A few days ago, camera industry guru Tony Northrup published a video arguing that in the age of digital photography, ISO is effectively meaningless and that it’s no different from dragging the exposure slider in Lightroom. Photographer Dave McKeegan has offered a response and argues that Northrup’s logic is completely wrong.
Like many others responding to Northrup’s video, McKeegan’s point hinges on the fact that the camera, in processing the signal from the sensor, is multiplying that data before it is converted from analog data into digital data. This is what Fstoppers' own Lee Morris was suggesting might be happening when performing his own tests last week, albeit without knowing the science behind it.
In effect, the exposure slider in Lightroom is dealing with completely different information than a camera’s ISO setting, thus creating a different outcome. As one of the comments on Northrup’s original video observes, sensor signal is to raw as raw is to JPEG. Essentially, exposing incorrectly and relying on editing software is definitely not recommended as a means of exposing your digital images, and adjusting your camera’s ISO setting is a better option.
If you’re interested in the technological aspects, be sure to watch all of McKeegan’s video. Beyond that, if you’re still keen to know more, you might want to deep dive into the comments on both videos. You will almost certainly want to check out the comments made in response to the Fstoppers article, paying particular attention to informed contributions from community members Gary Gray and Paul Gosselin.
Whatever the outcome of this discussion, it’s useful to have an awareness of how cameras and editing software deals with information differently, as well as having an insight into how ISO functions as an industry standard, albeit with various hangovers from the film era.
So now we know it's boosted analog in camera and digital in Lightroom
Last part was almost worse then anything.
What about expecting a camera will give correct exposure if you use a light meter and adjust camera accordingly?
ISO have on purpose, define the sensetivety of the media exposed to, so we can expose correctly. I don't care how Kodak did it, I don't care how it's done today, but I can not understand why Fuji underexpose 3/4 of a stop. And is not about t or f stop or shutter being off a little. I am with Tony on that issue.
I think this discussion is intended to make a New mirrorless Camera without ISO control.
It always seems like Tony enjoys taking positions that are controversial (shooting jpg vs RAW???). Is he just a contrarian trying to stir the pot or uninformed?
I'm a skeptical person and I like testing things that other people take for granted. FWIW there's no part in my original video that I feel the need to retract. The arguments against it seem to be, "Look, this camera isn't ISO invariant!" but I never said all cameras were ISO invariant, just that some were.
That’s a fair point, and an attitude that probably serves you (and the rest of us) well. Thanks for the reply, and I hope I didn’t come off as too flippant!
That's a weak argument. It's the same weak shitty rhetoric conservative use to stop immigration "I didn't say they are ALL criminals, just some of them"...
So the total information of your video was that some camera are iso invariant and some aren't? not exactly a big take away...
Mr Blah Wow, you are so unnecessarily angry in your response over camera ISO. Useless response.
Well a useless video should get the response it deserves.
Stoking fake contreversy than backtracking after realizing he said some bullshit just for clicks isn't exactly useful.
Tony,
Here’s my completely unscientific test but I think it’s a practical one in the context of your video.
Using an a7riii, I took one shot at 1/500, f/1.8, ISO 32000 (highest ISO without expanded range), then took the same shot at ISO 100. That’s a little over an 8 stop difference. Let’s set aside the appearance and quality of noise for a minute. The goal of this test is, if what you say is true and ISO has little to no effect on what the camera captures, in theory I should never be able to clip any detail with ISO alone. I should always be able to recover the same amount of detail in an image regardless of what ISO I use.
Now, the trouble with this test is that the dynamic range of this camera exceeds the 8 stops I can span with ISO alone, plus Lightroom only lets me boost the exposure by 5 stops. BUT, even then, the ISO 100 image showed significant color banding in the areas that register as “clipped”, which to me indicates that ISO is actually capable of doing something in the recording process.
Interested to hear your thoughts.
No camera is completely ISO invariant. Best practice is to expose correctly unless you need to underexpose to preserve highlights. But shooting everything at ISO 100? Just no.
"ISO is totally FAKE. Seriously." totally gives that nuanced "not all cameras are invariant" position. Stop backpedaling, you stir shit up with your half-truths in titles, skim over the important parts and then just claim that you "never said x". You knew what you were doing; it's manipulative and evil-spirited - I'd never want to work with you or a person like you. Yikes.
Yep, Tony is a contrarian. He's like the video version of Ken Rockwell and tends to say things that get attention. More attention is better, even if it's based on erroneous info and half-truths. It's a proven formula.
Tony should be more careful choosing the title.
It's a clickbait tactic.
