Optimum ISO for digital cameras....

[William W]

. . . So anyone know if anyone has done a test of cameras at various low ISOs- such as 50-100-200-400 and tested dynamic range? Wondering if nearly all sensors are ideal at 200ISO....

Yes. Done lots of testing. (Canon DSLRs and Fuji Mirrorless).

Results of that test don't matter much in the real world, unless you're ONLY EVER shooting absolute non-moving Subjects, with reasonable lenses in good daylight using a Tripod: because to 'display' the FINAL IMAGE usually (almost always) that Final Image Display will have LESS DR than the Camera, when the Camera is set at 'any' ISO: so you will need to finesse the raw file (i.e. Post Produce / Edit / Digital Darkroom) . . . and that finessing doesn't recon for much of whatever ISO you choose to use, so long as you nail the exposure.


***

. . . I thought the best ISO was the lowest ISO... many cameras the ideal ISO that has the most tonal range in RAW is iso200. This makes sense as Nikon in the past did not have ISOs lower than 200...

Not wishing to rain on the parade. But that’s theory: and theory is good to know.

Theory is REALLY good to know. I mean that.

But, just as a balance to answer the non-theoretical question “what is the best ISO” . . .

in the real world, the point is If you’re walking at night and you need to get the shot and you have an F/2 lens and a camera which is only capable of ISO 3200, and you NEED ISO 3200 @ F/2 to make the necessary Shutter Speed . . . then ISO 3200 is the “best” ISO to use:




***

. . . My Sony cameras go down to 50 ISO and this is helpful when trying to shoot fill flash and not having to use HSS- (high speed sync) which as you may know as the shutter speed increases becomes less effective as the power of the flash is greatly diminished.

If I understand the general meaning of that statement correctly: then the statement is generally incorrect or at the least displays a confused understanding.

For any given APERTURE, then Flash as Fill in daylight, will require a specific amount of FLASH POWER to make the desired Fill.

Assuming that one has reached the MAXIMUM Flash Sync Speed and that is what one desires (a fastest possible Shutter Speed) then the AMBIENT EXPOSURE will require a given APERTURE at that FLASH SYNC SPEED.

If one changes the ISO (to a lower ISO), ergo one MUST change the Aperture (open up Aperture).
Opening up the Aperture then requires MORE Flash power to act as Fill.

One might keep the ISO constant when opening up the Aperture but that would require slowing the Shutter Speed - which by implication in the statement is not what is wanted.

There are some minor nuances within various Camera Models, but generally if one is using Flash as Fill and one desires the fastest Shutter Speed WITHOUT engaging HSS, then (for most shooting scenarios) the BEST protocol is to set the MAX Flash Sync Speed; and then set (from what is possible) the desired APERTURE for DoF or for whatever other reason (for example best lens performance) and then check that Flash as Fill is within the realms of possibility and then set the ISO accordingly.

In this shooting scenario, the ISO is the least important factor and it is the last to be set.

Maybe I misunderstood the meaning of the statement - in which case please ignore this last part of the commentary.

WW

[DanK]

Bill,

You know far more about photography than I, and in the years I have been on this forum, I have never before had a reason to disagree with you, until today.

You wrote:

n the real world, the point is If you’re walking at night and you need to get the shot and you have an F/2 lens and a camera which is only capable of ISO 3200, and you NEED ISO 3200 @ F/2 to make the necessary Shutter Speed . . . then ISO 3200 is the “best” ISO to use:

Of course, but IMHO, that wasn't the OP's question. He wasn't asking, "should I forgo a shot if I need to use a higher ISO to capture it?" He was asking what the ideal is.

You wrote:

Results of that test don't matter much in the real world, unless you're ONLY EVER shooting absolute non-moving Subjects, with reasonable lenses in good daylight using a Tripod: because to 'display' the FINAL IMAGE usually (almost always) that Final Image Display will have LESS DR than the Camera,

Here it's me who is not sticking close to the OP's question. The OP's question was about dynamic range, and your answer focuses on that. But DR isn't the only issue in selecting an ISO setting. The other is noise, which of course is related to DR. I posted some test images here quite a long time ago taken with my 7D, which is a fairly noisy camera. What they showed is that IF the image has a fairly narrow DR and IF you ETTR, you can get away with boosting ISO quite a bit without severely damaging the image at screen resolution. However, often those conditions don't apply--either a wider DR in the scene forces you to include dim areas toward the left end of the histogram, or in a rush, you simply don't ETTR successfully. If for whatever reason you have substantial areas toward the left, the image degrades quite rapidly as ISO increases--again, dependent on the camera model.

