noise is the result of low levels of light hitting the sensor.

Both the effects of shot noise and read noise are included in the pixel's final digital value.

Now, when you raise the ISO in camera only the shot noise and Front End Read Noise are amplified when the voltage is amplified. The Back End Read Noise, if any, is introduced after the voltage has been amplified and during its conversion to a digital value.

If you raise the "exposure" in a raw converter you are essentially just multiplying the final digital value for each pixel which will include Shot Noise, Front End Read Noise and any Back End Read Noise.

Unless you are fully aware of the camera's noise properties (whether it introduces back end read noise or not), to minimise noise in the final image you are better off raising the ISO in camera (assuming you cannot widen the aperture and/or slow the shutter because of DOF, motion blur constraints) than shooting with a lower ISO in camera and increasing the image lightness in post because raising the ISO results in only the Shot Noise and Front End read Noise being amplified. In PP, raising the "exposure" results in the Shot Noise, Front End Read Noise AND Back End Read Noise being amplified.

Click to expand...

--
Jeff
http://www.flickr.com/photos/jck_photos/sets/

No problem Jeff.

My post was in the context of replying to someone else's comment -

"What it does is increase the amplification of the photons that the sensor receives."

D Cox said:

FingerPainter said:

fferreres said:

FingerPainter said:

D Cox said:

The job of ISO is merely to amplify the light signal that the camera receives.

Click to expand...

No. The job of the ISO setting is to control how large a digital number is stored in the camera's JPEG output for a given number of photons captured by a sensor element.

Click to expand...

Funny, I move the ISO setting in A-Mode and I see that right after changing the ISO, the shutter also changes.

Click to expand...

That's the autoexposure system at work, not the ISO on its own.

Click to expand...

Setting a bias on the auto exposure system is one of the main effects of raising the ISO number above the usual 100.

You can get the same effect either by adjusting the Exposure Compensation, or by changing the ISO number. The difference is that the "ISO" setting also does other things, not stated in the camera's manual.

Click to expand...

Yikes ... you're putting old film guys like me at risk of a stroke. Exposure compensation and ISO are different things!

ISO determines the how much light needs to be recorded at the surface of the medium to produce a middle grey tone. Exposure compensation has nothing to do with that ... set ISO and the mapping from recorded light to relative tone is fixed.

Exposure compensation is a setting that tells your exposure meter how to interpret the scene in front of you. If the scene averages to a middle grey tone, then your meter can just use the desired ISO to determine camera settings for the right exposure. If the scene in front of you does not average to middle grey, then you need to tell your exposure meter that fact. A small grey cat on a large white background may require a +1 or +2 on the EC dial. A small grey cat on a large black background may require a -1 or -2 on the EC dial. If EC is set correctly, then the grey cat will appear the same grey tone in the final image.

Now you could try to get the same effect by adjusting ISO (which, for film guys, means change films, or processing, or both) and leaving EC at 0. So for the grey cat on white background, use a use a higher ISO. But the only reason you're doing that is because you've metered your scene incorrectly. You're underexposing the image and making up for that by adjusting ISO higher to fix the problem. Two wrongs don't make a right.

In digital, adjusting for a high key scene by increasing ISO means that you're willing to use less exposure for the grey cat. So you're not using the full headroom of your sensor. That's why it would be better to fix the problem by adjusting EC rather than ISO.

EC and ISO are different things, as any old film guy will tell you. If you've got a high key scene, expose higher. If you have a low key scene, expose less. That way you can process for the same ISO and get a good result.

D Cox said:

D Cox said:

D Cox said:

I guess In need to go back and take classes with you, should you accept me as your pupil, which is highly improbable.

Click to expand...

Improbable for geographic reasons, perhaps.

Click to expand...

Click to expand...

--
Jeff
http://www.flickr.com/photos/jck_photos/sets/

David1961 said:

No problem Jeff.

My post was in the context of replying to someone else's comment -

"What it does is increase the amplification of the photons that the sensor receives."

Click to expand...

Thanks. I have a hard time following who replied to whom regarding what, so I just threw that in there. It was really meant to offer a different perspective on the thread, not a specific response to any one note.