Yes, 2 jpg pictures shoot with identical speed and aperture are the same internally (on a camera without dual iso). The RAW file is also identical in any iso settings (again for a given aperture and speed)
But that's not news. Anyone who is using RAW knows that from day one.
Nothing extraordinary here.
That do not mean that iso is useless and should be ignored.
He take the problem in reverse to create traffic.
I still don’t understand what I said in my video that was so wrong. It seems like I was right but the D850 was still incredible at boosting iso in post which made Tony’s claim seem legitimate.
The results you showed after pushing 5 stops was little to zero difference in noise. I’m not sure what people are talking about.
I think both Tony's and your video were honeypots for pedants, unintentionally of course. Simplifying technical details always generates excessive correction regardless of fundamental truth. That's just internet economics. Sure the noise comparison looked slightly different due to some of the interesting science behind it but hard to reasonably draw any conclusion other than "don't sweat Iso that much while shooting" just like we've all accepted you don't have to sweat white balance.
I've noticed that content that suggests photographers can or should rely on technology for a simpler shooting experience brings out a very angry dissenting mob that still think full manual is the answer to every shooting situation and if you didn't start on film you're not a real photographer.
Very well said.
I watched your video and I thought you did a good job. I didn't see that it showed any mistakes in my original video. I'm glad when people double-check things. I specifically said only some cameras were ISO invariant (a pretty minor point in my original video), and you showed a camera that was almost ISO invariant (though I think LR might not process underexposed photos the same; adjusting color noise reduction might have equalized your example shots better).
What I thought was interesting was that so many people i the comments thought you busted my myth... people just look at the title and thumbnail and jump to whatever conclusion they want... just like they will for this article.
Tony (and Lee), this might be a separate issue but do you think the results would look different processing in the camera maker's software? I've heard some people say that since they developed the raw format their raw converters will be the best at rendering them, particularly in edge cases like this. Might Sony's Imaging Edge or Nikons' Capture NX-D have given different results?
Ha just reading this comment in the airport...you are going to LOVE our next video coming out. :)
On your last point: perhaps creators should stop making such misleading thumbnails and video titles? It would have been way easier, if not as profitable, to make a title and thumbnail which conveyed that they were testing your argument instead it conveyed they were "myth-busting" you. Myth-busting is essentially a synonym for debunking.
Some would read that as deliberately creating drama for views and engagement.
fucking hell...
So you KNOW people will jump to conclusion from the title and still use titles like "ISO is totally fake. Seriously."???
the double thinking is strong with you. anything for clicks right?
Tony used a Nikon camera for his test, it must have been the D850. If you do this test with any Canon camera the results will not be the same. There are sensors and senators and each one will bring different results to this approach.
You didn't had to do the test yourself, dpreview have this done and available for many cameras, it's called ISO-Invariance test. You can see the test for the D850 here: https://www.dpreview.com/reviews/image-comparison/fullscreen?attr134_0=n...
Now change to a Canon camera to see the difference.
I think the main issue I have is that a lot of the conclusions being made about the practical uses of ISO (not even getting into the scientific inner workings of it) are the result of focusing on the wrong thing (no pun intended).
I think the problem with saying that ISO is inconsequential is that the tests being done all fall within a narrow range of light and within the dynamic range of the camera (which, for cameras like the D850 and a7riii, is larger than the ISO range) and the comparisons are using noise alone as the benchmark. But I think in order to truly prove whether ISO means anything we should be talking about detail, not noise.
In most practical scenarios, a scene does not fall within 1-2 stops of light between highlights and shadows. In most cases, a “proper exposure” doesn’t mean that everything in the scene lands directly in the middle of the histogram, it means that you’re capturing enough data to retain detail across the scene. The problem with saying “you can just shoot at any ISO and not care” is that when you’re talking about retaining detail in shadow areas, it does have an effect. The darker parts of a scene may fall towards the bottom of the histogram, even if the rest of the scene is “properly exposed”, but adding a stop or 2 via ISO may be the difference in retaining detail or clipping the extreme areas of the scene.
Since that does actually make a difference, I believe it’s impractical to say ISO doesn’t matter.
You mentioned that there was some extra noise, but the results were close. Which is accurate.
What we need is a disclaimer when Tony Northrup is right.
Well, I'll wait until all of you figure it out and then you will tell me. But so far my camera is working pretty good.
It's one of those things that make me want to say - So
Tony clearly states that video is not for beginners in the opening of the video.
I never said that all cameras were ISO invariant... I just said that some newer cameras were ISO invariant, and then showed an example of one. Showing that some cameras aren't ISO invariant is totally consistent with the content of my video. The fact that some cameras are ISO invariant was, I think, point #3 of 3, and fairly minor.
