Exposure triangle explanation please.

[AnthonyL]

Lots of references to the exposure triangle (ISO/Shutter Speed/Aperture). Now, I am comfortable with the relationship between the 3 elements, but has anybody got a good explanation of how to use the triangle to define the relationship and how to use the triangle?

 

[texinwien]

TTMartin wrote:

texinwien wrote:

TTMartin wrote:

texinwien wrote:

clack wrote:

texinwien wrote:

For starters, ISO is not a variable in the exposure equation. The three variables in the exposure equation are:

1. Scene Illuminance
2. F-stop (actually T-stop, to be 100% correct)
3. Shutter Speed

APEX camera exposure equation:

A: relative aperture (f-number)
T: shutter speed in seconds
B: average scene luminance ("brightness")
Sx: ASA arithmetic film speed (ISO goes here)
K: light meter calibration constant

K is a constant, B is not part of the camera settings.
This leaves the camera operator - or, camera control chip - with A, T and Sx.

Voila, the exposure triangle is part of 'the' exposure equation...

That's the wrong formula. It's an artificial formula built specifically to tell the photographer how to adjust the exposure values (aperture, shutter speed, scene luminance) to match the ASA / ISO.

Exposure (photography)

In photography, exposure is the amount of light allowed to fall on each area unit of a photographic medium (photographic film or image sensor) during the process of taking a photograph.

Note: Nothing about ASA or ISO above, and neither ASA nor ISO have any effect on the amount of light allowed to fall on each area unit of a photographic medium.

Exposure is measured in lux seconds, and can be computed from exposure value (EV) and scene luminance in a specified region.

What's EV, and how do we calculate it? Let's see:

Exposure Value

EV corresponds simply to a combination of a shutter speed and an aperture setting, independent of any ISO setting—independent even of whether there is film in the camera or any light available.

...

So, Exposure can be computed from scene luminance and EV (which, in turn is computed from the shutter speed and aperture setting). This leads us back to my previous post - the only 3 variables involved in the equation used to determine photometric exposure are scene luminance, shutter speed and aperture setting (or t-stop to be extra technical).

From the Exposure article linked above, this is the equation (photometric / luminous exposure) you need:

Hv is the exposure, Ev is the image plane illuminance (determined by the scene luminance and aperture setting) and t is time (aka the shutter speed).

et voila, APEX isn't the equation used to determine photometric exposure. It's simply a formula that photographers can use to adjust the three variables that make up the photometric exposure equation - shutter speed, aperture and scene luminance (where possible) - based on a given ASA or ISO.

Interesting how that Wiki article has been manipulated through the years, to imply that photographic exposure has nothing to do with the photographic medium. It has been changed from a definition of photographic exposure. To simply one of light exposure, which is incorrect.

That is the problem with Wiki a group of individuals can change it to suit their opinion regardless of fact.

What Wikipedia use to say about Exposure (Photography):

From Wikipedia, the free encyclopedia

In photography, the total amount of light allowed to fall on the film during the process of taking a photograph.

The correct exposure for a photograph is determined by the sensitivity of the film used. Film sensitivity is referred to as 'speed' and is measured as an ISO rating. Faster film requires less exposure and has a higher ISO rating.

Exposure is controlled in a camera by shutter speed and lens aperture. Longer shutter speeds and greater lens apertures produce greater exposures.

Exposure is measured in 'ev' with higher values denoting more light.

For example, an approximately correct exposure will be obtained on a sunny day using ISO 100 film, an aperture of f11 and a shutter speed of 1/100th of a second.

The Zone System is an excellent method of determining exposure.

Here's the thing - nothing you posted here contradicts what I have said. And I quote:

• Exposure is controlled in a camera by shutter speed and lens aperture. (Note: it is not controlled via ISO or ASA)
• The correct exposure for a photograph is determined by the sensitivity of the film used. (Note: read carefully - a film's sensitivity doesn't change or affect the exposure in any way, it just tells us what exposure [calculated solely based on shutter speed, aperture & scene luminance] is correct for that particular film)

And here's the big problem - ASA was useful with film, and ISO was somewhat useful at the dawn of digital photography, but, with increasing frequency, cameras are "ISO-less". In an age of mostly ISO-less sensors, ISO no longer has a place at the table.

I predict that ISO will mostly go away within a decade. For some of us, it already has. I, for instance, have a mostly ISO-less camera. I shoot RAW and almost always at base ISO, then apply the appropriate amount of gain in my post-processing step.

On an ISO-less sensor, that's all the labeled ISO does - it provides instructions that tell the JPEG engine (either the camera's or the third-party image processor) how much gain to apply to the captured image. The days of ISO being a useful concept are numbered. It'll eventually disappear, and with it the APEX equation, but the photometric / luminous exposure equation will stay with us.