I think this leaves the bottom line: keep the ISO as low as you can, given all the other constraints you have for a particular image. In your example, the constraint was low light, but it can be any number of things that leave you better off with a higher ISO than with a lower one.

Dan

[richardvallon]

the important point of what you said is that when the dynamic range in the scene is limited you can use a much higher ISO and still encompass the tones present - I was amazed in the DXO sensor tests to see how dynamic range fell as the ISO went up. I was recently photographing a stage presentation with a very flatly lit stage at iso3200 with a Sony a9. Color looks great and no loss of shadow detail or blow outs...

In a word, yes. I haven't noticed differences between 100 and 200 in practice (although I would if I tested in cases where the dynamic range is barely manageable at 100), but larger differences are often quite perceptible. When it is noticeable depends on a number of factors: the camera, the exposure level, the variation in luminance, and other aspects of the image. In the case of the Canon DSLRs I have used, it was most noticeable with my 50D, a bit less so with my 7D (first generation), and considerably less so with my 5DIII.

I posted a series of images here quite some time ago in which I showed that if the dynamic range is quite limited and you ETTR, you can often go quite high before the effects are really bad. However, those conditions often don't obtain.

[Manfred M]

Bill - I share Dan's experience here, not so much with my Nikon cameras, but with my Panasonic GX-7. a camera that reputedly has good noise characteristics (at least that is what the reviews that I read said). It has a base ISO of 200 and I try to shoot there because once I get to ISO 800 the sensor noise issues are definitely apparent in well exposed images.

That doesn't mean that I won't shoot it at higher ISO settings, but it does mean I have to deal with the side effects of shooting with more noise; more aggressive noise reduction and more aggressive sharpening, both of which do effect the quality of the final image.

[dem]

He was asking what the ideal is.

Naturally different people have different definition of what ideal or best or optimal is. Sure, ISO 200 will get you cleaner images than ISO 3200. But if setting ISO 200 means no image, it is hardly "ideal". Then there will be someone shooting ISO 3200 in day light because they do not like the clinical look of ISO 200 and want to introduce some grain.

Having said that, if someone wants to maximise dynamic range, they should shoot at camera's "native" ISO at which the signal is not amplified. The native ISO might be 173.5 and the nearest available might be ISO 200 at which the signal is amplified just a bit.

That assumes that there is plenty of light. If light is limited and you shoot raw using a camera that is less than 5 years old, 2-4 stops of underexposure will do the same damage to the image as cranking up ISO by 2-4 stops. There is no difference. Except for high DR scenes you are much more likely to blow highlights at high ISO.

[William W]

Bill, You know far more about photography than I, and in the years I have been on this forum, I have never before had a reason to disagree with you, until today.

. . .that wasn't the OP's question. He wasn't asking, "should I forgo a shot if I need to use a higher ISO to capture it?" He was asking what the ideal is.

Hi Dan,

The introductory sentence before the quote that you referenced is integral to the purpose of my response (bold and underlined now):

"But, just as a balance to answer the non-theoretical question “what is the best ISO” . . .


in the real world, the point is If you’re walking at night and you need to get the shot and you have an F/2 lens and a camera which is only capable of ISO 3200, and you NEED ISO 3200 @ F/2 to make the necessary Shutter Speed . . . then ISO 3200 is the “best” ISO to use:"

At the time of my response, the conversation was discussing the theory of DR and ISO - as I mentioned, I think knowing that theory is very good.

BUT - even though I did not address the question as written, the point that I was making as a balance to answer the non-theoretical question 'best' ISO cannot always be determined by the 'technical best DR': I don't know the OP's exact position; even less knowledgeable am I of the range of the vast readership here - sure I didn’t answer the question - just gave a bit of balance so folk don't get too engrossed in technical data at the expense of the shot that they might have made.

WW

PS - you know heaps more than I about Macro Photography.

[William W]

. . .
Here it's me who is not sticking close to the OP's question. The OP's question was about dynamic range, and your answer focuses on that. But DR isn't the only issue in selecting an ISO setting. The other is noise, which of course is related to DR. I posted some test images here quite a long time ago taken with my 7D, which is a fairly noisy camera. What they showed is that IF the image has a fairly narrow DR and IF you ETTR, you can get away with boosting ISO quite a bit without severely damaging the image at screen resolution. However, often those conditions don't apply--either a wider DR in the scene forces you to include dim areas toward the left end of the histogram, or in a rush, you simply don't ETTR successfully. If for whatever reason you have substantial areas toward the left, the image degrades quite rapidly as ISO increases--again, dependent on the camera model.

