[MrB1]

From the way that Dave has written his paper it seems clear that John is a beginner. Dave has outlined the basics in a form that should be accessible to someone starting out in digital photography.

Thankfully he has described "Photography's Triangle", rather than incorrectly calling it the "exposure triangle" as many often do. The triangle is a valid tool, although it would be more helpful and accurate to describe ISO as an "image brightening" factor, which can be applied either by increasing its value in-camera or by using post-processing software (or both).

With regard to ISO, it is my view that beginners should start by using in-camera ISO values as part of the triangle to produce an image of appropriate brightness. The reason is that developing the fundamental skills of vision and composition should be practised from the start, and viewing a mostly black image on the rear screen, because a low light shot was captured at base ISO for recovering later in software, is of little use or help to the learner.

Philip

[Des]

Nice work Dave. It would help, I think, to illustrate the differences in shutter speed, aperture and ISO with pictures. It helps novices understand the creative possibilities as well as the challenges.

The next step is to explain the interaction by reference to stops (or EV steps).

I would stick with the triangle concept. From a novice's point of view, it works like a triangle. In simple terms, same amount of light, same focal length, 1/125th f/5/6 100 ISO, 1/125th f/8 200 ISO and 1/250th f/8 400 ISO will give the same exposure. That is, you have a certain light "budget" in a given situation. If you spend it on a narrower aperture, a faster shutter or a lower ISO, you have to save the equivalent (converted to EV stops) somewhere else to get the same exposure. EV stops of shutter and ISO are easy enough to grasp (double or half is one stop), but aperture stops are not so intuitive.

[mhoule418]

Here is an image I found that sums it all...

What do you think?

[GUB]

Yes I agree that for the moment we need to adhere to the status quo and my apologies if it appears I have hijacked the thread.
But if the Iso-invariance feature continues to be a part of Sony/Pentax camera development then at some point we need to address that status quo at the entry level.

[GUB]

Even with ISO-invariant sensors, there's still a triangle. The first two sides of the triangle are shutter speed and aperture. One can make the final image brighter by slowing the shutter speed or by opening the aperture. And the third leg for adjusting image brightness is created either by adjusting ISO before the shot or by using base ISO plus boosting the brightness after the shot during post processing.

Consider this scenario of a dim scene where the meter says to use:

1/15, f/1.4, ISO100

But that's too slow shutter speed and may be too wide an aperture for the subject matter and photorgrapher's intentions. The photographer could shoot at 1/60, f/2.8, ISO100 to stop the subject's motion and improve the depth-of field but this creates a very very dark image -- 4-stops underexposed!

To compensate for the 4 stops of lost light caused by a faster shutter speed and narrower aperture, the photographer can either shoot:

1/60, f/2.8, ISO 1600 or

1/60, f/2.8, ISO 100 + 4EV in post

With an ISO-invariant sensor, both shots have the same final brightness and the same higher final noise level.

ISO invariance only means that the image looks the same whether you increase the ISO before the shot or boost the exposure after the shot in post. It simply means that either higher ISOs and higher post-processing exposure boost have the same effect (they both brighten the image and make it noisier).

You and I agree on the technical facts here.
What we disagree on is the practical visualisation of them.
I started out in the exposure triangle world of film like most of us here and carried it through to digital like all of us. So I understand it but now think (still correctly) differently and consider it a better path for learners to involve into. It is just a way of thinking-- take a look at how I now approach it--
My camera is set to Raw and Iso is set to base unless there is a large shortfall of light.
My personal habit is to be in M mode and I chimp. So my histogram is a key tool like most of you.
I set my shutter as slow as is appropriate for the situation.
I set my aperture as wide as appropriate for the situation.
I glance at my histo to check for either overexposure or the extent of underexposure.
I see that level of underexposure as a measure of noise and symptomatic of my two variables (aperture and shutter)
Iso has no need to come into the equation.
Should I consider the potential noise situation to be excessive I will readdress my two variables.
So quite simply look at your histogram as a noise indicator and forget Iso.
How simple can it be and I believe I am still at best practise.