JPEG only shooters have to make desired look straight out of the camera. If you cannot do that by increasing environmental light, decreasing shutter speed, using wider aperture...then do it by increasing ISO.

RAW shooters need to increase exposure until overexposure is over personal tolerances. More exposure is always better because it will decrease noise, but at the same time it shouldn't compromise anything you consider important. Only strict technical rule is that always increase exposure until something is about to get overexposed. There is no point to leave exposure room, it will just increase noise. Well, there is one exception and that is panorama shooting. If you want an uniform look from stitched photos, sometimes some photos are so bright that some other photos are then underexposed to get that uniform look. Maybe there is some post-processing methods to overcome this, but at least it will delay panorama shooting, if you are tweaking perfect exposure for every photo.

For RAW, if you cannot increase exposure to the "next 1/3 EV step means overexposure" level by increasing environmental light, decreasing shutter speed, using wider aperture...then compensate by increasing ISO. Some ISO invarience sensors might work differently, but pretty much every consumer camera benefits, if you increase ISO in those cases above.

Because most camera manufacturers don't care about RAW histograms, search for UniWB custom white balance to overcome that. Then you can optimize true exposure and decide, if higher ISO could help with noise.

Michael Fryd said:

Chato said:

bobn2 said:

Chato said:

jrtrent said:

Perhaps each camera is different. With mine, images shot any higher than ISO 400 are going to be unacceptably noisy to me, and images shot at ISO 200 (my camera's lowest setting) have more clipped highlights than those shot at ISO 400, so my camera stays at ISO 400 all the time. In light too dim for handheld exposures at that ISO, I either mount a flash or put the camera away. Following this procedure, noise is never an issue, and I never have to think about ISO.

Click to expand...

I'm sure they are.

But what was beaten into my head was that raising the ISO does NOT Raise the sensitivity of the sensor. What it does is increase the amplification of the photons that the sensor receives.

Click to expand...

Nope. Photons are quanta, they can't be amplified. It's actually quite an important point, because in the main noise is determined by the number of photons captured. The voltage output of the sensor represents that number of photons. Amplifying it doesn't change the number of photons, it just changes the 'exchange rate'.

The whole identification of ISO with gain is bogus. Unfortunately, it's also a brainworm. Once people adopt the idea, it's really hard to get rid of it.

Click to expand...

At this point I'll be happy to accept any technical explanation. But isn't what you say above, for all practical purposes the same thing?

Click to expand...

It's not uncommon on these forums for someone to say something that seems reasonable on the surface, but is fundamentally wrong "under the hood".

My standard example would be "umbrellas cause rain". If you didn't know how the weather worked, this is actually a reasonable rule of thumb. When you see people walking around with umbrellas, it is either raining, or it is about to. For many common situations, this belief gives you reasonable guidance.

Under the hood, it is wrong, Umbrellas don't cause the rain, however they are correlated to the rain.

Some people say the rule is fine as it gives the right answer

Click to expand...

The analogy does not fully work.

High ISO can be the cause or the consequence of low exposure depending on the mode.

With some modes, you raise ISO and this decreases exposure..

I have no way to open an umbrella which will cause the rain to start.

This is a major difference !

Christof21 said:

Michael Fryd said:

Chato said:

bobn2 said:

Chato said:

jrtrent said:

Perhaps each camera is different. With mine, images shot any higher than ISO 400 are going to be unacceptably noisy to me, and images shot at ISO 200 (my camera's lowest setting) have more clipped highlights than those shot at ISO 400, so my camera stays at ISO 400 all the time. In light too dim for handheld exposures at that ISO, I either mount a flash or put the camera away. Following this procedure, noise is never an issue, and I never have to think about ISO.

Click to expand...

I'm sure they are.

But what was beaten into my head was that raising the ISO does NOT Raise the sensitivity of the sensor. What it does is increase the amplification of the photons that the sensor receives.

Click to expand...

