Nikon FX / DX Observations and Questions

Nikon don't have a 24-70 f/4

Correct. (Dyslexic fingers on the keyboard, incomplete thoughts emanating from my brain.) My intent was to refer to the many options from various sources for lenses in the 24/28-70/85mm/2.8-4 range. A good, reliable, high IQ example in this category would fill the hole, with my current 70-300 picking up the balance of the range.

To elucidate further: I expect my D7100 will continue as my carry-around camera. My hopes for the D810 are to make outstanding landscape and architectural photographs that will stand up well to large prints, extensive cropping, and publishing (in the case of architectural subjects) as promotional materials. Therefore, I'm less worried about weight/size than I am about cost and IQ. While I prefer VR, I might do without given that most images will be made from a very stable tripod.

Understood, David. And yes, the D810 is an exceptional camera for those uses. I do claim that too much can be made of tripod requirements for the D8x0 bodies (they're less than 2x linearly sharper than, say, the D700; at worst, you should have to double the shutter speed to get an equivalently-sharp image). You shouldn't be scared off hand-holding them, and I can often get a sharp image at ISO 64. Would there be any difference from a tripod? Maybe, but only very slightly, and I'd have missed a lot of shots if I needed a tripod for all of them. I'm not beyond using a tripod when it's needed, though!

 

And, for what it's worth, the Tamron 24-70 does have "VC" vibration compensation (what Tamron call VR). I don't claim, at least in the one I have, that it's anything like as good as the 200-500's VR.

Andrew

I do NOT have a FX body, hence my point.

However, even if I had a FX body will my exposure change?

 

Let's not confuse the matter, I am primarily concerned with the correct exposure.

 

Is the exposure of a scene at a constant light level at a specified ISO level (say 200), with the same lens, the same in a DX body and FX body? Or is it different? IOW, do I need to increase the exposure on the DX body, to compensate for the crop factor? I do not have a FX body to make this test.

 

Or if I take a meter reading with a hand meter; ISO 200, f/8, 1/250 sec. Is that the same exposure for ISO 200 film in 4x5 view camera, 6x6 MF camera, 35mm camera, and a digital camera set for ISO 200 in FF/FX camera, DX camera, m43 camera.

 

Is that exposure affected by the lens. IOW is the correct exposure with a Sigma 50-100 f/1.8 at f/1.8, the same as with a Nikon 70-200 f/2.8 at f/2.8?

 

When I am shooting in low light, the first thing that I am concerned about is my exposure. Can I get a decently exposed shot? Everything else is sacrificed to get a decently exposed shot. Shooting wide open, I am not concerned with DoF. I shoot wide open at night under lights, because there isn't enough light to stop down the lens, and still maintain a decent shutter speed, without cranking up the ISO even higher. So aperture (wide open) and shutter speed are set first, ISO level is then adjusted to give me a correct exposure.

ISO specifically measures sensitivity to light per unit area. Relative aperture (absolute aperture divided by focal length) determines the rate at which light flows through the lens to a given point on the focal plane, irrespective of the sensor size behind it. Hence at the same f-stop and same sensor (or film) ISO, you get the same image by exposing the same amount of light for the same amount of time.

 

That is, to get the "correct exposure", you can ignore sensor or film area, and just worry about relative aperture, shutter speed, and ISO. This has to be the case - if you crop an image, that doesn't change the brightness of the remaining region. That applies whether you're cropping in Photoshop, whether you're cropping by using the DX section of an FX sensor, whether you're just comparing a DX sensor to an FX sensor, or whether you're using different sizes of film.

