ISO Invariance: What it is, and which cameras are ISO-less

ISO invariance is an exciting new frontier in photography, and I've been experimenting with it extensively over the last few days to see if it's a technique that I can add to my photography tool belt (Fine… I admit it.  I've never even worn a tool belt.  I'm the worst handyman on the planet.)

In essence, ISO invariance allows the photographer to entirely skip over the ISO when setting the exposure and never changing it, but still being able to take full advantage of the low light performance of the camera.  It's still a brand new concept (the first mention I could find of it online was in 2011), and really nobody started talking about it until DPReview started publishing more information about it 6 months ago.

What is ISO Invariance (An ISO-less Camera)?

The short answer is that ISO invariance means that a camera will produce the exact same image quality by staying at ISO (or whatever the base ISO is on the camera) and dramatically underexposing the photo and then brightening it up again in Lightroom, as if you had shot the camera at the proper ISO in the first place.

It makes your head spin for a minute, but don't worry.  We'll go through lots of examples and it'll make sense shortly.

Let's suppose you want to go out and shoot a photo of the stars.  As a starting place, suppose we set our camera settings to f/2.8, 20″ shutter speed, and ISO 3200.  Those are pretty standard settings for night photography, and they'll likely produce a nice image.  But what if you kept the shutter speed and aperture the same, but dropped your ISO to 100?  The resulting photo is dramatically underexposed, but if you take that underexposed file into Lightroom and bring up the exposure, it suddenly looks identical to the photo shot at ISO 3200.  Increasing the exposure in Lightroom introduces noise, but just the same amount of noise as the high ISO would have introduced.

Thus, we say this camera is ISO invariant (“invariant” means something that remains unchanged despite different circumstances), or sometimes photographers say the camera is ISO-less.

iso-invariance-test-small

What Cameras Are ISO Invariant?

First of all, a disclaimer.  Just because a camera fits in the “ISO invariant” category below doesn't mean it's precisely the same thing to shoot at ISO 100 and brighten vs shooting at a high ISO.  There are slight differences in how the files look with some benefits and drawbacks to each.  What we're focused on here is the levels of noise and reasonable dynamic range and highlight/shadow detail.

So you can't PROVE that any camera fits in any of these camps, but I made my best judgment call given the test results we got after I polled the Improve Photography Podcast listeners and asked them to send me their raw file tests using many many different cameras.

Cameras that are ISO Invariant

  • Sony A7RII (Much better highlight detail from shooting at base and brightening later, but lose a a slight amount of shadow detail.  I might even dare say that noise is handled JUST A TINY TINY bit BETTER on the brightened picture than on the high-ISO shot.)
  • Fuji XT1 (This is my personal camera.  I switched from Nikon to Fuji.  It's probably the most iso-less camera out of all those that I tested.)
  • Fuji X100
  • Fuji XE1
  • Nikon D810 (Relying on data from DPReview.  The Sony A7R uses the same sensor, so I would ASSUME that it is as well.)
  • Nikon D750 (Only did one test with this, but appears to be entirely ISO invariant.  Would like to test more)
  • Nikon D7100 (Tested only at base vs ISO 800, but the noise pattern is identical)
  • Nikon D5500 (I did not personally test this one.  Relying on data from DPReview)
  • Pentax K5 (At ISO 800 vs base ISO, you can't tell any difference.  Very high ISOs not tested)

Cameras that are Somewhat ISO Invariant

  • Olympus OMD-EM1 (Tough call.  Detail and contrast are definitely lost when brightening in post, but noise appears to be reduced quite a bit on the brightened image.  I'd like to do more testing.)
  • Olympus OMD-EM5 II (Difference is indistinguishable when zoomed out, but when you zoom in, the higher ISO photo is VERY VERY slightly better in terms of noise and contrast.  The 40mp mode brings the contest even closer.)
  • Sony A7S (I was interested to see this one.  The noise pattern on the brightened image is close to the high ISO shot.  However, the brightened image lost a SURPRISING amount of contrast).
  • Sony Nex 7 (This one is really close to being ISO invariant.  The noise is about the same, but contrast is lost on the brightened image.  Very close to being ISO-less.)