Nope. Photons are quanta, they can't be amplified. It's actually quite an important point, because in the main noise is determined by the number of photons captured. The voltage output of the sensor represents that number of photons. Amplifying it doesn't change the number of photons, it just changes the 'exchange rate'.

The whole identification of ISO with gain is bogus. Unfortunately, it's also a brainworm. Once people adopt the idea, it's really hard to get rid of it.

Click to expand...

At this point I'll be happy to accept any technical explanation. But isn't what you say above, for all practical purposes the same thing?

Click to expand...

It's not uncommon on these forums for someone to say something that seems reasonable on the surface, but is fundamentally wrong "under the hood".

My standard example would be "umbrellas cause rain". If you didn't know how the weather worked, this is actually a reasonable rule of thumb. When you see people walking around with umbrellas, it is either raining, or it is about to. For many common situations, this belief gives you reasonable guidance.

Under the hood, it is wrong, Umbrellas don't cause the rain, however they are correlated to the rain.

Some people say the rule is fine as it gives the right answer

Click to expand...

The analogy does not fully work.

High ISO can be the cause or the consequence of low exposure depending on the mode.

Click to expand...

High ISO is the consequence of processing a low exposure to produce a desired result.

There are many reasons why you might end up with a low exposure. Poor lighting, need for high shutter speed, or small aperture. Whatever. But high ISO is consequence, not a cause.

Member said:

With some modes, you raise ISO and this decreases exposure..

I have no way to open an umbrella which will cause the rain to start.

This is a major difference !

Click to expand...

Christof21 said:

Michael Fryd said:

Chato said:

bobn2 said:

Chato said:

jrtrent said:

....

Click to expand...

Click to expand...

Click to expand...

Click to expand...

Click to expand...

The analogy does not fully work.

High ISO can be the cause or the consequence of low exposure depending on the mode.

With some modes, you raise ISO and this decreases exposure..

I have no way to open an umbrella which will cause the rain to start.

This is a major difference !

Click to expand...

Yes. It's not a perfect analogy, but it is primarily intended to illustrate that correlation is not the same as cause and effect.

Perhaps something along the lines of focal length vs. subject distance. A longer focal length is normally associated with a longer subject distance, however the longer focal length is not always the cause of that longer distance, and sometimes the longer focal length will be used with the same, or a shorter distance.

====

Increasing the ISO setting can be an indirect cause of increased noise. As you mentioned the higher setting may result in a camera in automatic mode choosing a lower exposure. In fact, even if the camera is in full manual mode, a higher ISO setting may influence the photographer to change the settings to get a lower exposure.

We could also say that if a model gets an eye exam prior to the shoot, that can cause more noise. The model's eye's may be dilated from drugs applied during the exam. This makes the model more sensitive to light, and therefore the photographer uses less light to placate the model. Less light results in more noise.

While I would not say that the eye exam was a direct cause of increased noise in the image, I would say that it might be an "indirect" or perhaps a "contributing" factor.

But getting back to photography, we don't want to teach the fallacy that the ISO is a direct cause, and therefore noise always increases if you use a higher ISO setting. That leads people to the incorrect conclusion that they can reduce noise by reducing the ISO setting a few stops, adding a few stops of negative exposure compensation, and then lightening the image a few stops when the process the RAW file.

D Cox said:

bobn2 said:

fferreres said:

PhotoFactor said:

Great Bustard said:

Xasan said:

D Cox said:

You are attributing a misconception that does not exist

Click to expand...

You will need Houdini-level skills to explain the following differently from the misconception Bob describes (taken from https://www.colesclassroom.com/understanding-iso )

The job of ISO is merely to amplify the light signal that the camera receives. In the process, it can produce a similar effect as opening the aperture or using a slow shutter speed.

Click to expand...

Not even remotely true. Tell me, aside from lightness, what do the following photos have in common, and what is different, if photos are taken of the same scene with the same camera and lens:

1. f/2.8 1/200 ISO 100
2. f/5.6 1/200 ISO 400
3. f/2.8 1/800 ISO 400
4. f/5.6 1/800 ISO 1600

In what way is the ISO setting "amplifying the light"? In what way does it "produce a similar effect as opening the aperture or using a slow shutter speed"? Aside, of course, from resulting in photos with the same lightness.

