#ad #affiliatelinks #earncommission - as an Amazon Associate we may earn from qualifying purchases - more info
Strap in, there's rarely a binary answer, especially when considering Low Extended ISO
V1.1 - updated 25.04.2022
I recently uploaded an image to a Sony Alpha Facebook Group and got some interesting feedback regarding my use of low ISO (ISO 50) on a modern digital camera.
There seems to be a general consensus on these groups that shooting at lower ISOs will produce more noise than when shooting at the camera's native ISO. (In this case the Native ISO is 100, as the camera is a Sony A7R III)
I’ve seen this opinion before, and wanted to look in to it further before responding…
I've now read a fair amount on this subject and I have to accept that "in the lab" this may well be true, however, anyone who has been shooting for a long time will tell you what the science says, and what the film and camera manufacturers recommend, is rarely reproduced in the field.
So, the question is, does this technical wisdom carry over in to the field, when using modern full frame sensors?
This is the shot in question: It's a lovely lifestyle shot of a model in the lotus position, in a glorious summer environment here in the UK.
I would much rather talk about the shot set up, the styling and the make up, the model and brief, but it seems that people are more interested in the technical aspects of the image.
The Sony groups are disproportionally distracted with quality triangles and ISO speeds... So, here goes... 
I shot this lifestyle image of Cody at ISO 50.
The reason I selected a low ISO is because my priority was to keep the depth of field shallow by setting an aperture of f2, and I didn’t want to go any faster with shutter speed so I lowered the ISO to 50. In addition to this I was not carrying my ND filters with me.
It would be really easy of me to say that I chose ISO 50 because I knew that this will produce the least amount of noise, but at the time I shot this, it would be untrue. “The reason I used ISO 50 was because since “forever”, lower ISO speeds produce less noise, not more.”
Cody - The Lotus Position - Richmond Park - Shot at a Low Extended ISO - ISO 50
“The reason I used ISO 50 was because since “forever”, lower ISO speeds produce less noise, not more.”
I assumed that the transition from film to digital had not made any difference to this practice. For the last 20 years it certainly hadn't for me, but, could I actually make a claim that I knew that it was better to use a lower ISO? No.
Note: the actual reason I selected ISO 50 was for convenience, and a dinosaur habit.
I further researched ISO on the world wide web...
I found out that using low or high ISO in a digital environment is now referred to as using “Extended ISO”. This term is news to me. Amazing how I've been making a very good living for the last twenty five years + and this term had got by me. I have been too busy working as aprofessional photographer to notice. 
In addition to this, and without irony, it's called “Extended ISO”, even though, by setting the camera so that effectively, it has less ISO speed (gain), one would think that it would be called “Reduced ISO” when set to ISO 50, but no, just to keep us all guessing, it’s called "Extended ISO", whether we increase the ISO or reduce the ISO!
It turns out that there is quite a debate going on about using Extended ISO, with lots of differing views.
Much of this debate is simply not well informed. I make no specific expert claims here, but it seems to me that if I want to form an opinion, it maybe prudent to try to learn what I can before making sweeping statements that have been derived from "received wisdom" experts who propagate inaccurate memes about the subject. I thought that this may assist me in validating any conclusions that I make at the end of the Article.
A little "not too technical" background regarding ISO. Note: You may already know most of this stuff, but here it is:
The difference between the way that film works, and a simplified (for me, not you!) explanation of how camera sensors work.
Back in the day, cameras themselves did not have a way of recording images. Photographers had to choose the film to put in the camera. I know this sounds a little obvious, but most of us simply don't think about this stuff.
It was the film that was speed rated, using ISO speeds, not the camera.
Fuji Velvia 100 Transparency Film
Kodak Portra 400 Negative Film
So, if you were shooting a portrait you may choose a film rated at ISO 25 to 100. Film rated at these ISO speeds were usually fined grained and had lots of contrast. Nowadays, in a digital environment, this roughly translates to "less noise and a wider dynamic range".