Cameras that are NOT ISO Invariant

  • Canon 5D Mark III (Not even CLOSE!  Nick Page tested this one for us and it looks really bad when you shoot low and brighten later.)
  • Canon 6D (Not even close, and the camera did a horrible job of selecting the white balance in the under-exposed shot.)
  • Canon 70D and Canon 60D (Not too bad, but it's still much better to shoot at the higher ISO.  Horrible white balance in the underexposed shot.)
  • Canon 7D (Not nearly as bad as the 5DIII, and you can't tell the difference with the naked eye at ISO1600, but when you zoom in it's obvious that the higher ISO shot is cleaner.  White balance not good in the underexposed photo.)

Thank you to the following Improve Photography readers from all around the world who contributed raw files for this test: Dale Mellor, Nick Page, Darin Mellor, Linda Maier, A Zelkanovic, Ryan Fritsch, Tristan Davies, Benjamin Boynton, Julio Debeux, Jenn Dijk, Robert Connor, Stephen Tillman, Sarah Scully, Najib Mahmud, Mark Franks, Clarence Hagler, Nelson Tapias, and Wesley York. 

iso-dial-fuji-xt1

The ISO dial on the Fuji XT1, which is probably the most ISO invariant camera we tested.

Why Would Anyone Want to Use This?

There are lots and lots of reasons why this would be a fantastic technique in certain situations.  It's not something I'd use every day, but it's certainly a technique I will implement in my night photography.

One of the reasons is that by brightening the exposure in post, you can selectively brighten some areas and leave others darker.  This reduces the noise in some areas of the image, whereas if you set the ISO high in camera, it makes all of the photo have the higher amount of noise.

Another reason why you'd want to shoot an image at base ISO and brighten later is when you are concerned about preserving highlight detail.  Suppose you're in Time Square in New York and you want to take a shot of the colorful lights above the streets at night.  If you shoot at a high ISO to properly expose, there isn't enough dynamic range to properly preserve highlights and still expose the shadows. However, if you do this after the fact, you could easily brighten the image without blowing out highlight detail in the bright signs.

Another reason is when shooting night photography.  It can be difficult to judge how much noise is present in a photo when you're just looking at the tiny preview on the LCD of the camera.  Many times I've ruined night photography shoots because I pushed the ISO a little too high for the camera.  If I would have simply shot the whole night at ISO 100 and brightened later, this would not have happened.

The limitation to shooting at base ISO and brightening later, however, is that you'll see a black preview on the back of your LCD all night.  Makes it kinda tough to judge the composition!  So I plan on flipping back and forth.  Use high ISOs on my Fuji XT1 for getting the composition, and then once I have the shot looking how I want, I'll shoot one at base ISO (which is 200 on the Fuji XT1 but 100 on most cameras) and keep that one for the official shot.

What ISO Invariance (ISO-less Shooting) Is NOT

ISO invariance is not a measure of how good a particular camera or sensor is.  It simply means that a lot of the exposure is generated with the image processor rather than in the sensor.  It's simply a different approach to low light photography.

So while I do think ISO-less shooting is an exciting new technique for low light photographers (doesn't make a lick of difference for daytime photography), it's certainly not worth picking a camera simply based on its ISO-invariance.

 


About the Author

Jim Harmer

Facebook Twitter Google+

Jim Harmer is the founder of Improve Photography, and host of the popular Improve Photography Podcast. More than a million photographers follow him on social media, and he has been listed at #35 in rankings of the most popular photographers in the world. Jim travels the world to shoot with readers of Improve Photography in his series of free photography workshops. See his portfolio here.

Comments

  1. Great article. Looks like my email to you was too late for the Sony A6000. Sorry. Maybe next time.

    1. I also am interested in the a6000 but it should be even better than the NEX-7 as that is the camera that replace it (not to some fanboys, but technically it did to Sony).