Click to expand...

It say's "amplifying the light signal", not "amplifying the light.

Big difference.

Click to expand...

99.9999% (or 99,9999% if not American) of cases the discussion arises as a discussion of what happens where.

In G.B. the only thing common to all photos is the average light intensity we see in the final recorded (or observed) photo. The all seem to have the same amount of light intensity or "lightness" in terms of Church of No Triangle credo.

Click to expand...

'Light intensity' and 'lightness' are completely different things. Lightness is a perceptual measure, it locates a tone on a grey scale running from perceptual 'black' to perceptual 'white'. You can describe paint in terms of lightness, its very hard to do so with light intensity. The misidentification of lightness with light intensity is the heart of the 'light in-light out' fallacy, which Porky believes doesn't exist. The fact that that fallacy is embedded right in the middle of the 'triangle' is one of the many reasons it's a very poor teaching aid, and its use in general results in learners adopting multiple misconceptions. Not their fault - bad teaching.

Click to expand...

Tom Cornsweet's book "Seeing" explains all this, with diagrams.

Click to expand...

Thanks, I'll check it out.

D Cox said:

In a few cases, increasing the gain of the analog amplifier stage may reduce noise.

Click to expand...

In the majority of cases, that's why they do it.

mamallama said:

bobn2 said:

mamallama said:

PhotoFactor said:

Great Bustard said:

Xasan said:

mamallama said:

You are attributing a misconception that does not exist

Click to expand...

You will need Houdini-level skills to explain the following differently from the misconception Bob describes (taken from https://www.colesclassroom.com/understanding-iso )

The job of ISO is merely to amplify the light signal that the camera receives. In the process, it can produce a similar effect as opening the aperture or using a slow shutter speed.

Click to expand...

Not even remotely true. Tell me, aside from lightness, what do the following photos have in common, and what is different, if photos are taken of the same scene with the same camera and lens:

1. f/2.8 1/200 ISO 100
2. f/5.6 1/200 ISO 400
3. f/2.8 1/800 ISO 400
4. f/5.6 1/800 ISO 1600

In what way is the ISO setting "amplifying the light"? In what way does it "produce a similar effect as opening the aperture or using a slow shutter speed"? Aside, of course, from resulting in photos with the same lightness.

Click to expand...

It say's "amplifying the light signal", not "amplifying the light.

Big difference.

Click to expand...

Best not get into a word war with our forum experts. They will wordsmith you to death. They think only they know the right words for everything.

Click to expand...

The 'war' is one of understanding, not words. The reason for discussing words is that use of words impacts cognition (read a bit of Chomsky), and in particular using different words for different concepts helps the mind delineate those different concepts. Use the same word for different concepts and it's really hard to keep your thought processes straight. In a discussion, it's a good idea if there's some agreement on what the words mean in that context. You have every right to adopt your own meanings, but don't be surprised if it results in talking at cross purposes. In a field such as photography, a formal vocabulary has been established for many of the key words. It's not unreasonable to conduct a discussion on the basis of that vocabulary.

In the case above. PhotoFactor obviously believes that the addition of the word 'signal' makes a big change to the meaning of the sentence. That's fine, but it appears not to be evident to others in the discussion, so If he wants them to take what he says on board he needs now to explain why that word makes a substantive difference.

Click to expand...

If you are trying to make this forum like a technical journal you need to be more careful with your words. In the meantime you are just nitpicking others with a lot of word misuse yourself.

Click to expand...

I'm always happy to receive constructive criticism that helps me improve word misuse. Of course, I'm not trying to make this forum 'like a technical journal'. Any discussion here is far short of the typical requirements for such a journal. All I'm trying to do is help (believe it or not) people sort out the basic concepts of photography, and as I discussed, that means using different words (symbols) to denote different concepts.