There was wide range of films to choose from, including whether to shoot Negatives or Transparencies (Slides), Colour or Black and White and of course lots of choice when it came to ISO (Film Speed) and in addition to this, there were films for different colour temperatures. eg. Daylight and Tungsten!
On typical SLR cameras like the Nikon F3 or the Olympus OM2, you would select the ISO/ASA/DIN* to match the film that you had chosen.
This would set the sensitivity of the light meter in the camera, which in turn would give you the correct light readings. You then set the f stop and shutter speed to take the shot, and after processing the film your exposures would be reasonably accurately represented on the film. Job Done.
*We now simply use the term ISO, and have adopted new standards that no longer use either the ASA rating or the DIN rating
You can find out about the history of speed rating here
Digital Camera Sensors do it differently...
Modern digital cameras do not use film. They have a CCD or CMOS sensor. The sensor has a “Native ISO Rating”. The camera I am shooting with is the Sony A7R Mark III, so the “Native Speed” is calibrated @ ISO100** or thereabouts. So, if we make a comparison with a film SLR camera, the film, in the modern digital camera is always ISO 100.
a modern digital sensor on a Sony camera
**In fact the Sony A7R III has two native speeds ISO100 and ISO640. The ISO640 uses dual gain function that is built in to the camera.. So the camera has a selection of 2 "film" speeds, when compared to a film camera. You could say it comes with a choice of slow film and fast film if you need to compare to the older technologies.
Dave McKeegan explained it to me like this:
"Cameras like the Sony have dual gain ISO which is basically 2 separate circuits on the sensor, a low gain, and a high gain, so when the camera goes above ISO 500, rather than continuing to boost the low gain signal any further the camera switches to the high gain circuit, which is why you get less noise at ISO 640 than at ISO 500."
It can be quite a challenge to understand how the sensors actually work and really, it’s beyond the scope of this test, but Wikipedia says this:
"An image sensor or imager is a sensor that detects and conveys information used to make an image. It does so by converting the variable attenuation of light waves (as they pass through or reflect off objects) into signals, small bursts of current that convey the information."
See what I mean? Let’s just say that the sensor replaces the film... Not enough? OK.
"The sensor records the light that strikes it, in the form of Photons. These fall upon it during the exposure. The Photons are captured by little buckets, or wells. It converts the photons in to electrons that are placed on each pixel, on the sensor."
"When the exposure is finished, the information is converted from analogue electrons to digital information, and written to a camera file. Finally the digital information is transferred to the memory card on the camera."
(that's the abridged version. There are some other intermediate stages).
If the winter nights are dragging on, there's further information to be found at these links:
Excellent Information of Digital Camera Sensors from Cambridge Colour
Great Info on sensors from Wikipedia
Dave McKeegan has got some useful technical info here on both sensors and ISO. He's actually correcting Tony Northrup (I'm a big fan), but looking beyond this, his knowledge of the modern sensors and how they work makes interesting viewing if you want some quality information on this:
Since reviewing this video Dave has added some more useful information that really does explain this in greater detail.
You can watch lots of Dave's YouTube Videos by clicking here
Amongst the all too eager to post ill informed content brigade, my research showed up the usual hoards of “Received Wisdom” YouTubers. All well-meaning, but I’m not sure some of them have ever peered through a viewfinder, let alone presented a finished image to a paying client. Many of them insist that native ISO produces the least amount of noise and what they say sounds convincing…
So, what to do? Old skool snappers like me have a simple way of finding out… we do a test!
What's the test's purpose?
A simple comparison test to compare ISO 50 to ISO 100 RAW files - To see which ISO has the least noise.
I did 3 tests:
ISO Test One @ f22
ISO Test Two @ f8
ISO Test Three @ f2.8.
I did this to eliminate any variance, due to aperture, and to cover the entire aperture range on the lens. I shot these 3 tests at both ISO 50 and ISO 100
ISO Test One was shot when the sun was up and the sky was blue.
ISO Test Two & Three were shot later in the day, when the sun was on the horizon.