All of the variables in that equation are real, as opposed to the APEX equation, with its artificial ASA / ISO variable.

Your forcing your photographic technique on the definition of photographic exposure.

No, the definition of exposure is physics. It is clear and set it in stone. I am adapting my photographic technique to it.

You camera isn't ISOless you just don't use other ISOs.

My camera is (almost) ISOless and I also don't use other ISOs.

A camera where you can't change the ISO isn't ISOless, the photographic medium still has a sensitivity to light, it just may not vary.

I can change the ISO on my camera, but it has no effect on the sensitivity of the photographic medium - that's what ISOless means.

So, if changing the ISO doesn't change the sensitivity of the photographic medium, of what use is ISO?

Without a photographic medium you can't have an exposure.

If the exposure medium stays the exact same and has the exact same sensitivity no matter which ISO you choose, of what use is the concept of ISO?

Your definition of exposure works for a film camera with no film in it.

It works for all cameras - for film cameras with film in them and for working digital cameras.

All of the things you say are photographic exposure can be done, shutter speed and aperture, etc. And yes, the empty back of the camera was exposed to light, but, there was NO PHOTOGRAPHIC EXPOSURE.

Wrong - I just have an exposure on a highly insensitive medium

You can not have photographic exposure void of a photographic medium.

This is just wrong, and really, it's beside the point. See again the article on exposure and the equation that determines photometric exposure. All that is required is a plane on which the light will fall / be collected. Whether that plane is at all photosensitive is beside the point.

Again, I invite you to read the very informative article on Exposure vs. Brightening . You're laboring under a number of misconceptions that actually serve to muddy the waters and confuse the issue than they do to help.

Take heart, however - there was also a time when I labored under the same misconceptions. I also thought the Exposure Triangle was a real thing that had value. It took some reading, learning, thinking and discussing with intelligent people to put those confused notions to rest, but it was worth it, for me

 

[TTMartin]

texinwien wrote:

TTMartin wrote:

Your definition of exposure works for a film camera with no film in it.

It works for all cameras - for film cameras with film in them and for working digital cameras.

All of the things you say are photographic exposure can be done, shutter speed and aperture, etc. And yes, the empty back of the camera was exposed to light, but, there was NO PHOTOGRAPHIC EXPOSURE.

Wrong - I just have an exposure on a highly insensitive medium

In other words one with an extremely low ISO

You can not have photographic exposure void of a photographic medium.

This is just wrong, and really, it's beside the point. See again the article on exposure and the equation that determines photometric exposure. All that is required is a plane on which the light will fall / be collected. Whether that plane is at all photosensitive is beside the point.

Whether the plane is photosensitive is the entire point of PHOTOGRAPHIC EXPOSURE.

Again, you have defined exposure, just NOT photographic exposure.

 

[texinwien]

TTMartin wrote:

texinwien wrote:

TTMartin wrote:

Your definition of exposure works for a film camera with no film in it.

It works for all cameras - for film cameras with film in them and for working digital cameras.

All of the things you say are photographic exposure can be done, shutter speed and aperture, etc. And yes, the empty back of the camera was exposed to light, but, there was NO PHOTOGRAPHIC EXPOSURE.

Wrong - I just have an exposure on a highly insensitive medium

In other words one with an extremely low ISO

I'd call it an ISOless medium with a very low (and unchangeable) sensitivity. As a thought exercise, though just about every medium has some sensitivity to light.

If I ramp up the exposure by, say, using a magnifying glass to concentrate sun rays onto a small spot on a piece of ordinary wood, I can leave a mark on that wood. In a sense, the wood is photosensitive, although nowhere near as photosensitive as regular photographic film.

Now, if I put the same wood in my film camera and activate the shutter, it's highly unlikely that the wood will be visibly changed. The exposure took place, but the medium was so insensitive that no information was recorded.

You can not have photographic exposure void of a photographic medium.

This is just wrong, and really, it's beside the point. See again the article on exposure and the equation that determines photometric exposure. All that is required is a plane on which the light will fall / be collected. Whether that plane is at all photosensitive is beside the point.

Whether the plane is photosensitive is the entire point of PHOTOGRAPHIC EXPOSURE.

Practically all planes are photosensitive. Some are just more sensitive than others.

Again, you have defined exposure, just NOT photographic exposure.

There is only exposure. There is no special photographic exposure. There is light, there is a photosensitive plane on which it falls. We just have to make sure that we adjust the exposure parameters (shutter speed, aperture and scene luminance) such that the medium is properly exposed, given its photosensitivity.

And again, if changing the ISO setting on your camera doesn't affect the sensitivity of the light-gathering medium (the sensor), of what use is ISO as a concept? If ISO doesn't have any relation to the sensitivity of the medium, what good is it to us?