  2. Fuji’s hight ISO performance is a big lie because if you have a Canon or Nikon camera and compare the settings you will see that ISO 3200 on a Fuji X series is actually ISO 2000 on Nikon or Canon. It’s marketing.

    1. Author

      @Andy – While that may be true, it doesn’t really fit in this discussion. As I said in the article, ISO invariance doesn’t mean that a camera is some sort of low light miracle. It just indicates how the ISO works in the camera.

      Independent of how well or poor a camera does in low light, this is about ISO invariance.

    2. Great point Andy. Definitely deceives the unknowing photographer, which might be quite a few of us!

    3. X-T2’s ISO calibration is based on Standard Output Sensitivity (SOS) while the cameras from Canon, Nikon and Sony use a more liberal ISO standard named Recommended Exposure Index (REI). The two measures are not directly comparable. It is worth to study these methods before making judgement.

  3. Jim, you’ve once again made something that some may seem as complex and made it SUPER easy to understand. The example of the city lights is an excellent example. Great article!

  4. I shot the moon last night with my X-E1, and while I didn’t do much in the way of testing this theory, I did find that images shot at ISO 200 and “push processed” in Lightroom were a good bit noisier than those shot at higher ISOs. My test wasn’t scientific, and as I will mention in my blog post today, I went about the whole shoot “wrong,” but I think my point here is that it’s a “your mileage may vary” kind of thing.

    1. I own a Sony A7RII and it’s not ISO invariant. ISO 400 has noticable noise that ISO 100 does not. I don’t shoot above ISO 6400 since the noise gets really bad, but a photo shot at ISO 6400 looks much better than a photo shot at ISO 100 with the exposure boosted in post in low light situations. My low light photos would look considerably noisier if I used ISO 100 and underexposed.

      That said, the high ISO performance on the camera is excellent compared to other cameras I own.

      1. AFIK the A7RII is dual ISO. Is has two gains, one used for “native” ISO 640 and above, the other for anything below ISO 640 (native ISO 100).

  5. I have for years fought the noise vs. underexposure battle when shooting in my church. Entry-level crop-sensor camera and inexpensive glass I really struggled. I needed a high shutter speed and in low light, that meant increasing the heck out of my ISO. I hated the noise so much that I gave up and settled for underexposed photos that I lightened up in post as opposed to the noise from my sensor.

    Neither of my cameras are listed above; but this is how I’ve been shooting for years in certain circumstances just out of necessity.

  6. Very interesting, as Brian mentioned, you summed up a potentially complex topic really well. Kudos.

  7. So what I read here is that the ISO capabilities in the Canon 5D Mark III are superior compared to what the “ISO Invariant” cameras are. And it’s better at processing low light information than Lighteoom is too.

    1. No the Canon 5D Mark 3 is not superior. I have both the 5DMk3 and the X-E2. They are different cameras that process noise differently. The Fuji noise is more true to film granularity while Canon’s is more close to electronic noise. None is better or worse than the other one. Nevertheless, if I underexpose with the Fuji I can push it afterwards in post WITHOUT significant noise addition. That doesn’t work with the Canon – you either get the exposure correctly when you shoot or you’re toast – not that much recovery without adding noise. High ISO noise levels are similar. The full frame sensor in the Canon has more pixels and slighly better resolution than the Fuji.

  8. What setting in Lightroom increases the exposure across the whole range of values? Exposure mostly increases midtone values whereas increasing your camera’s ISO should increase all values. It doesn’t seem possible to replicate the camera’s exposure adjustments exactly in Lightroom.

