Every guide you find with camera ISO explained in the title tells you the same two things. Higher ISO gives you a brighter picture. Higher ISO gives you more noise. Both are true, and neither one answers the question you actually walked in with, which is how high your particular camera can go before the picture stops being worth keeping.
The spec sheet looks like it answers that. It does not. The top ISO number on a spec sheet is the end of a slider, not a score, and two cameras from the same manufacturer prove it in a way that is hard to argue with.
Two Nikon bodies, and the cheaper one has the higher ceiling
Take the D7500 and the D850. Both Nikon, both DSLRs, both still in plenty of working bags.
Nikon's published specification for the D7500 gives a DX sensor of 20.9 million effective pixels and an ISO range of 100 to 51,200, with expansion down to Lo-1 at an ISO 50 equivalent and up to Hi-5 at an ISO 1,640,000 equivalent.
Nikon's published specification for the D850 gives an FX full-frame sensor of 45.7 million effective pixels and an ISO range of 64 to 25,600, with expansion down to an ISO 32 equivalent and up to an ISO 102,400 equivalent.
So the D850, Nikon's high-resolution full-frame body, the one that cost a great deal more and that working photographers bought for exactly the jobs where image quality is the product, publishes a native top ISO that is half the cheaper camera's.
Nobody at Nikon made a mistake. The two ranges are pointed at different ends of the scale on purpose, and once you see why, you can never read that line on a spec sheet the same way again.
Camera ISO explained in one sentence
ISO is not sensitivity. It is how hard the camera amplifies what the sensor already collected.
The sensor gathers a quantity of light. That quantity is set by three things and ISO is not one of them: how wide the aperture was, how long the shutter stayed open, and how much light was there in the first place. ISO arrives afterwards and turns up the volume on whatever came in.
Turning up the volume turns up everything, including the random fluctuation in the signal that is always there. That fluctuation is the noise. It was in the file at ISO 100 too. At ISO 100 there was so much more picture than fluctuation that you never saw it.
Which means the real question is never how high the number goes. It is how much light your sensor collected before the amplifier got involved.
Why the expensive camera has the lower number
The D850's range starts at 64, and goes down to an ISO 32 equivalent. The D7500's starts at 100. That bottom end is the part people skim past and it is where the money went.
A base ISO below 100 means the sensor can take in more light before any pixel saturates. More light in means more highlight headroom, cleaner tones in the shadows, and more room to pull a file around afterwards. On a 45.7 million pixel sensor, where the whole point is that you can crop hard or print large, that bottom end is worth more than any amount of top end. Nikon built the range where that camera is good.
The D7500 goes the other way. Fewer, larger pixels on a smaller sensor, aimed at people shooting sport and wildlife and indoor events, where the shutter speed is not negotiable and the light is whatever the room gives you. A high native ceiling is the useful end for that job.
Two different ranges, two different jobs, one spec line that makes them look like a league table.
The expansion numbers are not ISO settings at all
Hi-5 on the D7500 is published as an ISO 1,640,000 equivalent. That word is doing a lot of work and it is Nikon's word, not ours.
Those settings are not sensitivities the sensor achieves. They are the camera applying more amplification after the fact, which is close to what you would get by lifting the exposure slider on a raw file at home, except baked in. Nikon says so plainly in its own documentation. The online manual for the Z50II, a current body with a native range of 100 to 51,200 and extended settings up to an ISO 204,800 equivalent, warns that the higher the sensitivity the more likely the image is to be affected by noise in the form of randomly spaced bright pixels, fog or lines, and that pictures taken at the Hi settings are particularly prone to exactly that.
Read a spec sheet with that in mind and a lot of impressive numbers go quiet. When you see a huge figure, look for whether it is inside the native range or past it. If a camera's native range ends at 51,200 and the headline says 1,640,000, the headline is five stops of marketing.
The entry-level body that matches the flagship on paper
Here is the comparison that finishes the argument off.
Nikon's published specification for the D3500, the entry-level DSLR at the bottom of its lineup, lists a 24.2 million pixel DX sensor and an ISO range of 100 to 25,600 in 1 EV steps, with no expansion offered at either end.
25,600. The same published top number as the D850.
