ISO seems like global brightness - True?

Sometimes you cannot set the ISO right. On one of my cameras max ISO is ISO 6400 that is quite insufficient. Or you have blown highlights so you have to reduce general brightness and then recover a part of the shadows. The old implementation was not so nice in contrasty situations where I saw an ISO 100 image that had in the shadows noise that was like at ISO 6400 and a ISO 1600 image without perceivable noise. With newer approach I saw less and less the shadow noise even when pushing the shadows quite a lot.

I don't want to turn this into a sterile controversy. I told that ISO was not similar with with a brightness control because in some cases made noise more perceivable depending on the approach.

As you know a lost highlight is not recoverable while a lost shadow is partially recoverable. Thanks to new developments we have more tools to protect highlights.

The Ghost of Caravaggio said:

mostlyboringphotog said:

...Well, I think Standard Deviation reflects the noisiness of the image file better than SNR.

SNR of the viewing image can be different depending on how the viewing image is derived.

Click to expand...

Nope.

The SNR determines the datas' maximum possible information content. Put another way the uncertainty in the data decreases as the SNR increases. What happens after the data is recorded is irrelevant. The limiting factor is always SNR.

Post acquisition it's impossible to add more information without making more measurements.

Granted sub-optimal data analysis/processing/display can be inefficient and not take full advantage of the datas' intrinsic information content. But wasting SNR doesn't mean SNR is not the single most important attribute of data.

This is a fundamental tenet of information theory. You can ignore, deny it or even refute it. But that doesn't change how the rest of the scientific and technical world uses information theory to solve a diverse range of practical problems when SNR is limited.

Click to expand...

+1

Sorry for tacking my post onto yours, but yours reads really well.

As for ISO affecting SNR . . . I was thinking . . .

If, for instance, you are shooting at ISO400 and you are getting the shots. But then, for whatever reason, you decide to turn up the ISO to ISO800. But then your shots come out "over exposed", so . . . instead of turning down the ISO back to ISO400, you decide to shut down the aperture a stop instead. The act of shutting down your aperture (and not changing your shutter speed) does reduce the light and thus your signal?

So . . . turning up the ISO itself does not affect the signal, but . . . if you end up changing your aperture or shutter speed to lower the light to compensate for the higher ISO, then . . . in a round about way . . . is changing the ISO affecting the end image?

LOL.

Take care & Happy Shooting!

baloo_buc said:

Sometimes you cannot set the ISO right.

Click to expand...

IMO "Right" ISO setting is not what folks use to think it is. To me, it is the the setting that minimizes unwanted highlight clipping and ensures the lowest possible artifact level after the shadows are opened. It is nearly never the base ISO in low-light situations.

--
http://www.libraw.org/

keano12 said:

I know how to work ISO, Shutter and Aperture. Thing to me is that I tend to view ISO as a global brightness. It makes the whole image brighter. Is this simple analogy close enough the way I view it?

Click to expand...

With GB and Iliah you've gotten the right answers. Allow me to summarize and make some additional points.

First, the term brightness refers to the apparent lightness of a displayed image, either on the camera's LCD, your (or other viewer's) computer monitor, or a print. There are many things that can affect this brightness: in-camera ISO, in-camera tone curves, in-camera LCD brightness setting, tonal adjustments during raw processing and/or post-processing, monitor or printer brightness controls, print display lighting.

Second, the term brightening refers to any of the means noted above that affect brightness, but it does not refer to the brightness itself. Brightness is a perceptual concept (the viewing), brightening is a technical concept (the doing).

Third, a raw file has no brightness since it is data and is not displayed. It may contain some brightening information: either directly by having altered data values due to in-camera ISO (most cameras) or indirectly by having in-camera ISO information included as part of the metadata (some few cameras). In either case, there is no brightness associated with the data. The brightness of any displayed image ultimately produced using these raw data is completely up for grabs. It may be made as dark or as light as desired (regardless of any in-camera ISO employed) and is determined during the processing of the raw data and through any controls of the display medium.