  9. Does it mean that it is easier to fix an underexposed photo with iso invariance camera file?

  10. Jim, Great read and something I have discovered as well. Obviously no where near to this level as I have one camera D7100 and not a dozen or more lol. Once I figured this out especially useful when shooting birds at f/8 to get a super crisp image @400mm on a min aperture 5.6 lens. Even in daylight conditions shooting at the proper ISO to expose the sky correctly then bringing up the bird in post is much easier than trying to bring back blown out areas of the sky or doing an entire sky replacement. So there are uses for it in Daylight as well, or you just make a high key looking image of a properly exposed bird. Thanks for sharing

  11. Very interesting especially as I have a Nion D7100. I do mostly studio portrait work but I’m often challenged to get enough light (as I use continuous lighting) and push my ISO to get an in-camera acceptable exposure. Since I generally am more concerned about Aperature and use Ap-Priority Auto, I guess this technique isn’t going to work for me though – since without pushing the ISO up, the camera is going to use longer exposures? I’d have to go fully manual right?

  12. It should be possible to tell for certain just graphing dynamic range vs. ISO. An ISO invarient camera will basically always sample the sensor output exactly the same in hardware. Most cameras are limited to 14-bit ADC (analog to digital converter), and as such, a 14-bit/ 14 EV DR is the most you’d expect possible out of a raw image, assuming of true raw image, no software messing with things.

    Traditionally, at least in higher end cameras (though I’ve personally seen thus same arrangement in cellphone sensors), there’s a variable gain amplifier between the sensor array and the ADC. Obviously, this is useful if you have reason to expect more than 14-bits of resolution out of the sensors (or you’re using a 12-bit ADC, as some do). For consumer P&S, maybe not so much. A JPEG is inherently 8-bit… if I capture a 12-bit raw at ISO100, I still have a full 8-bit image at ISO1600. You might include an amplifier and a cheaper 10-bit ADC, but no real need otherwuse in a cheap camera.

    For raw and a very large sensor, you’d certainly hope for a much wider range of overall ooeration than justca 14-bit DR across the whole ISO range. Though it’s true most systems are going to do some software processing, at least to extend range. In fact, that’s what we call it.. extended ISO range. There’s no hard definition for that. It migt be that the rating doesn’t meet the ISO:2006 standards, coukd be the manufacturer just doesn’t see mainstream quality at those ISO numbers. But most, like my 6D, offer a low ISO that’s actually below the hardware base ISO. You can onky get there via software magic… and you’ll see the 1 EV loss of DR in that mode, too, when looking at raw images.

  13. I’m just an on-again-off-again hobbyist, so there’s one crucial caveat I don’t understand here. Taking an extremely low ISO photo in low-light conditions and bumping it 5 EV in post will leave the entire photo, as-shot, in shadows! Doesn’t this mean detail in shadows aren’t picked up by the sensor, the exact same way details in highlights are lost when your exposure is too high??

    I’m failing to understand why you can bump EV 5 stops UP and recover shadow detail, but you cannot conversely bump EV 5 stops DOWN to recover highlight detail. How can the camera capture adequate recoverable detail in very very dark shadows when it cannot do so in very very bright highlights?

    1. Author

      @Douglas – Good question.

      The reason you can recover five stops UP is because cameras that are ISO invariant aren’t really using the sensor to become more sensitive to light. They are just taking a picture and then using the info from the sensor to expose the shot in the PROCESSOR of the camera. So it’s really just doing the same thing you’d be doing in Photoshop or Lightroom when brightening.

      By shooting at the base ISO, you’re telling the camera not to do the brightening for you.

      However, when a camera uses the SENSOR to increase the exposure (not the processor afterward), the brightness is baked into the shot the instant it’s recorded. That’s how most cameras work, and it means they are NOT ISO invariant.

      So that’s why you can go up but not down.

    2. Good info here regaridng why you can push shadows but not pull highlights as much, as said previous, sensors have more data in the higher end of the histogram (highlights) and less data in the lower end of the histogram (shadows) so you want to preserve the mid tones and highlights vs preserving below the midtone for the shadows.

  14. You reported on the Canon 7D. Is that the 7D Mark II? That’s what I am interested in…

  15. All last week I took advantage of this method heavily. I traveled up to Pictured Rocks National Lakeshore in Michigan’s Upper Peninsula and used the new EM5 Hi-res shot at base ISO in every situation. This eliminated the need for an ND filter as well.