As simple example, relevant to this thread. If you think that 'exposure' means both the light energy density at the sensor and how light or dark the output photo is, then it becomes impossible to understand what ISO is, since ISO deals in the relationship of one to the other. So, any discussion of ISO must use two different words to denote the two different concepts. 'Exposure' is a very well established term in photography, and every authoritative reference will tell you that it means the former concept, light energy density at the sensor (they may not use those exact words, but when you work through it, you find that is what they are telling you). So, that fixes the meaning of 'exposure'. Interestingly, it wasn't so long ago that definition would be hotly disputed in these forums, but now it seems that those insisting on other definitions (you were one, as I remember) have wilted under the weight of evidence.

Anyhow, once we have decided that 'exposure' should be the word for one of those concepts, what of the other? Candidates from the vocabulary of photography are 'value' (as in the HSV - Hue, Saturation, Value colour space), 'lightness' (as in the Lab, Lightness, a b, colour space) and 'brightness'. Of the three, I prefer to say 'lightness', and I have thought-out reasons for thinking that's a better word to help people sort out the ideas in their head. I don't think that 'value' is particularly descriptive (though it is the term ISO uses in the standard). 'Brightness' carries the connotation of emission of light, which tends to engender or reinforce the 'light in-light out' fallacy. Thus I think 'lightness' is a better term, even though it is an unfamiliar word to many. Really, it is no more or less jargonised than 'exposure', it's just less familiar. The other advantage of 'lightness' is its place in the Lab colour space, to which most of the operational spaces we use are referred. Thus, if people want to delve further into colour theory, they already have one of the key ideas in place.

I just wish that instead of being in attack mode all the time you'd give me some credit for meaning well and for thinking these things through. I might well have got it wrong, but I'm much more likely to respond to constructive criticism rather than constant mean-spirited attacks.

--
...because you know, sometimes words have two meanings.

Great Bustard said:

PhotoFactor said:

It say's "amplifying the light signal", not "amplifying the light.

Big difference.

Click to expand...

Three questions:

1. What is the "light signal"?
2. How is that signal being "amplified"?
3. In what was does the "amplification" of the "light signal produce a similar effect as opening the aperture or using a slow shutter speed" other than the lightness of the photo?

The answers to these questions are key in making the statements meaningful.

Click to expand...

Hi GB, long time no see, how are you going?

The answers you've been looking for:

1.Light signal is the electrical signal that comes out from the image sensor plate eg. CMOS in response to light. (Note: not talking about entire image sensor circuit board, just the sensor layer)

2.Signal amplified by transistors and other electronics, creating a stronger charge. Video cameras traditionally called this 'Gain'. 'ISO' term is a carry over from film sensitivity.

3.'Amplification' increases the image electrical signal from the sensor layer, ready for digital encoding into bits.

Jeff said:

D Cox said:

FingerPainter said:

fferreres said:

FingerPainter said:

Jeff said:

The job of ISO is merely to amplify the light signal that the camera receives.

Click to expand...

No. The job of the ISO setting is to control how large a digital number is stored in the camera's JPEG output for a given number of photons captured by a sensor element.

Click to expand...

Funny, I move the ISO setting in A-Mode and I see that right after changing the ISO, the shutter also changes.

Click to expand...

That's the autoexposure system at work, not the ISO on its own.

Click to expand...

Setting a bias on the auto exposure system is one of the main effects of raising the ISO number above the usual 100.

You can get the same effect either by adjusting the Exposure Compensation, or by changing the ISO number. The difference is that the "ISO" setting also does other things, not stated in the camera's manual.

Click to expand...

Yikes ... you're putting old film guys like me at risk of a stroke. Exposure compensation and ISO are different things!

Click to expand...

When I shot film, it was ASA, not ISO.

Jeff said:

ISO determines the how much light needs to be recorded at the surface of the medium to produce a middle grey tone. Exposure compensation has nothing to do with that ... set ISO and the mapping from recorded light to relative tone is fixed.

Click to expand...

They have the same effect on the metering system if you are using an auto mode. Note that I put "ISO" in quotes -- that acronym is used on cameras for a control that has little to do with any international standard.

Jeff said:

Exposure compensation is a setting that tells your exposure meter how to interpret the scene in front of you.