Fourth, in-camera ISO always affects the brightening applied to the camera's LCD image and/or OOC JPEG, but it may or may not affect the brightness of those images. As GB points out, when working in one of the semi-programmed modes, such as Av, changing ISO can result in images of the same brightness. Thus, in Av-mode at, say, f/4, if ISO is at 100 and the camera determines a SS of 1/250 secs., a shift to ISO 200 will result in a SS of 1/125 secs. The exposure is halved, and the brightening is doubled, but the brightness stays the same. That is, increased brightening has compensated for decreased exposure so as to leave brightness (LCD and OOC JPEG) the same – but, please note, accompanied by decreased s/n because of the decreased exposure. This is true only in programmed modes. In M-mode, where exposure is fixed by set values for both f-ratio and SS, in-camera ISO brightening will also directly affect the brightness of those images.

Fifth, applying brightening in-camera via ISO is not the same as brightening during raw processing or during post-processing. With almost all cameras, increasing in-camera ISO, up to some camera-specific point, will decrease read noise. This same benefit does not accrue to brightening applied afterwards. And as Iliah and J A C S point out, for some cameras there can also be non-beneficial anomalies that accompany differing in-camera ISOs that do not pertain to subsequent brightening.

For those interested in learning more on how brightening relates to (and differs from) exposure, please see Exposure vs. Brightening.

--
gollywop
I am not a moderator or an official of dpr. My views do not represent, or necessarily reflect, those of dpr.
http://g4.img-dpreview.com/D8A95C7DB3724EC094214B212FB1F2AF.jpg

TacticDesigns said:

Sorry for tacking my post onto yours, but yours reads really well.

As for ISO affecting SNR . . . I was thinking . . .

If, for instance, you are shooting at ISO400 and you are getting the shots. But then, for whatever reason, you decide to turn up the ISO to ISO800. But then your shots come out "over exposed", so . . . instead of turning down the ISO back to ISO400, you decide to shut down the aperture a stop instead. The act of shutting down your aperture (and not changing your shutter speed) does reduce the light and thus your signal?

So . . . turning up the ISO itself does not affect the signal, but . . . if you end up changing your aperture or shutter speed to lower the light to compensate for the higher ISO, then . . . in a round about way . . . is changing the ISO affecting the end image?

Click to expand...

Think of it this way.

When you do what the meter tells you to do increasing ISO decreases exposure,

What happens is: you reduce the light amplitude because the meter recommended you use a faster shutter time and, or smaller aperture compared to a lower ISO.

Usually this is good advice because you need to minimize subject an, or camera motion... or you need more depth of field. However less light (a.k.a. signal is measured. Increasing ISO by one stop decreases the signal recorded when the shutter is open. ISO amplification then increase the signal and the noise after the shutter closes. The global brightness (luminance) is identical but the SNRs are not. This assumes you're using a relatively new camera where the read noise is essentially independent of ISO. This technology is called ISO invariance. Then ISO amplification adds very little, if any, read noise. So when ISO increases the read noise is remains almost constant but the signal level decreases.

ISO invariance means IQ is not degraded if you use base ISO and just choose the whatever shutter time and aperture you want. Then during post-production you can increase the global brightness as needed. This strategy assumes you are willing to forgo viewing the jpegs in-camera, you use raw files and you don't over expose the sensor when the shutter is open.

The Ghost of Caravaggio said:

Think of it this way.

ISO invariance means IQ is not degraded if you use base ISO and just choose the whatever shutter time and aperture you want. Then during post-production you can increase the global brightness as needed. This strategy assumes you are willing to forgo viewing the jpegs in-camera, you use raw files and you don't over expose the sensor when the shutter is open.

Click to expand...

We are in the Beginners QF and I see some folks touting an ISO-invariant lifestyle; just set the ISO at whatever value is the point where the read noise levels out [stops increasing] and set whatever shutter duration and aperture is appropriate for the image.

I highlighted your last sentence because it is important. Many beginners NEED to chimp their shots. This is one of the most powerful aspects of digital photography; we don't need to wait days or weeks to see the pictures we took. It has made learning to take decent pictures easier. Of course 99% of the photographers just point their iPhone and take a snapshot. But for the 1% who aspire to something better, it has been a boon to have instant feedback!

Now, we have these amazing cameras that are ISO-invariant and some technologists think that characteristic can be used to take better images. Undoubtedly true. BUT, they really should not suggest to beginners that they adopt the ISO-invariant lifestyle w/o explaining that:

• There is no instant feedback, which means that they can't check to see if a picture was framed well or if it was focused well
• Manual exposure settings when in very dim venues works best
• Saving RAW files is required
• Every photo must be post processed

The only way I have found to apply an ISO invariant paradigm and use the expensive auto-exposure system in my D810 is to dial in lots of EC. I basically look through the VF and watch the exposure graph and spin the rear dial until it is where I want it [under-normal-over]. Is there another way?