 

But ISO isn't the only thing that determines noise. If you shoot an FX camera at ISO 6400 (and produce an image from the whole sensor area without cropping), the result is - barring big changes in sensor technology - less noisy than if you do the same thing with a DX camera. Over twice as much light is hitting the FX sensor at the same relative aperture because it's more than twice as large; this makes the result less noisy. Similarly, if you shoot medium format at the same relative aperture as 35mm, you get the same exposure at the same ISO - but the medium format image will look less grainy. The film grain per unit area is about the same (indeed, the actual film may be the same), but there's more of it contributing to the image. Put another way, with a larger film area, you're magnifying the grain less to produce the final image. This advantage goes away if you crop the image. And, to an approximation, ISO 400 film looks about as grainy as ISO 100 film does if you magnify it by a factor of two (quadrupling the area). By a similar argument, we typically compare image quality of a digital sensor for the whole frame, not pixel by pixel - otherwise sensors with a higher resolution are capturing less light at each pixel, and look much noisier.

 

However, if you're shooting a larger format with a longer lens (to match the field of view of a smaller format) and you use the same relative aperture, you get less depth of field from the larger sensor than the smaller one. The same aperture on a longer lens means the physical size of the entrance aperture is larger (hence the front element of a 200mm f/2 lens is a lot bigger than a 50mm f/2, roughly by a factor of four in diameter), and this defines depth of field in the scene. If you want the same depth of field from a larger format, you need to reduce the aperture by the same factor as the crop factor.

 

Correct the aperture to match the depth of field, correct the ISO so the noise levels roughly match, and everything balances out. Let's approximate micro 4/3 sensors as having a crop factor of two compared with full frame (because the numbers are nicer than if we look at DX): use the micro 4/3 system with a 50mm lens at f/2 and ISO 100, and to a reasonable approximation you'll get the same image as a full-frame body with a 100mm lens at f/4, at ISO 400 - you get roughly the same field of view, depth of field and noise. If you don't bother about matching depth of field and shoot the full-frame body at f/2 like the micro 4/3 body, it'll have less depth of field, and because it's capturing more light it'll be less noisy than the micro 4/3 system.

 

Similarly, let's approximate 5x4 (inch) large format film as having a crop factor of 4 compared with 35mm film (that's a bit off, and the aspect ratio doesn't match anyway, but let's pick a nice round number). Put a 50mm lens on the 35mm camera and, in our approximation, you get the field of view of a 200mm lens on the 5x4. If you put ISO 100 film in both and shoot both at f/8, you'll get the same exposure from both - but the 5x4 image will be much less grainy because the grains are a lot smaller relative to the film area. It'll probably also be sharper. However, the 5x4 also has a lot less depth of field when shot this way. To match the 35mm depth of field, you need the 5x4 to be at f/64 (8x4) - hence the correlation between the 35mm guidance "f/8 and be there" and the f/64 club. At f/64 the 5x4 camera is receiving 1/16 as much light as the 35mm, so - at the same shutter speed - you need to bump up the film ISO to compensate. And ISO1600 film is about as graiiny as ISO100 film is if you enlarge it by four, so the final print of the 5x4 image with ISO1600 film will look approximately like the 35mm print using ISO100 film - all assuming that the final print is the same size.

 

Compared with FX, DX has a crop factor of 1.5. That means an FX lens captures 2.25x as much light in the same time at the same relative aperture. 1.5 is close enough to the square root of two (1.4...) that we won't be far off if we say FX bodies tend to be about a stop better in terms of noise compared with DX ones, but reducing the aperture of the FX body by a stop gets you back to roughly the DX body's depth of field.

 

So: If you care about noise or depth of field, I contend that FX and DX are roughly identical so long as you can use a lens a stop faster on DX, and shoot the FX body one stop higher. Generally, a lens with that aperture is more likely to exist for 135 format than any other, but the 50-100mm f/1.8 is an example where DX has an equivalently fast lens (though with a shorter zoom range than the FX version). So a 50-100 f/1.8 wide open will produce the same image as a 70-200 f/2.8 wide open so long as you use an appropriately higher aperture on the FX zoom - but that higher ISO matches the noise advantage you had by using the FX sensor in the first place. On the other hand, there's no DX lens that matches FX f/1.4 (or 1.2) primes, or, say, the 14-24 f/2.8 or 120-300mm f/2.8, and the 18-35 f/1.8 and 50-100 f/1.8 don't cover the full range of the 24-70 and 70-200 f/2.8 lenses.

