Optimum ISO for digital cameras....

[richardvallon]

For a long time I thought the best ISO was the lowest ISO... and that seems to work fine- but recently read somewhere- wish I could remember- that 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...
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.
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....

[Manfred M]

Welcome to CiC Richard (?) Would you mind clicking on the "My Profile" button on the top of this page and entering at least your first name and where you are from into your profile?

As for the mysterious 200 ISO view, I've heard it before and frankly it makes no sense.

All sensors have a fixed "ISO" and any setting above that results from amplifying the signal. The moment gain is added, the dynamic range starts dropping, the amount of digital noise in your image goes up and the colour bit depth drops. In other words, the image quality will start to deteriorate.

That being said, one should shoot at an appropriate ISO setting to get the image. Sticking with the lowest ISO and introducing motion blur from camera shake or shooting a scene with too wide an aperture to get the critical parts of the image in focus don't make a lot of sense either. The closer you are to the base ISO (i.e. the lowest numeric ISO listed for your camera), the lower the impact of adding gain.

If you want to see what the measurements are for your camera, head over to the DxOMark website and look up your camera model.

As an example, this is the graphical representation for the various measurements for my camera, a D810:

https://www.dxomark.com/Cameras/Niko...--Measurements

[NorthernFocus]

Best I can tell the urban legend of the magic ISO200 started when that was the "native"/base ISO for the popular sensors that were on the market at that time. Like many things that was repeated so often that it became a rule of thumb that took on a life of its own. In reality as Manfred pointed out the base ISO of any sensor produces the cleanest images because any other ISO setting involves modifying the signal gain. Unfortunately it isn't always easy to find information on what the base ISO is for a given camera body/sensor. For some reason the manufacturers don't seem to publish that info. We often have to wait for it to appear in a technical review of the camera.

[pnodrog]

DX0 Mark have test results that will give you a guide. Basically the ISO 200 for best performance relates mainly to Canon sensors prior to about 2010. It is no longer of much (any?) relevance. Currently the lowest standard ISO setting (50-100) seems to give the best tonal and dynamic performance.

Here is a LINK to DX0 results on several newer cameras. You can select which cameras you wish to compare. Use tabs to select option you wish to view. Under Measurements you can have a look at graphs for Dynamic Range, Tonal Range etc.

P.S. Sorry Manfred I see you have already mentioned DX0 mark. Not surprising really, it is a useful site.

[Shadowman]

Nikon calls ISO 200-1600 the normal range, I'm sure every manufacturer will have a different range and as a photographer its up to you to find that normal range just as you would optimal aperture. I personally try to shoot as low as I can but I shoot at different times of the day, at times need a specific shutter speed so I push the camera to the level where the output is most efficient. One thing I've noticed that under the right circumstances, usually when the sky is in the composition or there are deep shadows; ISO 100 still gives you some level of noise so this suggests at least to me that ISO 100 is optimum for models that can get that low.

[MrB]

.....All sensors have a fixed "ISO" and any setting above that results from amplifying the signal. The moment gain is added, the dynamic range starts dropping, the amount of digital noise in your image goes up and the colour bit depth drops. In other words, the image quality will start to deteriorate.

That being said, one should shoot at an appropriate ISO setting to get the image. Sticking with the lowest ISO and introducing motion blur from camera shake or shooting a scene with too wide an aperture to get the critical parts of the image in focus don't make a lot of sense either.....

+1

Manfred, one of my cameras is an older model (a Pentax K-r) and one of the custom settings is called 'Expanded Sensitivity'. The default ISO range is 200 to 12800, which changes to 100 to 25600 when 'Expanded Sensitivity' is enabled. Just curious - do you know what is happening at the lower end here?

Cheers.
Philip

P.S. Just seen the previous message, posted while I was typing, which is a similar point.

[billtils]

Just curious - could it be that there is an element of urban myth in this, in that in the (good?) old pre-DSLR days the best films were in the 100 to 200 ISO range?

[dje]

Philip and Richard

It is fairly common these days with a variety of cameras to have an ISO setting lower than base ISO. But it's a bit of a trick. In most cases, when this lower setting is selected, it fools the camera metering system into thinking the ISO is lower and this results in increased exposure. What this means is that you lose highlights headroom and increase the risk of clipping.

If you look at the link Manfred provided to DXO, you will see on the ISO sensitivity screen that DXO Mark messures the same ISO sensitivity for the camera settings of 32 and 64. The dynamic range is better for ISO 32 but I believe this would be due to the extra exposure “allowed”.

This link http://blog.kasson.com/the-last-word...rii-fake-isos/ to Jim Kasson's site talks about ISO in Sony cameras.

Dave

[dje]

Ted just saw your post. My understanding is that the term base ISO generally refers to the ISO sensitivity as measured by the saturation method. I also believe DXO Mark use the saturation method for their measurements.

The camera makers probably prefer the extra “flexibility” allowed with other methods as outlined in the ISO spec.

Dave

[richardvallon]

Thanks Manfred- that was a great answer- I need to look at DXO Mark to get optimum ISO - esp for finding the greatest latitude. Yep- I get appropriate ISO but wanted to know what is the best for various cameras to shoot things like flowers and architecturals from a tripod where you have the luxury of using the best possible ISO... I think you are a great teacher... I completed my profile... Did not really think about DXO Mark as although I am pretty technical I've never been into making small comparisons between cameras - just buying those that get the job done.