You didn't go into much detail when you mentioned it. You just skimmed over the point and focused on backing up your hypothesis (which is good). But it does lead to people making a fuss because you didn't spend half the video talking about that point.
"Try and please everyone and no one will like it."
The video is 100% accurate, but the clickbait title got people confused
Clickbait titles are his standard procedure.
Your title isn't at all nuanced as you make it out to be. You claimed that all ISO is fake.
stop back pedaling and own up to your lack of nuance and clickbait title...
I'm an academic. If you do a double-blind randomized study and find an effect in a subset of the data (for example, epithelial cells stop dividing after 70 divisions), you don't write an article claiming all the data follows that rule, even if it's only in the title (CELLS STOP DIVIDING AFTER 70 DIVISIONS). While it's technically true, it's also not. It applies to a small subset of cells and if you pulled of that shit in a serious setting, you'd be out of a job in no time. But hey, lucky you, you're not an engineer, but a youtube person. Just remember that.
So maybe you can pull that shit off for the general public (people are morons on average), but you should know that there are more knowledgeable people than you that look at your stuff and - laugh. Yes, they laugh. Your reputation in high-end circles (and I can vouch for that because I'm part of them) is laughable. You're a meme, Tony.
The pedantry and snobbery coming off this comment is astounding.
I understand the arguments but it doesn't change the fundamentals of digital photography and taking a well exposed photograph.
Devils Advocate: It kind of does when you realize that this is no longer standardized and that the in-camera meter is lying to you to exaggerate manufacturers' claims. Trying to take a well-exposed photograph with a meter that isn't set to any actual standard is like measuring things with a ruler that uses the manufacturer's proprietary definition of inches or centimeters. Of course you'll probably adapt to it as long as you keep using the same ruler just like you'll learn to compensate for your camera, but it's stupid for there to be variance across manufacturers in the first place. ISO 400 should mean ISO 400 whether you're shooting Canon, Nikon, Sony, Fuji, Leica, or a Mickey Mouse Camera.
The key word is "should". The fact is that it doesn't. I come from the telecommunications world. Most service providers built their networks by clustering equipment from the same vendor in network nodes. This was true even though the equipment from all vendors was built to the same standard specifications. It was too much risk mixing equipment because individual manufacturers handled signals differently. I believe that standards organizations try to get things roughly right, but there will always be a lack of precision.
This is true with any electronics really. Just look at ram. Typically if you want 32 gigs of ram for your system you buy a 32 gig kit since if you buy 16 gigs now and grab another 16 gigs later you can't guarantee they will play nice together even when you buy the same brand and series.
The difference on a modern sensor is close enough that it doesn't matter all that much unless you're planning to blow up a print to some absurd size and look at it at nose distance. Nobody is going to watch Tony's video and suddenly go out and start purposely underexposing their images by 5 stops either so i think we can all stop getting our panties in a twist over minor details.
Tony used a Nikon camera for his test, it must have been the D850. If you do this test with any Canon camera the results will not be the same. There are sensors and sensors and each one will bring different results to this approach.
I more liked the "senators" one!
Can we please, please, please, stop using those obnoxious facepalm expressions in video thumbnails? If you want to challenge the guy's argument, fine, but do you need to be so patronizing about it?
Sounds to me like we each need to calibrate our own equipment and our own method of processing the images--like back in the Zone days. At least cameras these days keep their own exposure notes.
Maybe just go out and make images instead of videos about something so unimportant. Man, I sound grouchy.
Watched both and learned from both. Very interesting and educational. I am definitely going to do a test on all my cameras and see for myself.
Bottom line. Does it really matter?
Yes? Carry on discussing.
No? Carry on making images.
I’m off to do a job. Invariant variant boosted non boosted ISO, or whatever.
BOOM!
Great video. Simple explanation of a question I've pondered many times. Essentially he's describing signal to noise ratio. In the audio world clipping your signal is like clipping highlights - you can never get them back. Does it make sense to record at a conservative gain and boost later if necessary (not all amplifiers or RAW converters being created equal)? yes. Does it make sense to record a unnecessarily low signal and in turn amplify all the noise incurred along the signal path? No. A 24 bit 192k audio recording has a much lower signal to noise ratio than tape which allows you to cleanly record very low signals. Digital cameras today also have very low signal to noise ratios that allow us to not only shoot in the dark but also give us a comfortable cushion that allows us to preserve the highlights and boost the brightness without any noticeable degradation in image quality. Good times.