 

So FX has an advantage if you want to trade depth of field for lower noise, for the majority of lenses that don't have a correspondingly fast DX equivalent (and some faster crop lenses do exist for formats other than DX, though they're often not very good - and there are also speed boosters). If you don't need that functionality, the DX lenses tend to be smaller and cheaper (partly because they're shorter, partly because they need less coverage, partly because we're usually comparing effectively slower DX lenses with faster FX ones once equivalence is considered) - which are perfectly acceptable reasons to choose DX.

 

At least, that's how I think about it all. Just because I can justify it doesn't make it right, let alone useful!

Thanks for raising this topic. I've had the Nikon 12-24 f.4 DX for 10 years and got good use from it with D200 and D7100. I've not used it since getting into FX. If I understand correctly, on my D850 dialed in DX format, I've got a 18-35 with 20 MP, and using full FX I get a functional 18-24 with 45.7 MP ? Is this right? I just might start using it again while I wait for the 14-24 f.2.8 replacement.

Thanks for raising this topic. I've had the Nikon 12-24 f.4 DX for 10 years and got good use from it with D200 and D7100. I've not used it since getting into FX. If I understand correctly, on my D850 dialed in DX format, I've got a 18-35 with 20 MP, and using full FX I get a functional 18-24 with 45.7 MP ? Is this right? I just might start using it again while I wait for the 14-24 f.2.8 replacement.

 

In DX mode you would see what you did on your D7100 (18-35 FX equiv).

In FX mode you see a 12-24, BUT you may get heavy vignetting because the smaller DX image circle of the lens won't cover the full FX sensor.

 

You can put a FX lens on a DX camera. Because the FX lens image circle is larger than the DX sensor.

But not a DX lens on a FX camera. Because the DX lens image circle is SMALLER than the FX sensor.

Thanks for raising this topic. I've had the Nikon 12-24 f.4 DX for 10 years and got good use from it with D200 and D7100. I've not used it since getting into FX. If I understand correctly, on my D850 dialed in DX format, I've got a 18-35 with 20 MP, and using full FX I get a functional 18-24 with 45.7 MP ? Is this right? I just might start using it again while I wait for the 14-24 f.2.8 replacement.

 

I suspect this got misplaced from the "D800 and 12-24mm lens" thread.

 

Gary: I believe the caveat is that some DX zooms, this one included, do have a image circle large enough for the FX frame over some of their range - hence the 18-24 comment.

 

Dan: I think the issue is that there's a difference between producing an image and producing a sharp image. Given how demanding the D850 is, I'd be surprised if you found the FX edges of the 12-24 acceptable at any focal length.

 

The 14-24 is possibly the Nikon lens I've owned longest (I've forgotten when I got the 28-80 and 50mm AF-D). The biggest issue I have with it is field curvature, but I usually shoot it at f/7.1 if I care about the corners. It's certainly imperfect out of the camera, but DxO cleans it up pretty well. I'm not sure when Nikon will get around to updating it (or rather, I've no guess how high it is on their priority list) - they might be a bit distracted with mirrorless glass, and I'm not sure I'd put it ahead of the 200mm macro, maybe the 200/2, the 135/2, the 50mms etc. if I were scheduling things.

 

Sigma have a fairly new 14-24 f/2.8. It's not perfect either, but I'm not sure from reviews that it's actually inferior to the Nikkor, and it's cheaper. Is that worth a look?

I'd meant to add: Dan, if you're after a placeholder for a new 14-24, I believe the Samyang 14mm is supposed to be decent (if not quite up with, say, the 15mm Zeiss). Just a thought.