Nobody who has used both cameras believes those two produce the same file at ISO 12,800. They do not. But on the one line of the spec sheet that people use to compare low-light ability, they tie. If a number can rank the entry body level with the high-resolution flagship, that number is not ranking anything.
We made a related point when we compared the D3500 and the D5600, where two bodies one generation apart differ by far less than the spec sheets imply and by quite a lot in the places the spec sheets do not cover.
Find your own ceiling, because nobody can publish it for you
Your usable top ISO is not a property of the camera on its own. It depends on what you do with the file: a picture posted at 1,000 pixels wide forgives noise that the same picture printed at 16 inches will not, and a frame you intend to crop to half its width is being asked to perform a stop or so worse than the same frame uncropped.
So measure it yourself. Twenty minutes, once, and you never guess again.
- Pick a scene at home with a dark corner in it and something with fine detail, a bookshelf or a patterned fabric. The shadows are where noise shows first, not the bright parts.
- Put the camera on a table or a tripod. Shoot the same frame at every full stop from your base ISO up to the top of the native range. Raw, not JPEG, so the camera's own noise reduction is not deciding for you.
- Do not look at them on the camera screen. Everything looks fine on a three inch screen.
- On a computer, open them at 100 percent and look at the darkest part of the frame first. Then look at the fine detail. Noise is the obvious failure; the quieter one is detail that has been smeared away.
- Find the last frame you would be happy to deliver. Then find the last one you would accept if the alternative was no photograph at all. Write both numbers down.
Two numbers, not one. The first is your working ceiling. The second is your emergency ceiling, and it exists because a noisy photograph of something that happened beats a clean photograph of nothing.
Put the number to work
The number is only useful if the camera knows it. Set your maximum sensitivity in Auto ISO to your working ceiling and the camera will never quietly wander past it in a dark room. That setting, and the minimum shutter speed that goes with it, is the whole subject of our guide to setting Auto ISO properly, and it is the single change that turns ISO from something you worry about into something you have already decided.
One more thing worth saying out loud, because high ISO gets blamed for it constantly. A blurry indoor photograph is usually not an ISO problem. It is a shutter speed problem wearing an ISO costume, and we took that apart in why indoor photos come out blurry. Grain is survivable. Motion blur is not.
How to read the ISO line from now on
Four habits, and they take about ten seconds each.
- Read the bottom of the range before the top. A base of 64 tells you about the sensor's headroom. A base of 100 is normal. A base of 200 tells you the camera has been pushed somewhere else.
- Check whether the top number is native or extended. If the manufacturer calls it an equivalent, it is not a sensitivity.
- Compare sensor size before you compare ISO numbers. Collected light is the thing that actually decides this, and sensor area is most of collected light.
- Ignore the ISO line entirely between two cameras in the same family. It will tell you less than the autofocus, the viewfinder and the buffer will.
If you want somewhere to start before you have run your own test, our chart of camera settings by scenario gives a sensible ISO for the situations most people are standing in when they start searching for this.
And the thing holding the camera while you decide
That photograph is one of ours, taken on a real shoot rather than against a backdrop, and it is what the low-light argument looks like in practice: a long lens, a hand under the barrel, and a camera that has to be steady because the shutter speed is already as slow as it dares go. Everything in this article about not needing a higher ISO ends in the same place, which is holding the camera still.
The HiiGuy camera strap is the one in that frame. It comes in two variations and sizing is the only difference between them: the Regular adjusts up to 32 in on the body and the XXL adjusts up to 44 in. Both are heavy-duty polyester webbing with a wide padded shoulder pad, a stainless steel quick release and a safety tether sitting behind it, so the body comes off in one movement when you put it on a tripod and stop arguing with ISO altogether. HiiGuy has made these in Los Angeles since 2016 and backs them for three full years. See it on Amazon.
The short version
The top ISO on a spec sheet is the end of a range, and the manufacturer chose which end of that range to spend its engineering on. A cheaper camera can publish a higher ceiling than a better one and both numbers can be honest. Anything labeled an equivalent is amplification, not sensitivity. Your own ceiling is two numbers you find in twenty minutes at a kitchen table, and once you have them, you put the first one into Auto ISO and get on with photographing something.