I think this leaves the bottom line: keep the ISO as low as you can, given all the other constraints you have for a particular image. . . .

Yes.

I concur on both points.

WW

[Ken MT]

Testing ISO and Dynamic range
Does this add to the argument

http://www.photonstophotos.net/Charts/DXOPDR.htm

[DanK]

Testing ISO and Dynamic range
Does this add to the argument

http://www.photonstophotos.net/Charts/DXOPDR.htm

It shows graphically what was described in a few of the earlier posts: an increase in ISO leads to a decrease in dynamic range.

If you fiddle with it, you will find some newer cameras for which the function is linear. I believe that is what is misleadingly called "ISO-less" cameras. AFAIK, the real impact of a linear trend is that for such cameras, it makes no difference whether you brighten an image by boosting ISO or by brightening it in postproduction. For most cameras, the nonlinearity at the left-hand end indicates that you damage the image less by boosting ISO at low values than by brightening in post.

[xpatUSA]

http://www.photonstophotos.net/Charts/DXOPDR.htm shows graphically what was described in a few of the earlier posts: an increase in ISO leads to a decrease in dynamic range.

That's Bill Claff's site. When he tested one of my cameras, he only put up a single point "curve" at 100 ISO. He may have changed his presentation since the Sigmas disappeared from his site. Durn those 'minority' cameras, eh?

If you fiddle with it, you will find some newer cameras for which the function is linear. I believe that is what is misleadingly called "ISO-less" cameras.

Out of interest Dan, since I own four ISO-less cameras, could you expand on "misleadingly" please?

[DanK]

Out of interest Dan, since I own four ISO-less cameras, could you expand on "misleadingly" please?

I don't know anything about Sigma cameras, but look at the graph for the Nikon D800, which I believe is one of the cameras often labeled as "ISO-less". The disadvantage of higher ISO is the same as with, say, the Canon 5DIII (my camera): raising ISO decreases DR. There is no magic in this. For the same reason, higher ISO will mean more noise in many photos. The difference is simply the lack of nonlinearity in the left-hand end of the curve. The linearity at the left end is handy--it means that you can amplify the signal either in camera or in post, with no major effect of the choice, but it doesn't alter the fact that raising ISO has its costs.

[Manfred M]

Out of interest Dan, since I own four ISO-less cameras, could you expand on "misleadingly" please?

Misleading only in so far as your Sigma cameras are very much outliers in terms of the technology they use. I don't think members should have any issues at all couching their answers in terms that are relevant to the main stream users on this site. I don't see a lot of people mentioning film cameras (those are truly ISO invariant, unless you look at "push-processing"), the Leica Monochrom line (no Bayer array), medium format cameras (larger sensors), etc. when making postings. I own three working film cameras; perhaps I should be offended whenever someone forgets to mention film. I shot one of them last week, so its not like I haven't used them recently.

I would tend to suggest all cameras are really"ISO-less". What we refer to as ISO is nothing more than an engineering / marketing kludge. All sensors are going to only have a single sensitivity and that will be related to the design of the sensor itself; how effectively it measures the photons that enter it. Anything else is related to how that data is processed. Most of what we look at as ISO is basically little more than amplifying the data which gives the illusion of higher sensitivity.

I find it interesting that when I Google ISO-invariance; Sigma does not even warrant a footnote...

https://improvephotography.com/34818/iso-invariance/

[DanK]

This:

Most of what we look at as ISO is basically little more than amplifying the data which gives the illusion of higher sensitivity.

That's really the core of the answer to the OP's question. There an be lots of reasons to amplify the signal, several of which have been noted above, but image quality is never one of them. (Unless you want noise in the photo--but then again, you can always add that in post, with more control.) Raising ISO will at best leave image quality unchanged and more often will degrade it, depending on factors mentioned about.

Even though I don't own one, I can see some advantage to the characteristics of the misnamed ISO-less bodies. In principle, they allow you to ignore underexposure and to correct for it in post with results very similar to those one gets by boosting ISO. However, for my work, that is not an advantage worth giving much weight to.

[xpatUSA]

Misleading only in so far as your Sigma cameras are very much outliers in terms of the technology they use. I don't think members should have any issues at all couching their answers in terms that are relevant to the main stream users on this site. I don't see a lot of people mentioning film cameras (those are truly ISO invariant, unless you look at "push-processing"), the Leica Monochrom line (no Bayer array), medium format cameras (larger sensors), etc. when making postings. I own three working film cameras; perhaps I should be offended whenever someone forgets to mention film. I shot one of them last week, so its not like I haven't used them recently.