    I also underexposed for panoramic shots, finding that I didn’t need any of my bracket and HDR captures. The more I use this camera, I’m finding that I can worry more about everything else, and not the ISO. Amazing. Every situation is different, but for most situations, I now “set it and forget it.”

  16. This does not appear to be an actual ability, just a tolerance of working on the extreme of the sensors dynamic range. For a fully representative range of exposure/detail in an image you are giving up a lot at whichever end you are clipping off to salvage some form of exposure in the final image. I am very suspicious that it is anything else but accepting the lowest level of quality in the final image.

  17. On the Fuji x-e1 (I imagine this applies to all x-cameras), go to set-up menu 3 and change “Preview Exp. in Manual Mode” to OFF. This will brighten the display making it possible to compose under very dark conditions. Keep in mind that this display is no indication of the actual exposure which is normally displayed with the setting to ON.

  18. Hello everyone!
    The question “recovery of Shadows” is one of the issues that most passionate and light up discussions between Canon and Nikon users.
    The ability to recover underexposed shots seem a considerable comfort and a clear advantage for photographers who operate in poor lighting conditions, but how far is it?

    Dpreview on, I played a little ‘with the various combinations ISO / recovery for the latest SLR on the market.
    Well, if you look at the comparisons to 100-200 ISO the situation Canon is nothing short of embarrassing recover 4-5 stop leads to results dramatically affected by electronic noise.
    I compared then two comparable models of the two houses, as the D750 and 6D at higher ISO. Raising the ISO recovery can obviously lower, but the differences are very thin, to be virtually indistinguishable between 800 and 1600 ISO. Even at ISO 3200 with the recovery of a stop 6D Canon 750D seems better.

    What does that mean? What if the light conditions are good and I can use medium or low ISO, Nikon allows me to take almost without looking at the meter, while Canon wants more precise exposure; But by operating in lower lighting conditions, when we have to necessarily use at least ISO 1600 for shutter speeds fast enough, the differences are canceled and we will have to take virtually carefully exposure even with the Nikon.

    I am deductions that I did, but I might have missed something.

  19. OK. I’ve been using an X-E2 for about a year. Never heard of ISOless before today. What I’ve found myself doing is choosing a convenient ISO according to lighting conditions to get a starting point, then then rotating the exposure compensation dial while watching the histogram to achieve an ETTR solution. (Sometimes choosing to go a little further to the right if blowing out some highlights makes sense.) Then take the raw file home to LR and twiddle the sliders (B&W, Sh&Hl, Exp&Cont) until it come right.

    I think this just means “exploiting the ISOless capabilities of the sensor” without resorting to a technobabble essay. 😊😊

  20. There was a question from September last year whether Fuji X-T10 was ISO invariant. Cant´t find an answer herein, just curious because I own one and thought that this article was really interesting.

    Thanks for an answer in advance on the subject.

    Lars

  21. I just tested for the D7100, ISO100 vs ISO800 in my living room with the curtains closed. Check the comparison at the link:

    https://dl.dropboxusercontent.com/u/18460539/iso%20100%20vs%20800.PNG

    Clearly the ISO100 overexposed image shows banding, a slight color shift and a bit more noise. IMHO this method should be avoided on the D7100, and when highlight clipping is an issue, underexpose only the necessary amount to prevent it. On D7100, the underexposure results on banding on the final corrected image.

    Cheers
    Thanks

    1. I have the D600, D7100, X-T1 (and a few more). The D7100 has a fair amount of lattitude, but isn’t completely ISO invariant. The reason I believe is that it’s a Toshiba sensor. The D600 and X-T1 sensors are Sony, although the X-T1 has a different color filter in front of it (X-Trans). I think you’ll find in general that Sony sensors (at least APS-C and greater) are ISO invariant. Canon doesn’t use Sony sensors. Pentax also uses Sony sensors. (I never ran the test with any of the 1″ sensors. My RX100 unfortunately does not reliably extend retract lens anymore–it did have a full life though–and I don’t think I’ve tested my FZ1000 yet, but I’m willing to bet it passes.