 

[Jeff]

There certainly is an elegance to the Hv = Ev * t formulation, but I don't think we'll be seeing an iso-less world for a very long time.

First off, few shooters want to wait until PP to set image brightness. Most want an image now, out of the camera and straight to facebook. So somewhere the captured light has to be translated into an image. The camera might set the gain automatically assuming the user wants a 'middle grey', or the user might do it him or herself. But either way, there's a gain knob that has to be calibrated and set. Historically the gain knob is calibrated in ISO units. You prefer to do it PP, which is fine and actually quite logical, but you're still turning a knob, too. Call it whatever you like, it's just a proxy for ISO.

Second, not all cameras are created equal. I want to be able to compare cameras based on their ability to light into images as measured under controlled conditions. That control is the gain. It's got to set somehow, why not ISO?


Somewhere in the chain you need to establish the relationship between Hv hitting the sensor and the image brightness. Historically that's been measured as ISO. Sure, it's an historical artifact from the world of film, but it does the job and everyone 'gets it'. It's not going away anytime soon.

 

[Jeff]

texinwien wrote:

TTMartin wrote:

texinwien wrote:

TTMartin wrote:

Your definition of exposure works for a film camera with no film in it.

It works for all cameras - for film cameras with film in them and for working digital cameras.

All of the things you say are photographic exposure can be done, shutter speed and aperture, etc. And yes, the empty back of the camera was exposed to light, but, there was NO PHOTOGRAPHIC EXPOSURE.

Wrong - I just have an exposure on a highly insensitive medium

In other words one with an extremely low ISO

I'd call it an ISOless medium with a very low (and unchangeable) sensitivity. As a thought exercise, though just about every medium has some sensitivity to light.

If I ramp up the exposure by, say, using a magnifying glass to concentrate sun rays onto a small spot on a piece of ordinary wood, I can leave a mark on that wood. In a sense, the wood is photosensitive, although nowhere near as photosensitive as regular photographic film.

Now, if I put the same wood in my film camera and activate the shutter, it's highly unlikely that the wood will be visibly changed. The exposure took place, but the medium was so insensitive that no information was recorded.

You can not have photographic exposure void of a photographic medium.

This is just wrong, and really, it's beside the point. See again the article on exposure and the equation that determines photometric exposure. All that is required is a plane on which the light will fall / be collected. Whether that plane is at all photosensitive is beside the point.

Whether the plane is photosensitive is the entire point of PHOTOGRAPHIC EXPOSURE.

Practically all planes are photosensitive. Some are just more sensitive than others.

Again, you have defined exposure, just NOT photographic exposure.

There is only exposure. There is no special photographic exposure. There is light, there is a photosensitive plane on which it falls. We just have to make sure that we adjust the exposure parameters (shutter speed, aperture and scene luminance) such that the medium is properly exposed, given its photosensitivity.

And again, if changing the ISO setting on your camera doesn't affect the sensitivity of the light-gathering medium (the sensor), of what use is ISO as a concept? If ISO doesn't have any relation to the sensitivity of the medium, what good is it to us?

What I care about is the final image. I don't have a photograph until I have a final image.

I need enough exposure to render a usable image on whatever medium I'm working with - whether that's film, wood, halide covered bedsheets (actually saw this done once), or a digital sensor. How much exposure do I need to render a usable image?

For better or worse, ISO is the calibration of choice. What would you propose in its place?

 

[Press Correspondent]

AnthonyL wrote:

Press Correspondent wrote:

So here you go, the answer to your triangle question geometrically is a hyperboloid that looks cool and represents the well known dependencies among the exposure parameters. I hope this helps

That's a lot of work to design a pretty coffee filter

Coffee filter? I thought he made you a hat!

Because it is over your head?

 

[Limburger]

Jeff wrote:

There certainly is an elegance to the Hv = Ev * t formulation, but I don't think we'll be seeing an iso-less world for a very long time.

First off, few shooters want to wait until PP to set image brightness. Most want an image now, out of the camera and straight to facebook. So somewhere the captured light has to be translated into an image. The camera might set the gain automatically assuming the user wants a 'middle grey', or the user might do it him or herself. But either way, there's a gain knob that has to be calibrated and set. Historically the gain knob is calibrated in ISO units. You prefer to do it PP, which is fine and actually quite logical, but you're still turning a knob, too. Call it whatever you like, it's just a proxy for ISO.

Second, not all cameras are created equal. I want to be able to compare cameras based on their ability to light into images as measured under controlled conditions. That control is the gain. It's got to set somehow, why not ISO?