I really think this amazing sensor ability is being used to lead beginners into the ocean.

chuxter said:

The Ghost of Caravaggio said:

Think of it this way.

ISO invariance means IQ is not degraded if you use base ISO and just choose the whatever shutter time and aperture you want. Then during post-production you can increase the global brightness as needed. This strategy assumes you are willing to forgo viewing the jpegs in-camera, you use raw files and you don't over expose the sensor when the shutter is open.

Click to expand...

We are in the Beginners QF and I see some folks touting an ISO-invariant lifestyle; just set the ISO at whatever value is the point where the read noise levels out [stops increasing] and set whatever shutter duration and aperture is appropriate for the image.

I highlighted your last sentence because it is important. Many beginners NEED to chimp their shots. This is one of the most powerful aspects of digital photography; we don't need to wait days or weeks to see the pictures we took. It has made learning to take decent pictures easier. Of course 99% of the photographers just point their iPhone and take a snapshot. But for the 1% who aspire to something better, it has been a boon to have instant feedback!

Now, we have these amazing cameras that are ISO-invariant and some technologists think that characteristic can be used to take better images. Undoubtedly true. BUT, they really should not suggest to beginners that they adopt the ISO-invariant lifestyle w/o explaining that:

Click to expand...

+1

I tend to agree with that.

I've set up my old Pentax ist DS with an old manual focus lens with an aperture ring and, if my kids are interested, I was going to show them how to use my incident light meter to set up the camera to shoot.

I'm going to show them how to follow the meter first.

IMHO I think it helps to start with how things are "set-up" to be used. Once you got that working (kinda), then start learning how to do it differently.

chuxter said:

• There is no instant feedback, which means that they can't check to see if a picture was framed well or if it was focused well
• Manual exposure settings when in very dim venues works best
• Saving RAW files is required
• Every photo must be post processed

Click to expand...

+1

And that last one is sometimes a biggy. If you are taking shots and you don't want to process them afterwards, like if it is casual vacation shots, or you are doing a lot of shots for something, then . . . not having to process the shots may be more desirable.

chuxter said:

The only way I have found to apply an ISO invariant paradigm and use the expensive auto-exposure system in my D810 is to dial in lots of EC. I basically look through the VF and watch the exposure graph and spin the rear dial until it is where I want it [under-normal-over]. Is there another way?

Click to expand...

I guess just going full manual exposure, which is what I have been leaning more towards when taking portraits.

chuxter said:

I really think this amazing sensor ability is being used to lead beginners into the ocean.

Click to expand...

IMHO I think you have a point there.

Take care & Happy Shooting!

gollywop said:

Fourth, in-camera ISO always affects the brightening applied to the camera's LCD image and/or OOC JPEG, but it may or may not affect the brightness of those images. As GB points out, when working in one of the semi-programmed modes, such as Av, changing ISO can result in images of the same brightness. Thus, in Av-mode at, say, f/4, if ISO is at 100 and the camera determines a SS of 1/125 secs., a shift to ISO 200 will result in a SS of 1/250 secs. The exposure is halved, and the brightening is doubled, but the brightness stays the same. That is, increased brightening has compensated for decreased exposure so as to leave brightness (LCD and OOC JPEG) the same – but, please note, accompanied by decreased s/n because of the decreased exposure. This is true only in programmed modes. In M-mode, where exposure is fixed by set values for both f-ratio and SS, in-camera ISO brightening will also directly affect the brightness of those images.

Click to expand...

Please note that in this example I mistakingly reversed the shutter speeds. The paragraph should read as corrected in bold above.

--
gollywop
I am not a moderator or an official of dpr. My views do not represent, or necessarily reflect, those of dpr.
http://g4.img-dpreview.com/D8A95C7DB3724EC094214B212FB1F2AF.jpg

A minor point. In some cases ISO is implemented as darkening (or negative brightening ), for example Canon 5D Mk II and some other Canon cameras scale raw data down for certain intermediate ISO settings.

I wasn't recommending any particular technique for making photographs.

All I was trying to do is explain how come increasing ISO above a camera's base ISO decreases the signal-to-noise ratio.

The behavior of ISO invariant cameras helps make this point since increasing ISO does not add read noise.

I agree new photographers have different needs than experienced photographers. Eventually new photographers become experienced photographers who can make their own decisions.

