How much does ISO matter?

[Lanidrac]

I think the last thread by BAK needs to be continued. There was a lot wrong information by folks stating brain farts.

 

[beatboxa]

I think the last thread by BAK needs to be continued. There was a lot wrong information by folks stating brain farts.

Why? Those threads come up all the time, and usually have the same issues.

• People may confuse causality with correlation
• People may misunderstand each other due to imprecise language
• People genuinely may be stubborn and refuse to learn
• People may not realize that different cameras have different properties

In any event, it's probably not worth reviving a thread unless you have specific objections. Even then, might be best to let people who refuse to believe facts go on their merry way in blissful ignorance.

 

[Lanidrac]

I think the last thread by BAK needs to be continued. There was a lot wrong information by folks stating brain farts.

Why? Those threads come up all the time, and usually have the same issues.

• People may confuse causality with correlation
• People may misunderstand each other due to imprecise language
• People genuinely may be stubborn and refuse to learn
• People may not realize that different cameras have different properties

In any event, it's probably not worth reviving a thread unless you have specific objections. Even then, might be best to let people who refuse to believe facts go on their merry way in blissful ignorance.

I can't agree more. But when you see stuff posted that is so bogus and wrong, it's annoying.

 

[jshen808]

I think the last thread by BAK needs to be continued. There was a lot wrong information by folks stating brain farts.

In today's digital photography, "ISO" matters greatly (just my opinion).. Cheers!

--
Cheers, John
Photography is my hobby.
http://www.pbase.com/johnshenphotography

 

[Mike_PEAT]

I think the last thread by BAK needs to be continued. There was a lot wrong information by folks stating brain farts.

Wrong to who? First of all there are multiple methods manufacturers use to measure ISO.

Second, people believe what they believe, and it's tough to tell them the truth.

For instance, people wrongly believe that any sensor smaller than 35mm/135 format has a "crop factor", but if you are using a lens designed for the smaller sized sensor there is no cropping taking place. There IS a multiplication factor IF you are comparing it with a different sized sensor, but if you are using a medium or large format lens on a poorly named "full frame" camera, then IT has a crop factor!

The world isn't flat, it's almost spherical.

 

[Mark S Abeln]

Lanidrac wrote:.

I can't agree more. But when you see stuff posted that is so bogus and wrong, it's annoying.

Unfortunately, the primary source document defining ISO, the ISO 12232:2006 standard, is not online. It's expensive to purchase, and very few libraries have a copy.

Of course people are going to argue, it's not that very many people have actually read the document, and its reported contents is mainly hearsay. If even the camera manufacturers get it wrong in their documentation, you know we're in trouble.

--
http://therefractedlight.blogspot.com

 

[Aaron801]

... unless it isn't the kind of thing that you particularly care about.

 

[biza43]

As much as the other two vertices of the exposure triangle.

 

[Bob A L]

As much as the other two vertices of the exposure triangle.

--
www.paulobizarro.com
http://blog.paulobizarro.com/

We have a winner - congratulations

 

[Janoch]

As much as the other two vertices of the exposure triangle.

--
www.paulobizarro.com
http://blog.paulobizarro.com/

We have a winner - congratulations

Bummer!

I would have said 42%...

 

[D Cox]

I think the last thread by BAK needs to be continued. There was a lot wrong information by folks stating brain farts.

Why? Those threads come up all the time, and usually have the same issues.

• People may confuse causality with correlation
• People may misunderstand each other due to imprecise language
• People genuinely may be stubborn and refuse to learn
• People may not realize that different cameras have different properties

In any event, it's probably not worth reviving a thread unless you have specific objections. Even then, might be best to let people who refuse to believe facts go on their merry way in blissful ignorance.

I can't agree more. But when you see stuff posted that is so bogus and wrong, it's annoying.

Only for a moment. There are real problems and bad things in the world.

 

[Lightpath48]

My most recent camera is far beyond my earlier ones, in being able to render good quality high ISO files.

 

[digidog]

I think the last thread by BAK needs to be continued. There was a lot wrong information by folks stating brain farts.

Wow, another 140 odd possibilities for some to post more misinformation on ISO. That would make perfect sense if you suspend all rational thought.

IF only those that promise to provide data points on such a subject actually did, and allowed others to peer review their opinions, we'd uncover a lot more facts and bury a lot more misinformation here. Alias, such posters who promise such data usually cannot deliver such data. Peer review isn't a flower that grows in their gardens.

How much does ISO matter? This much:

Loading…

www.guillermoluijk.com

--
Andrew Rodney
Author: Color Management for Photographers
The Digital Dog
http://www.digitaldog.net

 

[beatboxa]

I can't agree more. But when you see stuff posted that is so bogus and wrong, it's annoying.

I didn't follow that thread, but what was posted that is so bogus and wrong?

Here are my 2 cents on the topic.