  22. Can i make a comment and correction. Yes i can and yes i will.
    All digital camera’s are ISO less. ISO is the sensitivity of a film emulsion.
    On digital camera’s there is no sensitivity changing when turning the ISO dial.
    The dial should have been called GAIN!! It is an amplifier that amplifies the signal and noise levels.
    It is not ISO. ISO does not exist in a digital camera.
    The differences between cameras you will see when doing this is because of differences in signal to noise ration.
    A cheaper cam might have a worse signal to noise ratio. If the signal is amplified also the noise is amplified. If there is more noise then there will be more noise amplified and thus more noise visible in the image. That is logical. So you can do this with any old digital cam.

  23. I have the a7RII and was wondering about this a while back before Inread your post. I disagree a bit. I did tons of pics to prove to myself and I have to say there is definitely more noise when I increase exposure in LR vs if I increased ISO in camera. Now, granted this was across 5 or higher steps.

    I absolutely agree that in 1-4 or so stop ranges, it is really really hard to tell the difference. It’ll be much more important to make sure you don’t blow the highlight and take a chance with possible tiny more noise when LR when increasing exposure.

  24. Thanks for a very interesting article Jim.
    Does this mean there is no longer a need for “exposing to the right” (ETTR) on such a camera?
    Does it also mean there is no effect on the dynamic range of the image?
    Thanks again for all your tips and advice.

  25. I agree with Maz. I have the A7RM2 and can confirm that noise level is lower on higher ISO then after increasing exposure in Photoshop. Usually I use ISO at 1 stop down against camera’s normal settings.

  26. No mention of the Canon 7D Mark II. I am comparing cameras for a possible purchase soon and would like to know if it is ISO invariant.

  27. In photography you can quickly tell the artist brains from the engineer dominated brains. People ,it is a simple thorny just go test it then use it or lose it!

  28. It seems to me that ISO “invariance” is really a misnomer. Even in your discussion you point out that there are differences between images shot at the base ISO and pushed, vs. shot at the higher ISO. Those 4 stops of latitude are really just an example of higher dynamic range in the Sony sensor compared to Canon or some of the others.

  29. Tested a canon 1D X Mark II. Plus 5 stops in lightroom has the two images looking almost identical (probably plus another third stop or so would match them). You can have the files.

  30. Nice to see your open uncensored forum.

    What makes you believe the Sony A7RII is ISO invariant? It’s far from it. I HAVE seen a Nikon crop sensor camera that actually IS ISO invariant but I assure you after 8 months of use that the A7RII is NOT ISO invariant. I wish it was but it’s not. Test it if you don’t believe me. I’m not interested in an argument. But I’m an electrical engineer as well as an amateur photographer and I’m very certain you’re information is incorrect. This camera certainly allows me to recover shadow detail much better than my Canon 6D due to the high dynamic range. But increasing the ISO in low noise situations creates FAR less noise than shooting at a lower ISO, underexposing the photo and attempting to increase the exposure in post using Lightroom. I can provide examples.

    Respectfully submitted,

    Stephen Cole

    stephencolephoto@gmail.com

  31. dpreview agrees with you about the A7RII but looking at the examples here the ISO 6400 image looks FAR better than ISO 100. My own experiments show it’s much worse. I shoot only uncompressed RAW. I also do astrophotography. Forgetting to “pump up the ISO” will ruin my photos. I DO consistently underexpose with this camera by 1 to 2 stops, but I can’t shoot at base ISO 5 stops below proper exposure and get anything even close to using a higher ISO. I’m “intimate” with this camera. The author of the dpreview article is wrong IMO. That’s all I have. Cheers.

    https://www.dpreview.com/articles/7450523388/sony-alpha-7r-ii-real-world-iso-invariance-study

  32. Poor Andy is a Cannon fanboy and they cannot accept that Canon sensors have been left in the dust by everyone else over the last ten years.

Leave a Comment