[richardvallon]

wonderful quote from Kuhn. - I will have to look up Saturation method. for a long time I simply thought the lowest ISO a manufacturer put in the camera had the greatest latitude. Of course shooting RAW is the way to preserve this latitude. So now I'm wondering - is the Base ISO related to latitude? BTW- years ago found a firmware Hack fro a Sigma SD 14 camera that reduced "ISO" to 50. The camera shot with incredible latitude - a parade in New Orleans on a deep blue sky day- and I also did fill flash. Images were great- but I had to color correct by color in Lightroom- asphalt was slightly green- and any small changes in exposure resulted in color changes- odd ones not across the whole exposure - but green shadows, warmer highlights... After color correcting each image they looked great... and now I know why most cameras at the time- 10 years ago - did not offer 50 ISO.

[richardvallon]

I want to thank everyone who replied to my question about ISO. I posted this question to an American based forum recently and received the answer of " Make America Great Again" - just kidding- actually received no answers... So this was more than a pleasant surprise - and i'm a bit overwhelmed and I need to do some homework. I own a Sigma SD1, Sony a9 and a trusty Canon 5dMKIII- so cannot wait to look these up and see what DXO Mark says. So I got into this question ISO - esp in regards to latitude- as I noticed with my Sony a7rII that at iso 200 I was getting more leeway in my highlight detail than at iso100... not scientific- just seat of my pants observation- that began my quest as I was part of the lowest ISO must be the BASE ISO and must be Best crowd.... wow- you guys are incredible- so nice to be in a group where I am not the one providing all the answers...

[richardvallon]

thanks so much for your informative reply... I saw your portfolio and truly admire your work. wow- what a span of continents and subjects... I posted my replies below not directly to your reply in error....

[DanK]

I think this thread may be over-complicating the answer.

You wrote:

is the Base ISO related to latitude?

I assume by latitude you mean dynamic range. Then the answer is yes. The simple answer is that you will get the best results on all dimensions, including dynamic range, at the base ISO. As Manfred pointed out, higher ISOs entail amplification, and that reduces dynamic range. To see a concrete example, there are charts here for the Canon 5DIII, which you and I both shoot, and the Nikon D800.

The sole exception to this, AFAIK, are the cameras that have an artificially low ISO setting, that is, a setting below their base. for cameras that don't, like the 5DIII, the bottom line is that if highest quality is the goal, one should raise ISO only when there is some other reason to. (For example, I shoot candids of kids at ISO 400 with TTL flash. The reason is that this increases the impact of ambient lighting. The degradation of the image at that ISO, with that lighting and that camera, isn't enough to be troubling.)

[richardvallon]

Thanks Dan- yes I mean dynamic range when I say latitude- which is possibly a very old film term...
so - all of your points are excellent- so many smart people like you here on this forum... so once question I did not ask is there is perhaps an ISO for common cameras where 100 is artifically low and the chip is actually BASE at ISO 200? Yep- I really should test my Sony cameras at ISO 50 as it seems you may be in more danger of losing highlight detail even when in RAW,,,

[rpcrowe]

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?

[xpatUSA]

Ted just saw your post. My understanding is that the term base ISO generally refers to the ISO sensitivity as measured by the saturation method.

Yes, in an old Kodak application note, it says:

"There are two basic types of ISO: saturation-based and noise-based. The saturation-based ISO is also referred to as the ‘base ISO'.

from: kronometric.org/phot/exp/ISOiawKodak.pdf

It's an old Note, the other methods introduced later by ISO (SOS, REI) are not mentioned at all.

The note is gone from the web, glad I copied it when I could.

I also believe DXO Mark use the saturation method for their measurements.

The camera makers probably prefer the extra “flexibility” allowed with other methods as outlined in the ISO spec.

Dave

Indeed so, Dave. Thus it was that Canon was allowed to reduce the headroom afforded by saturation-based, according to this well-respected Canonista:

http://kronometric.org/phot/exp/Expo...alibration.pdf

You probably know that you can get a clue of what your camera is up to by shooting a gray card or white card that fills the frame in Aperture Priority, infinite focus, in daylight, then converting to an sRGB image.

Theoretically, saturation-based ISO gives about 98/255, standard output sensitivity gives 118/255. That difference being due to about a half-step of sensor exposure.

In practice, these exact figures are rarely obtained. Quite frustrating for a pedant.

[DanK]

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?

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.

[Shadowman]

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?

With my shots the difference can be seen when the sky is included in the shot or when the subject is backlit, for my shots done between ISO 100 to 400 it just means a bit more processing is needed to clean up the photo.

[NorthernFocus]

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?

I can definitely see the differences. But only after having read an article telling me it's there

Joking aside I suppose much depends on how you define "normal day to day shooting". If you're talking SOOC jpegs used only for web or small prints, highly doubtful any difference will be evident at relatively low ISO values. Where I can see the difference is if I really push PP to recover shadow detail, pull out sky detail, etc. Then yes on any of the Nikon bodies that I shoot it's likely there would be a noticeable difference from ISO200 to 400. Certainly from 200 to 800.