[photoptimist]

You and I agree on the technical facts here.
What we disagree on is the practical visualisation of them.
I started out in the exposure triangle world of film like most of us here and carried it through to digital like all of us. So I understand it but now think (still correctly) differently and consider it a better path for learners to involve into. It is just a way of thinking-- take a look at how I now approach it--
My camera is set to Raw and Iso is set to base unless there is a large shortfall of light.
My personal habit is to be in M mode and I chimp. So my histogram is a key tool like most of you.
I set my shutter as slow as is appropriate for the situation.
I set my aperture as wide as appropriate for the situation.
I glance at my histo to check for either overexposure or the extent of underexposure.
I see that level of underexposure as a measure of noise and symptomatic of my two variables (aperture and shutter)
Iso has no need to come into the equation.
Should I consider the potential noise situation to be excessive I will readdress my two variables.
So quite simply look at your histogram as a noise indicator and forget Iso.
How simple can it be and I believe I am still at best practise.

That seems like a sensible workflow for someone who works in RAW with an ISO invariant sensor, especially if they want to preserve any overlooked highlights.

However, if the image is underexposed even by some amount, you do presumably take steps to brighten it in post by some number of stops. And if you have to brighten it by many stops, the final image is noisier. With an ISO invariant sensor, a different photographer could have bumped the ISO by the same number of stops as you used in post and gotten the same results with the same noise levels (with the key exception of any blown highlights that you protected by shooting at lower ISO).

And you will admit that if you had picked a slower shutter speed or a wider aperture, then that would have reduced the amount of under-exposure and reduced the noise in the final image. So there is a trade-off between shutter speed, aperture, and some sort of image brightening whether done as a post process or done as a pre-process with ISO gain.

The triangle simply expresses the trade-offs between shutter speed, aperture, and a third variable (either the ISO setting of the camera or the EV boost in post) that all produce the same overall gray levels in the final image (albeit with noticeable differences in motion blur, DoF, and noise).

In photography, you can freeze motion, have a deep DoF, or have a clean image but you often must pick two out of the three.

[GUB]

And if you have to brighten it by many stops, the final image is noisier.

As indicated by the in-camera histogram

The triangle simply expresses the trade-offs between shutter speed, aperture, and a third variable (either the ISO setting of the camera or the EV boost in post) that all produce the same overall gray levels in the final image (albeit with noticeable differences in motion blur, DoF, and noise).

In photography, you can freeze motion, have a deep DoF, or have a clean image but you often must pick two out of the three.

And once more it is all in the visualisation of this.
For me the trade offs are;
dof and diffraction ------- Aperture
Motion blur or motion freezing -- Shutter
Noise ------- The combination of above

[K-Three]

An image worth a 1000 words(they say...)

This graphic has always bothered me, I initially thought it's connecting ƒ/32, 1/1000s & ISO 50 (super nova as a light source?) or ƒ/1.4, 1/2s & ISO 2560 (getting the canary in the coal mine?). I guess I'm trying to make the triangle
After looking at it longer, I realized that it's just meant to be read left -> right, with no connection between each row, but I keep trying to get more out of it.
Sometimes too much knowledge is a bad thing

[stevebrot]

Not to be critical of Dave, but I would have avoided the notion of triangle or tripod, despite the popularity of the concept.

There is no triangle. The only things that affect exposure are shutter speed and aperture. Increasing ISO amplifies both signal and noise but does not change the sensitivity of the sensor, that is fixed at the base ISO setting.

I know there is a good argument upcoming but I am with steephill here. There is only two variables to input.
EDIT that is presuming you are using a modern Pentax (iso-invariant).

Yes, while ISO is important, its importance is the same when it refers to film sensitivity as when it refers to digital capture sensitivity.* You choose what you want for the task at hand and unless you don't care leave auto-ISO off.

That out of the way, it is pretty obvious that ISO is not a component of exposure (how "exposed" the sensor/film is to the subject's illuminance). There is nothing about the ISO setting that increases or decreases light to the medium. I used to teach the exposure triangle to digital noobs until I realized they seemed confused and that it was not part of how I learned how to operate a camera. I changed how I taught and was amazed at how easily concepts were picked up once ISO was placed in the background and left fixed in much the same way as film speed was fixed once the camera was loaded. There is a reason film photography instructors have the students shoot with Tri-X the entire first term. There is very little in reasonable light that cannot be addressed at ISO 400.