Somewhere in the chain you need to establish the relationship between Hv hitting the sensor and the image brightness. Historically that's been measured as ISO. Sure, it's an historical artifact from the world of film, but it does the job and everyone 'gets it'. It's not going away anytime soon.

 

[Guidenet]

Jeff wrote:

Pyramids, Hats, and Coffee Filters, flux capacitors ... we can call it the pointy head approach to exposure, lol.

Rereading what I wrote above, I realize now that my terminology was a bit confusing. So here's another cut at it, hopefully a bit clearer and using somewhat more standard terminology ...

Use B to refer to the scene brightness. Think of it as the light incident on the scene. Using N for f-number, t for shutter duration, and S for the sensitivity of the camera to light (a function of the sensor, camera electronics, and camera settings), then

B = N^2/(t*S)

The idea here is that to achieve a properly exposed image. A more brightly lit scene (larger B) requires a larger f-number (i.e., smaller f-ratio), a shorter shutter duration, or lower S.

For this to work as an honest engineering equation, B, N, t, and S need consistent units. If you want to use SI units to measure luminance and ISO to measure S, then you'd throw in a conversion factor, but that's not necessary for what we're doing. Just accept B as scene brightness measured in weird but consistent units.

The next step is logarithms. Why logarithms? We talk about logarithms all the time in photography by referring to N, t, or S in terms of equivalent 'f-stops'. That's the same as measuring these quantities in base 2 logarithms.

So fooling around a bit with high school algebra,

log2(B) + log2(S) = log2(N^2) + log2(1/t)

On the left side are scene brightness and ISO. On the right are f-number and shutter speed (1/t being shutter speed, the inverse of shutter duration). Using log2 means that we're measuring these quantities in terms of 'f-stops'.

The right side is the Exposure Value (EV).

The equation is a like a titter-totter. Changes on the left have to be matched by changes in exposure value on the right. One stop more brightness or a one-stop increase in S on the left has to be balanced on the right by a one stop increase in f-number or a one-stop increase in the shutter speed.

The exposure titter-totter

Another way of writing this is

Bv + Sv = Av + Tv

where Bv is the 'scene brightness value', Sv is 'Sensitivity value' of ISO, Av is 'Aperture Value', and Tv is 'Shutter Speed Value'. All measured 'f-stops'.

Creatively, you make the left side larger by using more real light. Or if that's not possible, artificially boost what light you do have by increasing Sv (ISO) in the camera. Unfortunately, Sv comes with noise.

On the right side, you can use that 'light' to increase depth of field by making Av larger, or freezing motion by increasing Tv.

You're free to make creative tradeoffs so long as the titter-totter is kept in balance.

 

[DJF77]

You have all succumbed to a WUM

The o.p understands exposure as he says but can't see the relevance of calling the three elements of exposure - that are all inter-connected a "triangle"

I feel sorry for those of you that actually bothered to right a decent explanation for this guy who is seriously pulling your tail!

 

[Limburger]

DJF77 wrote:

You have all succumbed to a WUM

The o.p understands exposure as he says but can't see the relevance of calling the three elements of exposure - that are all inter-connected a "triangle"

I feel sorry for those of you that actually bothered to right a decent explanation for this guy who is seriously pulling your tail!

Why you think I added cowbell?

 

[texinwien]

Jeff wrote:

There certainly is an elegance to the Hv = Ev * t formulation, but I don't think we'll be seeing an iso-less world for a very long time.

Physics is often quite elegant, and I think (and hope) you're wrong about ISO.

First off, few shooters want to wait until PP to set image brightness. Most want an image now, out of the camera and straight to facebook. So somewhere the captured light has to be translated into an image. The camera might set the gain automatically assuming the user wants a 'middle grey', or the user might do it him or herself. But either way, there's a gain knob that has to be calibrated and set. Historically the gain knob is calibrated in ISO units. You prefer to do it PP, which is fine and actually quite logical, but you're still turning a knob, too. Call it whatever you like, it's just a proxy for ISO.

Actually, I'd say ISO is a proxy for gain, and I think it's a confusing proxy for most beginners, and even for many advanced photographers (including a few professionals) who think that adjusting ISO has some effect on their camera's photosensitivity.

The point is that we are moving from ISOful sensors to ISOless sensors. When all sensors are ISOless, I think we should start talking about the saturation sensitivity of a camera's sensor and the gain we wish to apply to a specific exposure. Just call it gain or brightening. Allow the user to set it before they record the photo (or have it set automatically by the camera) and allow the user to adjust it (in-camera) after the exposure has been recorded.

The main point, for me, is that the topic of ISO always seems to be surrounded by a good deal of confusion, including this infernal exposure triangle. Get rid of it and replace it with an easier-to-understand and physically correct concept rather than the flawed conceit of ISO.