There is a big difference between film & digital process flow and how ISO plays a part. In the film world, ISO is effectively set per roll--it is not a variable; so exposure must accommodate for the ISO. In the digital world, the opposite can be true: the ISO can accommodate the exposure.

So the first thing to determine is "is available light a constraint?" ie. Are we using 'exposure priority' or 'ISO priority'?

• "Exposure Priority": When light is limited, ISO setting is not a variable we can control.
  The total light, a combo of sensor area x exposure (scene lighting + t-stop + shutter speed) is the primary constraint;
  and this total light combined with desired brightness determines ISO. In this sense, the ISO is like "brightness balance," conceptually similar to "white balance". It's just compensating for the difference.
• "ISO priority": When light is not limited (eg. shooting still landscapes on a tripod), ISO can be controlled. The max SNR is the primary constraint;
  and setting the lowest ISO will subsequently determine the exposure.

That's step 1. On film, you are 'always' ISO-priority, while digital gives the option. This also implies that if you have limited lighting & a desired image brightness, you simply cannot control ISO.

For a given exposure (ie. light limit) & rendered brightness:

• On some cameras, ISO does not matter. "ISO Invariance" Either setting the highest ISO or brightening a lower ISO in post will result in the same image.
• On other cameras, the highest ISO that doesn't clip will be the best to shoot at. Brightening in post has a markedly different (and worse) result.
• I am not aware of any camera where brightening the lowest ISO will provide better IQ than using a highest ISO.

'ISO rendering' is roughly the same per total amount of light, as determined by the combo of photosensitive area x exposure.

As I mentioned earlier, people often confuse causality and don't control variables--and this is where many misconceptions can come from. Here are examples:

2 stop ISO difference (1600 vs. 6400), from an ISO-invariant camera, with limited lighting. Lower ISO brightened in post:





Here is a difference frame: subtracting one of the above from the other:



Almost identical.

Here's another example, illustrating ISO invariance; and the effect of higher ISO vs. lower ISO. All of these images have almost identical exposure & sensor area.



(Was one of the most extreme examples I could find). ISO 6400 looks pretty much the same between the cameras. On the D750 (ISO Invariant), the ISO 100 additionally looks similar to ISO 6400 except a bit of color cast. But on the 5DIII, the ISO 100 looks the worst....at the same exposure. It is not ISO invariant.

Similarly, if we look across sensor areas, in this example, the Olympus has 1/4th the sensor area of the Canon. This is 2 stops.



We see that ISO 12800 is better on the Canon than Olympus. But the Olympus has 2 stops less total light at the same exposure.

And so (predictably), when we give the Olympus 2 stops more exposure so that the total light between the Canon @ 12800 is the same as the Olympus @ 3200, the image quality is almost identical.



EDIT: I just wanted to add in one more point for emphasis: In all examples & both 'priorities' above, the common element is that you are maximizing exposure for your scene, equipment, and scenario. A vast majority of the time (except when I have 'unlimited' light), I cannot control ISO. It is determined for me. So why bother? Control the exposure and shoot at the ISO you need.

And so (no offense to the OP), I think the ISO question gets asked too much and is too often misunderstood; while a much more impactful question is:

"How much does total light matter?"

 

[AwesomeIan]

The importance of ISO performance is going to depend on the type of photography. A landscape photographer isn't going to push the ISO very often because he is on a tripod and does long exposure. However, the person who photographs birds in flight will push the ISO because they need 1/4000 shutter speed. So it all depends on who you ask.

 

[Iliah Borg]

if you know what ISO speed is.

 

[Lanidrac]

if you know what ISO speed is.

Please expand this statement. It's a bit short.

 

[Lanidrac]

What's your point? Some 5D2 samples that show lower ISO worse than higher ISO's?

 

[robert1955]

if you know what ISO speed is.

Please expand this statement. It's a bit short.

you were in that thread

 

[fishy wishy]

There's an unnecessary verbiage in these threads by people thinking they're very wise in pointing out that on digital sensors the ISO involved does not have such a profound influence as on film where high iso films were physically different.

If you go by the rule of thumb of high ISO is expressing low exposure values, you don't need to fret about the wording any more- you are free. I highly recommend this instead of getting bogged down in perpetual silliness with people on an ego-trip to prove they are right on every last pettifogging detail.

Normal people expose a dim scene with a high iso so they can actually see the photo they've taken after they've shot it. For reasons of quality it's better to avoid high iso if you can. That is very simple; it does not take 149+ posts to argue it.

You would soon find in your testing that you begin to lose a lot of detail by shooting past a certain point of high iso.

It's possible the original post was fishing for the answer to whether shooting at iso 100 and increasing exposure by +5 on a computer afterwards makes a real difference compared to shooting at iso 3200. So, why would this question preoccupy them enough to make a new thread about it? Does anyone think regularly massively underexposing shots by 5 stops so you can't effectively review them on the camera afterwards is a realistic way to proceed in photography? And to what advantage?