Click to expand...

So does the "ISO" setting. The difference is that the "ISO" control also tries to reduce the bad effects of underexposure, in various ways depending on the camera model.

Jeff said:

If the scene averages to a middle grey tone, then your meter can just use the desired ISO to determine camera settings for the right exposure. If the scene in front of you does not average to middle grey, then you need to tell your exposure meter that fact. A small grey cat on a large white background may require a +1 or +2 on the EC dial. A small grey cat on a large black background may require a -1 or -2 on the EC dial. If EC is set correctly, then the grey cat will appear the same grey tone in the final image.

Now you could try to get the same effect by adjusting ISO (which, for film guys, means change films, or processing, or both) and leaving EC at 0. So for the grey cat on white background, use a use a higher ISO. But the only reason you're doing that is because you've metered your scene incorrectly. You're underexposing the image and making up for that by adjusting ISO higher to fix the problem. Two wrongs don't make a right.

In digital, adjusting for a high key scene by increasing ISO means that you're willing to use less exposure for the grey cat. So you're not using the full headroom of your sensor. That's why it would be better to fix the problem by adjusting EC rather than ISO.

Click to expand...

Matrix metering solves all those cat-and-snow problems. Nobody should be using averaging metering today.

Jeff said:

EC and ISO are different things, as any old film guy will tell you. If you've got a high key scene, expose higher. If you have a low key scene, expose less. That way you can process for the same ISO and get a good result.

Click to expand...

If you need a shorter exposure time, give less exposure. In the case of film, you can put a higher ISO film in the camera. In a digital camera, you just increase the ISO setting. Or, you can underexpose with the EC control. The advantage of the ISO setting is that it can bring in any features the camera has for getting acceptable results from underexposed shots.

If you are saving raw files, the end result will not be all that great, depending on the implementation of analog gain.

David1961 said:

bobn2 said:

In a field such as photography, a formal vocabulary has been established for many of the key words. It's not unreasonable to conduct a discussion on the basis of that vocabulary.

Click to expand...

Yes but people also need to balance the vocabulary they use so that it is not too technical for the intended audience. Obviously that is much harder in online forums like this.

Click to expand...

See my discussion in my answer to mamallama. Simply, to discuss ISO we need two different words to denote two different concepts. One is the light energy density at the sensor. The other is how light or dark the output processed image looks. I'm open to suggestions. I don't see that saying that 'ISO defines the relationship between exposure and exposure' is an option.

bobn2 said:

D Cox said:

In a few cases, increasing the gain of the analog amplifier stage may reduce noise.

Click to expand...

In the majority of cases, that's why they do it.

Click to expand...

But they don't always do it. I may be wrong, but my impression is that current Sony cameras don't increase the amplification any more above ISO 640, for instance.

David1961 said:

bobn2 said:

That's fine, but it appears not to be evident to others in the discussion...

Click to expand...

I see your comment as simply your opinion and not an established fact.

Click to expand...

I never tried to present it as anything other than an opinion, that's why I included the word 'seems' in there. In the end, when it comes to whit is and is not evident to other people, all of us are speculating, unless we're solipsists.

David1961 said:

It say's "amplifying the light signal", not "amplifying the light."

My interpretation is that there is a huge difference in the two meanings.

Click to expand...

Sure, but what is that.

David1961 said:

"amplifying the light signal " to me clearly means the signal from the sensor pixel generated by the light.

"amplifying the light." to me clearly means amplifying the light itself.

Click to expand...

Personally, I wouldn't go with that interpretation. The 'signal from the pixel' isn't a 'light signal', it's an electronic signal, the value of which represents an amount of light. By amplifying the electronic signal you change the representation, but not the amount of light. Whilst you might be clear on the difference, from previous conversations I know that many aren't.

The interpretation depends on what you think the phrase 'light signal' means, and that's far from clear. I'd hazard a guess that if you looked in the wayback machine you'd find an earlier version that did say that the light was amplified, and the word 'signal' was added in response to a comment pointing out the error. I don't think it completely corrects it.

--
...because you know, sometimes words have two meanings.