In my view your arguments against an ISO-invariant workflow for new photographers also apply to the use of auto-ISO.

The Ghost of Caravaggio said:

The behavior of ISO invariant cameras helps make this point since increasing ISO does not add read noise.

Click to expand...

ISO "variant" cameras (all of them) actually reduce the read noise with increasing the ISO!

J A C S said:

The Ghost of Caravaggio said:

The behavior of ISO invariant cameras helps make this point since increasing ISO does not add read noise.

Click to expand...

ISO "variant" cameras (all of them) actually reduce the read noise with increasing the ISO!

Click to expand...

ISO invariant cameras do not add negative read noise

Actually, we are discussing partial truths here. I'm afraid we are leaving the impression that read noise is the only image quality factor to consider when deciding on the ISO setting.

The Ghost of Caravaggio said:

I wasn't recommending any particular technique for making photographs.

Click to expand...

+1

I meant to reply to your other post . . .

The Ghost of Caravaggio said:

All I was trying to do is explain how come increasing ISO above a camera's base ISO decreases the signal-to-noise ratio.

The behavior of ISO invariant cameras helps make this point since increasing ISO does not add read noise.

Click to expand...

I've got the Nikon D750 and reading the review here, and the airplane museum example really got me rethinking ISO.

The Ghost of Caravaggio said:

I agree new photographers have different needs than experienced photographers. Eventually new photographers become experienced photographers who can make their own decisions.

Click to expand...

IMHO, the reason I want to start with following the light meter for my kids is that IMHO it gives a starting point / framework to start adding things to.

The Ghost of Caravaggio said:

In my view your arguments against an ISO-invariant workflow for new photographers also apply to the use of auto-ISO.

Click to expand...

Yes and no?

The thing about manual exposure and auto-iso is that if you are choosing both the aperture (to determine depth of field) and then shutter speed (to control motion blur), then to get a "proper" exposure, auto-iso is just letting the camera choose the iso to get the "proper" exposure.

Is that different than choosing the ISO, Aperture and then letting the camera choose the shutter speed?

Take care & Happy Shooting!

TacticDesigns said:

The thing about manual exposure and auto-iso is that if you are choosing both the aperture (to determine depth of field) and then shutter speed (to control motion blur), then to get a "proper" exposure, auto-iso is just letting the camera choose the iso to get the "proper" exposure.

Click to expand...

You mean: auto-ISO is just letting the camera choose the ISO to get the brightening needed for achieving an "appropriate" brightness (in the eyes of the camera manufacturer) for the OOC JPEG – but not necessarily one relevant to the best raw capture.

TacticDesigns said:

Is that different than choosing the ISO, Aperture and then letting the camera choose the shutter speed?

Click to expand...

Indeed it is. How do you know, a priori, what ISO is relevant to having the camera choose the SS you've deemed appropriate to controlling motion blur?

--
gollywop
I am not a moderator or an official of dpr. My views do not represent, or necessarily reflect, those of dpr.
http://g4.img-dpreview.com/D8A95C7DB3724EC094214B212FB1F2AF.jpg

gollywop said:

TacticDesigns said:

The thing about manual exposure and auto-iso is that if you are choosing both the aperture (to determine depth of field) and then shutter speed (to control motion blur), then to get a "proper" exposure, auto-iso is just letting the camera choose the iso to get the "proper" exposure.

Click to expand...

You mean: auto-ISO is just letting the camera choose the ISO to get the brightening needed for achieving an "appropriate" brightness (in the eyes of the camera manufacturer) for the OOC JPEG – but not necessarily one relevant to the best raw capture.

Click to expand...

+1

Exactly.

You're letting the camera decide what to do to get to the program line. But giving up a bit of control?

If you are determining that the aperture and shutter speed is more important to you, then the iso setting is the one to give control to the camera?

gollywop said:

gollywop said:

Is that different than choosing the ISO, Aperture and then letting the camera choose the shutter speed?

Click to expand...

Indeed it is. How do you know, a priori, what ISO is relevant to having the camera choose the SS you've deemed appropriate to controlling motion blur?

Click to expand...

+1

Exactly.

You're letting the camera decide what to do to get to the program line. But giving up a bit of control?

If you are determining that the aperture and iso is more important to you, then the shutter speed setting is the one to give control to the camera?