See here for more discussion of the matter. There have been other discussions here about how to relabel and redesign the 'gain knob' on ISOless cameras. I think it'll happen, one of these days.

Second, not all cameras are created equal. I want to be able to compare cameras based on their ability to light into images as measured under controlled conditions. That control is the gain. It's got to set somehow, why not ISO?

Not ISO because too many people are confused about what it means. Gain or some other term that is accurate and easier to understand, that doesn't confuse people into thinking that adjusting it somehow magically makes their camera's sensor more sensitive to light.

Somewhere in the chain you need to establish the relationship between Hv hitting the sensor and the image brightness. Historically that's been measured as ISO. Sure, it's an historical artifact from the world of film, but it does the job and everyone 'gets it'.

That's the point, time and again in the discussions I see here, people don't "get it." Far too many people don't "get it." The "Exposure Triangle" doesn't "get it." We have a broken and confusing model based on historical facts that are increasingly irrelevant. Just look at some of the earlier discussion in this very thread for numerous examples of people who don't "get it" and who are confused by the ISO concept.

It's not going away anytime soon.

It's going away technically, and has been for some time. Once it's technically gone, I believe it will eventually be replaced with a simpler terminology and concept, removing it from the "exposure equation" and pushing over into the "image brightening equation", where it belongs.

It's not a part of exposure, and forcing it into discussions about exposure causes way too much confusion, in my experience.

--
Jeff
http://www.flickr.com/photos/jck_photos/sets/
You don't make a photograph just with a camera. You bring to the act of photography all the pictures you have seen, the books you have read, the music you have heard, the people you have loved.” -- Ansel Adams

 

[Limburger]

texinwien wrote:

Jeff wrote:

There certainly is an elegance to the Hv = Ev * t formulation, but I don't think we'll be seeing an iso-less world for a very long time.

Physics is often quite elegant, and I think (and hope) you're wrong about ISO.

First off, few shooters want to wait until PP to set image brightness. Most want an image now, out of the camera and straight to facebook. So somewhere the captured light has to be translated into an image. The camera might set the gain automatically assuming the user wants a 'middle grey', or the user might do it him or herself. But either way, there's a gain knob that has to be calibrated and set. Historically the gain knob is calibrated in ISO units. You prefer to do it PP, which is fine and actually quite logical, but you're still turning a knob, too. Call it whatever you like, it's just a proxy for ISO.

Actually, I'd say ISO is a proxy for gain, and I think it's a confusing proxy for most beginners, and even for many advanced photographers (including a few professionals) who think that adjusting ISO has some effect on their camera's photosensitivity.

The point is that we are moving from ISOful sensors to ISOless sensors. When all sensors are ISOless, I think we should start talking about the saturation sensitivity of a camera's sensor and the gain we wish to apply to a specific exposure. Just call it gain or brightening. Allow the user to set it before they record the photo (or have it set automatically by the camera) and allow the user to adjust it (in-camera) after the exposure has been recorded.

The main point, for me, is that the topic of ISO always seems to be surrounded by a good deal of confusion, including this infernal exposure triangle. Get rid of it and replace it with an easier-to-understand and physically correct concept rather than the flawed conceit of ISO.

See here for more discussion of the matter. There have been other discussions here about how to relabel and redesign the 'gain knob' on ISOless cameras. I think it'll happen, one of these days.

Second, not all cameras are created equal. I want to be able to compare cameras based on their ability to light into images as measured under controlled conditions. That control is the gain. It's got to set somehow, why not ISO?

Not ISO because too many people are confused about what it means. Gain or some other term that is accurate and easier to understand, that doesn't confuse people into thinking that adjusting it somehow magically makes their camera's sensor more sensitive to light.

Somewhere in the chain you need to establish the relationship between Hv hitting the sensor and the image brightness. Historically that's been measured as ISO. Sure, it's an historical artifact from the world of film, but it does the job and everyone 'gets it'.

That's the point, time and again in the discussions I see here, people don't "get it." Far too many people don't "get it." The "Exposure Triangle" doesn't "get it." We have a broken and confusing model based on historical facts that are increasingly irrelevant. Just look at some of the earlier discussion in this very thread for numerous examples of people who don't "get it" and who are confused by the ISO concept.

It's not going away anytime soon.

It's going away technically, and has been for some time. Once it's technically gone, I believe it will eventually be replaced with a simpler terminology and concept, removing it from the "exposure equation" and pushing over into the "image brightening equation", where it belongs.

It's not a part of exposure, and forcing it into discussions about exposure causes way too much confusion, in my experience.