Teach how to use the shutter and aperture for creative effect and to control exposure and the rest takes care of itself.


Steve

* Actually digital ISO has more to do with metering than the sensor behavior, but that is another discussion. Anyone with doubts can reference the CIPA makers standards for the SOS (used by Pentax) option for digital ISO.

[Kiwizinho]

I know there is a good argument upcoming but I am with steephill here. There is only two variables to input.
EDIT that is presuming you are using a modern Pentax (iso-invariant).

You are correct but...

...how people use their camera makes a big difference to whether they can do the same amplification post-production as increasing ISO does in camera. Shoot RAW in M mode, and only aperture and shutter speed matter.

Shoot jpg and use any mode where the camera is allowed to set parameters automatically there can be issues, as the camera will adjust shutter speed or aperture or both according to each other and the ISO setting which may not give the desired result, and in any mode, if you shoot an underexposed jpg, the camera discards a lot of raw image information so shadow information is lost.
With a RAW file, all the sensor information is there, so it doesn't matter whether you amplify it in camera or in post processing, but jpg discards a lot.
At 8 bit per channel jpg can only represent 256 brightness levels, whereas current Pentax models K-70 and up are 14 bits per channel, or 16,384 brightness levels, so what are actually 64 shades of grey in a raw file get compressed to just one shade of black in a jpg.
The problem doesn't mean lack of ISO invariance in the raw data, it's just how the camera compresses tonal range down when it converts to jpg, as it's effectively doing the post processing and discarding unused data before you get to it.
I've seen this effect with some of my images. RAW files have enabled recovery of usable images from what would have been total rejects as jpg, by increasing exposure in post processing.

I guess ISO on a modern camera is a bit like the notion of 'centrifugal force' which doesn't exist. There is no force flinging a spinning object outwards. It is actually just trying to continue in a straight line, and there is actually a force pulling inward preventing it from doing so, but the concept of 'centrifugal force' is convenient.
In the same way the concept of ISO increasing sensitivity isn't true in a physical sense but is a convenient concept, especially when it comes to automatic metering when producing jpg images.

[beachgardener]

I don't understand all this no ISO thing at all. You say put the camera on 100 iso (base) and choose any shutter speed or aperture, shoot any scene, and fix it in post processing? All sounds like rubbish to me.

[GUB]

In the same way the concept of ISO increasing sensitivity isn't true in a physical sense but is a convenient concept, especially when it comes to automatic metering when producing jpg images.

And I agree that we are stuck with this concept while camera settings are presented as a triangle (by the makers). Iso was an elegant term to hijack at the start of the digital age when things were iso variant and so needed the three variables.
But imagine now with an Iso invariant camera if you had your two true variables and Iso was presented as EV (in stops) gain required to make what the camera thought was a correct image. One glance at that EV number and you soon learn whether it is acceptable noise for your requirements. And there is no need to unlearn as you progress.
And imagine if that camera had an option to save to Raw with or without that EV applied. That would keep everyone happy.
You may say that there is no difference apart from the name (Iso and EV) but it breaks the link that iso inevitably has with sensitivity. (which creates the triangle)
Of course the big BUT here is will cameras continue to be iso invariant? If I remember correctly the best practise with the new accelerator units in the K1ii would be to increase to iso 400 and utilise iso invariance (if you wish) from then on.

[GUB]

Yep rubbish is my second name
Iso invariance - PentaxForums.com
Iso-invariance indisputably exists - whether you choose to utilise it is up to you.

[macman24054]

dof and diffraction ------- Aperture
Motion blur or motion freezing -- Shutter
Noise ------- The combination of above

Need to check your info. Aperture and shutter speed have nothing at all to do with noise.

[GUB]

Aperture and shutter speed are the two variables that dictate the light levels on the sensor. If they are unnecessarily restrictive then the noise to signal ratio gets unnecessarily increased and further magnified by downstream gain. Noise is totally to do with the two variables.