By shooting the tests under different conditions it allowed me accommodate the wider apertures on ISO Test Two and Three. It also varied the sky, and the light that falls on the building. This eliminates conditions on the day as the reason for noise.
I did not move the tripod, or use the camera between tests.
The test is conducted in calm, dry conditions.
Camera: Sony A7R Mark III
Lens: Sony FE 70-200 GM f2.8
Sony Electronic Cable Release
Tripod: Manfrotto 028B Pro Tripod with Manfrotto 808RC4 Pro head
The Test Images
The image below is the full test image. I have marked the area where I have expanded all of the other images, This expansion has been set to 400%.
The reason I chose this area is because it's easier to see the noise in the shadows. However, it doesn't matter where you look across the entire image. The noise is clear to see. I chose 400% otherwise you won't see the difference on these compressed JPG images. Every test has a downloadable file. These are the original .ARW RAW files. So that you can use your own environment and draw your own conclusions. I have added my comments and observations under every image.
Test One - ISO100 f22 @ 1/40s
My Comments and Observations:
Remember to look for random noise, not uniform pixels. The noise in the shadow cast by the window frame is clear to see,
Now compare this to the noise on the ISO 50 image on the next test below...
_CAP7259
Note: The downloadable file is the original Sony RAW file in .ARW format. It is archived into a zip file format for easy download.
This will be repeated for every test image.
Test One - ISO50 f22 @ 1/20s
My Comments and Observations:
Look for random noise, not uniform pixels.
The noise in the shadow, cast by the window frame, is significantly reduced in this image when compared to the ISO 100 shot above it.
_CAP7258
Note: The downloadable file is the original Sony RAW file in .ARW format. It is archived into a zip file format for easy download.
This will be repeated for every test image.
Test Two - ISO 100 f8 @ 1/25s
My Comments and Observations:
Look for random noise, not uniform pixels.
The noise in the shadow, cast by the window frame, is clear to see, Now check the noise on the ISO 50 image and file below...
_CAP7263
Note: The downloadable file is the original Sony RAW file in .ARW format. It is archived into a zip file format for easy download.
This will be repeated for every test image.
Test Two - ISO 50 f8 @ 1/13s
My Comments and Observations:
Look for random noise, not uniform pixels.
Once again, the noise in the shadow cast by the window frame is significantly reduced in this image when compared to the ISO 100 shot above it.
_CAP7262
Note: The downloadable file is the original Sony RAW file in .ARW format. It is archived into a zip file format for easy download.
This will be repeated for every test image.
Test Three - ISO100 f2.8 @ 1/160s
My Comments and Observations:
Look for random noise, not uniform pixels.
The noise in the shadow cast by the window frame is significantly reduced in this image when compared to the ISO 100 shot above it.
_CAP7261
Note: The downloadable file is the original Sony RAW file in .ARW format. It is archived into a zip file format for easy download.
This will be repeated for every test image.
Test Results:
When viewed in Adobe Camera Raw 13.1 the results above are consistent and crystal clear, in both the highlight, and in the blue sky on test 1 (f22), and in the shadows areas of all of three tests (f22, f8 & f2.8).
"Setting the camera to ISO 50 renders “significantly” less noise than when the camera is set to its native ISO 100"
It’s the same whether I accept the Adobe applied default noise reduction in ACR, or manually remove it. I have not applied any unsharp masking during this test. I have simply left all of the RAW import settings at the Adobe defaults, including the default sharpen settings.
FYI: The import is using the Sony Lens Profile in the Optics panel:
"In this instance, and I believe in almost every instance, it is simply incorrect to state that a modern, full frame digital camera has less noise when shot at their Native ISOs. It’s certainly not true in the case of RAW files produced by the Sony A7R III"
This phenomenon may well have been a fact during the long development of full frame digital sensors, but it's no longer true today.
Why is this? - Well, I don’t know?
I repeat, I make no claim to expert status here. In fact, I simply do not have a definitive answer. However, I’m pretty sure that the RAW files speak for themselves?