 

[Guidenet]

I read the article several times and though I respect the writer, I think it is another regurgitated explanation using the Wiki article. I think it analogous with the "The Emperor's New Clothes" tale.

https://en.wikipedia.org/wiki/The_Emperor's_New_Clothes

I think we try to take definitions from physics and try to apply it to a different scenario in photography. The idea that it quacks and has webbed feet makes it a duck, might be closer to the mark. When I use my camera, I can use the control of the ISO dial in such a way as to change the saved RAW file. That's a duck, call it what you will. I can also change EV-L or scene brightness and all else being equal, I also affect the RAW capture. I realize this might be brightness applied after the event, but the overall effect is the same. It quacked again. I'm sure it's a duck.

The idea of ignoring ISO as one of the tools to control exposure seems to be a poor way to both explain it to novice photographers and to understand it in general use as well. Fair enough?

 

[Tan68]

texinwien wrote:

TTMartin wrote:

... In photography, the total amount of light allowed to fall on the film during the process of taking a photograph.

The correct exposure for a photograph is determined by the sensitivity of the film used. Film sensitivity is referred to as 'speed' and is measured as an ISO rating. Faster film requires less exposure and has a higher ISO rating.

Exposure is controlled in a camera by shutter speed and lens aperture. Longer shutter speeds and greater lens apertures produce greater exposures. ...

Here's the thing - nothing you posted here contradicts what I have said. And I quote:

• The correct exposure for a photograph is determined by the sensitivity of the film used. (Note: read carefully - a film's sensitivity doesn't change or affect the exposure in any way, it just tells us what exposure [calculated solely based on shutter speed, aperture & scene luminance] is correct for that particular film)

...

What TTMartin wrote was a little different from how you presented your argument. You accepted TTMartin's point and added your caveat as a note to remember sensitivity doesn't affect exposure in any way but it does tell us the correct exposure... OK.

In this way, ISO/film speed/sensitivity indirectly affects exposure. We have to cater to the film loaded and it determines the amount of light we can throw at it and still have a correct exposure for the image. Call sensitivity a second order effect.

You seem to be Wiki Table 1-centric and TTMartin seems to be Wiki Table 2-centric.

At the link you provided is Table 2 for "Exposure values (ISO 100 speed) for various lighting conditions". This is what is useful to me and, I suspect, TTMartin. Table 1 is just a look-up table for settings. Without ISO or whatever name it shall have as specified in Table 2, the Table 1 has little use.

What should I do with Table 1 alone... I can use EV results from a light meter and look up exposure settings. But I have to input ISO to the light meter... I can read a picture was taken at EV 0 and look up settings. I can see equivalent settings... But I have no idea how bright the environment was until I know how much gain was needed. Not until I know the ISO. Table 1 alone is adrift until it is nailed down with the sensitivity of your medium...

For a photographic image, it isn't much use to describe light without taking into account either the sensitivity or, at least, the gain necessary to present a properly exposed image.

And here's the big problem - ASA was useful with film, and ISO was somewhat useful at the dawn of digital photography, but, with increasing frequency, cameras are "ISO-less". In an age of mostly ISO-less sensors, ISO no longer has a place at the table.

I understand ISO-less sensors. One of the first thoughts after reading my last paragraph was probably 'there you are bringing ISO into this again'. Something like that.

With your ISO-less camera, I imagine you shoot at base ISO and use the slowest shutter and widest aperture you can for the demands of the scene. You maximize light. That's good.

I imagine your images sometimes look under exposed ('exposed' v. proper brightness). You don't care to use gain in the camera and prefer to use gain in software. I infer this from what you have written; I could be wrong.

But there will always be some gain used at some point. I am not saying you deny this; it is noted in the flow diagram.

If someone must use more gain in PP than you did for a certain scene, they made a poor exposure (meaning physical light) decision. But, for a properly exposed image (meaning for viewing), gain will always be a part. More or less depending upon the exposure to light... More or less depending upon the photographer's exposure (physical again) decision at the point of capture.

My software only allows me to increase exposure/gain by 4 stops. I better take that into account with my exposure. Even without blocked shadows, I sometimes might have trouble raising exposure in software to get a properly exposed/bright image. Yeah, if I have troubles then I made a poor decision setting my exposure/Tv/Av. But my point is I have to keep it in mind. ISO/Gain and its limits is a force to be reckoned with...

If someone has set ISO 6400 in their camera on a sunny day on South Beach (yeah, WTF, right) then they must take this gain into account. It does affect the amount of light they can capture. Odd decision, 6400, but it is interrelated.

I predict that ISO will mostly go away within a decade. For some of us, it already has. I, for instance, have a mostly ISO-less camera. I shoot RAW and almost always at base ISO, then apply the appropriate amount of gain in my post-processing step. ...

I don't think the concept will go away. Let's stop calling it ISO and I don't care. There will always be some need to describe the amount of gain applied. With film, we needed to know the sensitivity of the film. With digital, we need to know the gain applied.