Of course . . . the other option is to use auto-iso and set the desired shutter speed. This is available on my Nikon cameras, and . . . I have to admit that it works pretty well. It IMHO is really useful when using a zoom lens without a constant aperture.

Take care & Happy Shooting!

TacticDesigns said:

Yes. And No?

All 3 of those settings can "brighten" the image. But each one is tied to another aspect that can affect the look of the picture.

Aperture, if you open it up, can let in more light, but that affects the depth of field that is captured.

Shutter speed, if you slow it down can let in more light, but that affects the amount of motion blur (or lack of) as well.

Increase the iso can increase the brightness of the final image, but (as some one has pointed out) with the potential of more image noise and perhaps with reduced dynamic range?

Take care & Happy Shooting!

Click to expand...

I enjoyed your response, as well as others you have made in this thread. You mentioned in one of them a Pentax *ist DS. I happen to use Samsung's clone of its successor (that is, my GX-1S is a clone of the *ist DS2), which also has a 6 mp sensor with a lowest ISO setting of 200.

There are a lot of ways to approach photography, and mine hasn't changed much since my slide film days. I use JPEG output and expose for the midtones, letting highlights and shadows fall where they will; just as with slide film, I don't give thought to adjusting things like this after pressing the shutter release. In contrast to the OP, my primary control for adjusting overall brightness is shutter speed.

I always leave ISO at 400 because ISO 200 clips highlights a bit more, and ISO 400 is virtually indistinguishable to my eyes in terms of noise. ISO 800 has both visibly higher noise and reduced detail, so I don't like to go there. Having used ISO 100 Fuji Astia for many years as my only slide film, keeping my DSLR permanently at ISO 400 doesn't feel like a limiting factor. I like deep depth of field and almost always have my aperture set to f/8 (which is also my kit zoom's sweet spot for across-the-frame resolution and reduced chromatic aberrations at my preferred 35mm focal length setting). That leaves shutter speed as the variable I turn to first to adjust my camera for the prevailing light conditions.

My approach is simple because my needs are simple. Photography for me is mostly typical holiday/vacation snapshots taken outdoors in daylight. On exposure value charts, I rarely encounter conditions below EV 11, which is described as sunsets or subjects in deep shade. Keeping my ISO at 400 and even using f/8, that still gives me a hand-holdable shutter speed of 1/100 second.

Over the years, friends have sometimes suggested that a simple box camera is more appropriate to my needs than an SLR or DSLR, but I do love that optical, through-the-lens, pentaprism viewfinder.

keano12 said:

I know how to work ISO, Shutter and Aperture. Thing to me is that I tend to view ISO as a global brightness. It makes the whole image brighter. Is this simple analogy close enough the way I view it?

Click to expand...

That's really not the way I look at it. I would say ISO, shutter speed and aperture ALL control brightness, or more correctly, exposure. Exposure wise, each is roughly equivalent and linked, but produce different results:

For instance, higher ISO produces more noise or "grain" in the image. Lower ISO gives you a finer tone.

A wider (bigger) aperture, gives you a shallower depth of field, a smaller aperture more depth of field.

A faster shutter speed helps freeze action, a slower shutter speed lets you blur.

In all three cases you are changing the "brightness" (exposure) of the image up or down. Most cameras have 1/3rd or 1/2 click stops or marks on these dial mechanisms so you can easily keep track.

--
photojournalist
http://craighartley.zenfolio.com/

hotdog321 said:

higher ISO produces more noise or "grain" in the image.

Click to expand...

When just the ISO setting is increased, most of the cases noise goes down or stays the same.

hotdog321 said:

keano12 said:

I know how to work ISO, Shutter and Aperture. Thing to me is that I tend to view ISO as a global brightness. It makes the whole image brighter. Is this simple analogy close enough the way I view it?

Click to expand...

That's really not the way I look at it. I would say ISO, shutter speed and aperture ALL control brightness,

Click to expand...

yes, at least of the OOC JPEG.

hotdog321 said:

or more correctly, exposure.

Click to expand...

no

hotdog321 said:

Exposure wise, each is roughly equivalent and linked, but produce different results:

Click to expand...

ISO is not equivalent to aperture and shutter-speed in determining exposure. Indeed ISO has no role in defining exposure.

--
gollywop
I am not a moderator or an official of dpr. My views do not represent, or necessarily reflect, those of dpr.
http://g4.img-dpreview.com/D8A95C7DB3724EC094214B212FB1F2AF.jpg