Please download them and run them through your RAW processing workflow to see for yourself in your work environment.
What do others say?
Others are saying that it is not the ISO setting that causes the noise, Native ISO or not. They are saying that when closely observed, it’s the shutter speed that is effectively causing the noise.
Why would that be? - What's the difference?
All three tests require the frames shot at ISO 50 to be exposed for twice as long at the ISO100 frames. This means that the sensor has twice the amount of time to interpret and record the data, culminating in greater accuracy, and less noise. That seems to make sense to me?
What do you think?
Tony Northrup explains it here:
I do have another thought. Although I have no evidence to support it. It's just a flight of fancy. I wonder if Sony are applying some noise reduction before, during, or after writing the RAW file, when shot at ISO 50?
If my research is correct. When the camera is set to ISO 50 it simply shoots at ISO 100, but indicates that it is set to ISO50 and adds metadata to the file that tells the RAW development software to open up one stop brighter. So effectively, it is a "fake" setting. If this is the case then the setting the ISO to 50 itself has no effect on the the noise at all!
Maybe a "real" expert can enlighten us mere mortals?
There's some interesting reading at DxO regarding shooting at lower ISOs here
I recommend that you test it for yourself, so that you can see this in the real-world. Your real world. This test is something that everyone of us can do in 15 mins.
A sting in the tail?
“He hasn’t considered highlight clipping!”
It’s well known that shooting at a lower ISO will add more recoverable detail in the shadows and remove recoverable detail from the highlights. Once again, common thinking here is that the highlights may blow out if you are not careful with your exposure. We're told to keep an eye on the histogram to stop clipping in the highlights.
Of course, this is not in dispute, however in an “in the field" test this concern may be considerably overstated?
Is it something that we would ever actually encounter, when shooting RAW files, with a modern, full frame, digital camera?
The dynamic range of these sensors, and the subsequent RAW files are so wide, (around 14.7 stops from black to white for my camera) I would really have to over expose the image to produce a file that would come anywhere near seeing this problem.
Even with images that have bright highlights in the frame. And, how many photgraphers even print their work today? Most photographers are now wedded to the sRGB environment of the world wide web, and for the time being, this means viewing images that have a maximum dynamic range of around 8 stops. That's 8 stops from black to white from a RAW file that has 14.7 stops of available tonal range. Now, dependant on where your exposure falls within the 14.7 stops, it's unlikely you'll be 3 stops over or under exposed isn't it?
And whilst it has nothing to do with image quality, most photographers will end up processing an sRGB JPG that is limited to displaying in an 8 stop environment. It's unlikely these photographers would ever see clipping on their shots.
With the launch of the Sony A1 there's even more Dynamic Range to fall back on in the future, as that technology trickles down to new cameras in the range.
In addition to this, I'm not sure that it is possible to even see this "clipping" on anything but the most esoteric monitor, in a controlled dark room environment? Certainly not on any reflective, paper based medium, which can be averaged out to having around 6 stops dynamic range.
I guess if I were shooting a pure white cat, on freshly fallen snow, outside a white house, in flat light, it may be prudent to use a higher, or native ISO, and keep an eye on the histogram during the shoot. Of course, there will be times when due consideration should be given to this issue, like when shooting into the sun, or dark environments with bright spot lights. Attention may be needed to under expose a little, but the RAW file has to be so far out to cause an actual issue of clipping.
FYI: I gave zero consideration to highlight clipping when I shot the test files above. Yet I can recover so much detail in the highlights and white areas simply because the tools at our disposal are so advanced, and the RAW files have so much information recorded within them from the sensor capture. I'm not saying that a very bright light will not burn out to white, clipping the image highlights, but I'm not sure I've ever seen a shot with a bright white light that didn't burn out to white and clip the image and that looked bad. That's what a specular highlight is isn't it, a burned out highlight?
Below is an examplar: Both of the images have clipped highlights. I have highlighted the clipped areas with red circles.
The image on the left reads R255 G255 B255 in the foam on the coffee and in the key lights in her eyes.