This talk of exposure excluding ISO/gain meaning physical light versus exposure including gain meaning 'proper' exposure... It makes me think of debates about APS-C v. FF equivalency. Some like to ignore field of view. I maintain field of view is important because I have a particular framing in mind.

Anyway, I feel like there is a movement afoot to divorce ISO from exposure. In the context of physical light, exposure is separate from ISO. However, only in a strict sense. To prepare an image, ISO by whatever name is interrelated...

 

[Jeff]

texinwien wrote:

Jeff wrote:

There certainly is an elegance to the Hv = Ev * t formulation, but I don't think we'll be seeing an iso-less world for a very long time.

Physics is often quite elegant, and I think (and hope) you're wrong about ISO.

First off, few shooters want to wait until PP to set image brightness. Most want an image now, out of the camera and straight to facebook. So somewhere the captured light has to be translated into an image. The camera might set the gain automatically assuming the user wants a 'middle grey', or the user might do it him or herself. But either way, there's a gain knob that has to be calibrated and set. Historically the gain knob is calibrated in ISO units. You prefer to do it PP, which is fine and actually quite logical, but you're still turning a knob, too. Call it whatever you like, it's just a proxy for ISO.

Actually, I'd say ISO is a proxy for gain, and I think it's a confusing proxy for most beginners, and even for many advanced photographers (including a few professionals) who think that adjusting ISO has some effect on their camera's photosensitivity.

The point is that we are moving from ISOful sensors to ISOless sensors. When all sensors are ISOless, I think we should start talking about the saturation sensitivity of a camera's sensor and the gain we wish to apply to a specific exposure. Just call it gain or brightening. Allow the user to set it before they record the photo (or have it set automatically by the camera) and allow the user to adjust it (in-camera) after the exposure has been recorded.

The main point, for me, is that the topic of ISO always seems to be surrounded by a good deal of confusion, including this infernal exposure triangle. Get rid of it and replace it with an easier-to-understand and physically correct concept rather than the flawed conceit of ISO.

See here for more discussion of the matter. There have been other discussions here about how to relabel and redesign the 'gain knob' on ISOless cameras. I think it'll happen, one of these days.

There is confusion, no doubt. But just replacing ISO with gain is rather minor tweak. It'd be nice, sure. But recasting it all in terms of what's happening at the sensor plane might not be the best way to resolve the confusion for most people.

What I really care about are the image attributes, like image brightness relative to scene brightness, depth of field, motion blur, perspective, dynamic range, color. That we control them in terms ISO, sensor size, saturation, sensor efficiency, f-stop, focal length, etc., are all artifacts of the equipment and medium.

If you're going to restructure the camera interface, why not do it terms of final image attributes. Sort of a stripped down version of the Lightroom Develop panel where I can set brightness, shadows, highlights, color and tint, clarity, etc. Now that would be truly useful, imho.

 

[texinwien]

Limburger wrote:

It's a matter of formulae and definitions. But photographically by whatever ISO will be replaced it will be in relation with shutterspeed and aperture just in the fashion it is now.

You need not know how the internal combustion engine works in order to drive a car.

 

[JTC111]

Press Correspondent wrote:

AnthonyL wrote:

Press Correspondent wrote:

So here you go, the answer to your triangle question geometrically is a hyperboloid that looks cool and represents the well known dependencies among the exposure parameters. I hope this helps

That's a lot of work to design a pretty coffee filter

Coffee filter? I thought he made you a hat!

Because it is over your head?

Says the guy who thought a triangle meant a triangle. ;-)

 

[texinwien]

Jeff wrote:

texinwien wrote:

Jeff wrote:

There certainly is an elegance to the Hv = Ev * t formulation, but I don't think we'll be seeing an iso-less world for a very long time.

Physics is often quite elegant, and I think (and hope) you're wrong about ISO.

First off, few shooters want to wait until PP to set image brightness. Most want an image now, out of the camera and straight to facebook. So somewhere the captured light has to be translated into an image. The camera might set the gain automatically assuming the user wants a 'middle grey', or the user might do it him or herself. But either way, there's a gain knob that has to be calibrated and set. Historically the gain knob is calibrated in ISO units. You prefer to do it PP, which is fine and actually quite logical, but you're still turning a knob, too. Call it whatever you like, it's just a proxy for ISO.

Actually, I'd say ISO is a proxy for gain, and I think it's a confusing proxy for most beginners, and even for many advanced photographers (including a few professionals) who think that adjusting ISO has some effect on their camera's photosensitivity.