I was not at all offended and asked a genuine enough question which was answered satisfactorily by Dan in post #33.

I would tend to suggest all cameras are really"ISO-less". What we refer to as ISO is nothing more than an engineering / marketing kludge. All sensors are going to only have a single sensitivity

If only it didn't say "All"

www.photonstophotos.net/Aptina/DR-Pix_WhitePaper.pdf

and that will be related to the design of the sensor itself; how effectively it measures the photons that enter it. Anything else is related to how that data is processed. Most of what we look at as ISO is basically little more than amplifying the data which gives the illusion of higher sensitivity.

Nothing new there. Unfortunately, your much-vaunted majority doesn't think that way.

My personal view is, similarly to yours, that ISO-sensitivity should never have been applied to digital cameras. I've always like the idea of a single knob, marked IQ and going backward from 1 to some small number like 0.

I find it interesting that when I Google ISO-invariance; Sigma does not even warrant a footnote...

https://improvephotography.com/34818/iso-invariance/

Is that, by any chance, an intentionally demeaning comment Manfred?

[Manfred M]

Nothing new there. Unfortunately, your much-vaunted majority doesn't think that way.

My personal view is, similarly to yours, that ISO-sensitivity should never have been applied to digital cameras. I've always like the idea of a single knob, marked IQ and going backward from 1 to some small number like 0.

When it comes to ISO, camera manufacturers were selling to photographers, so ISO would have been in their comfort zone. With my Panasonic video camera, I have something called a "Gain Switch", which is a far more accurate assessment as to what it does.

Is that, by any chance, an intentionally demeaning comment, Manfred?

No Ted - it just goes along with my comment that most people know Sigma makes lenses, and some fine ones too. Their cameras are very much a niche product that most people are likely unaware of. You have a habit of commenting on the uniqueness of your cameras, and to me that is somewhat irrelevant other when there is something that is handled differently than one other makes.

I did follow the development of the Foveon sensor and was disappointed that its performance characteristics were never resolved to the point where it could compete with the Bayer array sensors. The Foveon sensor has inherent advantages over the Bayer (no need for Bayer or AA filter and associated demosaicing in raw conversion).

I'm just saying that Sigma cameras are not on most people's radar, as you have pointed out many times yourself.

[xpatUSA]

When it comes to ISO, camera manufacturers were selling to photographers, so ISO would have been in their comfort zone.

Yes, I am well aware of that.

With my Panasonic video camera, I have something called a "Gain Switch", which is a far more accurate assessment as to what it does.

Excellent!

No Ted - it just goes along with my comment that most people know Sigma makes lenses, and some fine ones too. Their cameras are very much a niche product that most people are likely unaware of. You have a habit of commenting on the uniqueness of your cameras, and to me that is somewhat irrelevant other [than] when there is something that is handled differently than on other makes. <>

I'm just saying that Sigma cameras are not on most people's radar, as you have pointed out many times yourself.

Equally so.

In future, I will try hard not to post comment when I feel that some Sigma property might be of interest to some people - now that you have made it quite clear (on behalf of "most people") that they are likely unaware and, by extension, uninterested.

[pnodrog]

In future, I will try hard not to post comment when I feel that some Sigma property might be of interest to some people - now that you have made it quite clear (on behalf of "most people") that they are likely unaware and, by extension, uninterested.

While you are at it Ted I would appreciate it if you would arrange a block on any reference to any camera that uses the terms Tv or Av in the mode selection - it is just too confusing and of no interest or relevance to me (and most people?.... ) Thanking you in advance.

[dje]

While you are at it Ted I would appreciate it if you would arrange a block on any reference to any camera that uses the terms Tv or Av in the mode selection - it is just too confusing and of no interest or relevance to me (and most people?.... ) Thanking you in advance.

Ah spoken like a true Nikonian. Thank you bro

[pnodrog]

Ah spoken like a true Nikonian. Thank you bro

Thank you for your support bro. Mind you being my bro makes you ineligible to stand for Parliament.

[JBW]

Has any CiC member done his or her own testing to see if in "normal day to day shooting" there is any perceivable difference between say, ISO 100, ISO 200 etc. In other words, rather that reading some interesting testing results, have you actually SEEN the difference in the size images you use?

Absolutely. In real life as soon as I move off ISO 100 my IQ starts dropping.