The image on the right clips on the brightest part of the gold necklace at R255 G255 B255.
both of these image have clipped highlights, circled in red
Does the coffee look any less like coffee because of the clipped highlights? And does it not look like a gold necklace because of the clipped highlights?
Is it possible that the issue of highlight clipping may be theoretically accurate but visually insignificant in many cases?
Me personally, I am not prioritising my attention on clipping when shooting RAW on my Sony A7R III. There are a hundred more important things on my mind...
I'm thinking about composition, I'm looking at the "the look". I'm checking the clothes, Making sure the hair falls correctly. Checking that the assistant has the reflector in the best position to lose the nose shadow. Anguishing over the choice of make up. Making sure the model is having a good time and making sure the client can get the wifi password from the location we are at. I'm not sure I've ever worried about highlight clipping since I started using digital cameras? This also applies to most other genres, with very few exceptions, Astrophotography being one of them where I believe highlight detail is uber critical. But, with all respects to the Astro guys and girls the Milky Way was not top of the list when shooting these terrestrial shots.
Does anyone have a good example of shooting ISO 50 and it causing a problem with highlights? Clearly, despite my words, this must be a issue for some shooters. It would be great to place an image right here to see the other side of this coin?
Let me know if you have appropriate image and associated story that illustrates it?
If the winter nights are still dragging, further information can be found at this link:
Some great info about dynamic range here from Cambridge Colour
Using the "speed" metaphor thoughout this article.
Newbies and Intermediate shooters may not realise that ISO is no longer really about "speed" as it was when we used to shoot film. Technically it's now about GAIN. I have used the term "speed" throughout this article, but of course, this is technically an archaic term, however ubiquitous.
However, as a metaphor, or framework, it still works, and when applied, will be accurate in most cases.
In the 21st century ISO should technically be measured in, and thought of as electronic gain, similar to a volume control on an amplifier.
So, when we select a higher ISO we "turn up the volume on the sensor", or, we "boost the signal" and this increases the exposure, and the associated electronic noise that the camera produces. Once again, this can reasonably be compared to a radio signal being boosted "after" it reaches your radio. If you turn up the volume it will be louder, but you will also hear the pops and crackles louder. This is audio noise. Noise on the image is electronic noise produced by the camera. At this time speed and gain can both reasonably be used and interchanged in much the same way we use "noisy" and "grainy". Clearly there are no film "grains" in a digital image, and no digital image has ever shown a grainy look. they show a noisy look, but we understand what it being said and it's not cool to correct people when we perfectly understand what they are saying.
My Conclusion
I would think it’s time to put to bed the notion that shooting at a camera's native ISO will always give the best results.
If you are shooting on a modern, full frame digital camera, and you are looking for ultimate image quality... and why would we be try to achieve anything else?
The notion is simply not true.
I would conclude that it’s best to think old skool. I acknowledge that this may not have been true when digital was first introduced, but the sensor manufacturers have been very busy, and times have changed, and the ideas that we espouse should adapt too.
The old adage that less ISO means less noise has made a come back. (maybe, with a little help from longer exposures )
A final word on highlight clipping: I would guess that this is already such a small problem for the vast majority of us, and easily side stepped by most modern, high resolution, high dynamic range sensors and RAW file processing. I suspect within a year or two, it will simply be consigned to the museum of photographic challenges, helped along its way by further sensor improvements, new in camera image processors, software and firmware improvements. It's going exactly the same way that film and Polaroids and Kodak themselves went. And, cameras without multi autofocus options, and even the way optical viewfinders and DSLRs are going, as we speak. Some of us will resist (I am not exempt from this. I just sold my entire Hasselblad 503cw kit only last year!) but we all know for whom the bell tolls...
APS-C and JPG shooters... That's another story altogether.
If you have some thoughts or experience with this, or would like to comment please do so below.
I hope the test and information is useful for some of you - Colin
related products on Amazon - curated by Colin Anthony Photography
0 Comments