The point is that we are moving from ISOful sensors to ISOless sensors. When all sensors are ISOless, I think we should start talking about the saturation sensitivity of a camera's sensor and the gain we wish to apply to a specific exposure. Just call it gain or brightening. Allow the user to set it before they record the photo (or have it set automatically by the camera) and allow the user to adjust it (in-camera) after the exposure has been recorded.

The main point, for me, is that the topic of ISO always seems to be surrounded by a good deal of confusion, including this infernal exposure triangle. Get rid of it and replace it with an easier-to-understand and physically correct concept rather than the flawed conceit of ISO.

See here for more discussion of the matter. There have been other discussions here about how to relabel and redesign the 'gain knob' on ISOless cameras. I think it'll happen, one of these days.

There is confusion, no doubt. But just replacing ISO with gain is rather minor tweak. It'd be nice, sure. But recasting it all in terms of what's happening at the sensor plane might not be the best way to resolve the confusion for most people.

What I really care about are the image attributes, like image brightness relative to scene brightness, depth of field, motion blur, perspective, dynamic range, color. That we control them in terms ISO, sensor size, saturation, sensor efficiency, f-stop, focal length, etc., are all artifacts of the equipment and medium.

One of these things is not like the others. On an ISOless sensor, ISO has no effect on anything but the image brightness relative to scene brightness. That is, it's simply gain, and it's something that you should be able to simply adjust after the fact.

Being able to adjust gain after the fact allows you to focus on the only parameters that really matter, shutter speed, aperture and scene luminance. ISO just muddies the water on an ISOless camera.

If you're going to restructure the camera interface, why not do it terms of final image attributes. Sort of a stripped down version of the Lightroom Develop panel where I can set brightness, shadows, highlights, color and tint, clarity, etc. Now that would be truly useful, imho.

Sure - it's already possible in some cameras. Still, my point is that you shouldn't have to choose your gain BEFORE you take the picture. It's an extra, unnecessary variable that you shouldn't have to keep in mind when choosing settings. Given static scene luminance, the only variables you should think about are shutter speed (freeze motion) and aperture (DOF). Why throw the third variable into the mix, when it can just as easily be changed after the fact? Let the camera choose the gain automatically, for all I care, and allow the user to adjust it easily after the fact.

No need to force the user to choose the gain before he takes the picture.

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

 

[Tan68]

TTMartin wrote:

You're forcing your photographic technique on the definition of photographic exposure. You camera isn't ISOless you just don't use other ISOs.

In my previous post, I said I understand 'ISO-less' cameras. It is a confusing term to me and I better say I believe I understand it.

ISO camera = my K20 where it really was better to raise ISO in camera; in some situations, I got lower noise using a higher ISO.

ISO-less camera = K5 where the noise levels are about the same whether use ISO in camera or in PP. I am still not convinced this is true through the entire range... My K20 only showed benefits to 1600 and after that, PP was the same.

So, yeah, they choose to not use other ISO. I still prefer to lift as little as possible in PP. Maybe it makes no diff, but I figure if I have to apply gain somewhere... may as well be in the camera. Anyway.

ADDED: and for anyone wondering why I wouldn't add exposure (Tv/Av-style), I mean a situation where I no longer care to increase physical exposure by lengthening shutter or whatever. The constraints of the image require the depth of field or shutter I have set. So, gain must be applied. Or, ISO increased.

... Without a photographic medium you can't have an exposure.

How can you have any exposure if you don't load your medium ?

Your definition of exposure works for a film camera with no film in it.

mmm... You sound like a true Table 2 guy instead of a Table 1 fellow.

All of the things you say are photographic exposure can be done, shutter speed and aperture, etc. And yes, the empty back of the camera was exposed to light, but, there was NO PHOTOGRAPHIC EXPOSURE. You can not have photographic exposure void of a photographic medium.

Why not ? It's digital :^|

- - - - PS - loved the flux thing

 

[JTC111]

texinwien wrote:
I predict that ISO will mostly go away within a decade. For some of us, it already has. I, for instance, have a mostly ISO-less camera. I shoot RAW and almost always at base ISO, then apply the appropriate amount of gain in my post-processing step.

You're defeating your own argument. A "mostly ISO-less camera" is an ISO camera. You shoot "almost always at base ISO..." Both of those statements indicate that you, like most everyone else that has a good working knowledge of photography, adjusts their ISO when necessary. Like you, I like to keep my ISO at the base number, 100 in my case. I only increase it if the amount of available light is insufficient for me to get the shot I want. I don't know anybody who has a preference for a high ISO setting if a lower one is up to the task at hand.

ISO has been a factor in photography for a long time. The nice thing about ISO on a DSLR is that we no longer have to make a commitment to a particular ISO for 36 shots. But without the ability to adjust ISO, there are many pictures for which we'd have to sacrifice artistic sensibilities or we wouldn't be able to take at all.