Is DX superior to FX?

michael66

Nope - the DOF becomes wider (increases/deeper) as the crop factor increases. You can play with the Calculations here

Well, that site says the opposite. For a DX camera, using a 50mm lens set at f/16, the DOF is 9 ft. For an FX camera, same lens and f-stop, the DOF is 17.2 ft. Or do you mean that the crop factor increases as you keep more of the image; going from DX up to FX? But my brain says all that is settled solely with the lens as the image hits the focal plane. The sensor just keeps more or less of the image. Sigh. Serious confusion. I'll have to figure out a way to test this for myself to get it to make sense for me. For now, I'll just take what you are saying on faith and move on. Thanks.

manhattanboy

here
Well, that site says the opposite. For a DX camera, using a 50mm lens set at f/16, the DOF is 9 ft. For an FX camera, same lens and f-stop, the DOF is 17.2 ft.

I am not sure exactly what he meant, but hopefully I can help with the confusion. The comparison is for same FOV i.e. a 100mm DX versus 150mm FX, which would give the same image view when printed. At those distances, the DX has more DOF than the FX does at the equivalent focal length. Some people like this for the FX camera as if you are taking up close portraits, then you can use a longer lens than you could on a DX camera, and as a result you have the ability to throw the background into greater blur from the more shallow DOF. Some people prefer the opposite, however, where the DX has increased DOF at an equivalent focal length allow more of the target to be in focus in case the focus is not 100% spot on.

heartyfisher

As for me I now have the best of both worlds I have an FX and a DX camera ... and if you have 1 prime lens. On the Dx camera it will have shallower DOF.

spraynpray

I actually think the DX/FX thing is really another horses for courses question. I like the greater working distance and greater depth of field of any given focal length on DX for macro, along with the crop factor 'apparent magnification' when shooting telephoto. In good light, there is a smaller and decreasing difference between them for normal photography and in low light the FX sensor has an advantage too although that is less than it was now.

If I could afford it, I would have a D610 and my D7100 but I don't like the nasty surprise of the extra megapixels making my lenses look soft and my computer look slow. I hope it stays at 24mp for the foreseeable future.

@PB_PM: you need to get over your problem with me. I couldn't use more emoticons in my previous posts without it looking ridiculous. You are in danger of creating another TTJ/Ade situation.

Ade

@TTJ, "That is actually the same difference. The inner part of the image is projected by the inner part of the lens. "

No, I claim that is incorrect. All parts of the FOV use all parts of the lens.

Hmm, I've been thinking about this -- I don't know if all parts of the FOV use all parts of the lens.

But I believe @BabaGanoush is generally correct that the outer parts of a lens do contribute to the inner parts of the image. So, if we mount an FX lens on a DX camera, the DX sensor will still "use" or capture light from "all" parts of the FX lens -- even though the field of view is restricted. This is something I've never thought of before but appears to be correct.

Intuitively we know this must be true because aperture controls on an FX lens work normally when mounted on a DX camera. If a DX sensor only captures light from the "inner part of the lens" then the aperture would be ineffective until it is sufficiently stopped down to reach the "inner part".

For a slightly more technical analysis, in optics a "marginal ray" is can be defined as a ray of light reflecting from the subject (at the center of the optical axis), then hitting the very edge of the (FX) lens, before striking the image plane, again at the center of the optical axis -- which is also the middle of the sensor be it FX or DX.

So even the very middle of a DX image would have some contribution from the very edge of an FX lens. Or in other words, the entire lens is used.

(I'd draw a diagram if there's interest, but it's past my bedtime now)

spraynpray

Interesting @Ade: - so how does your post relate (if at all) to the often quoted point that DX bodies use the centre or 'sweet-spot' of an FX lens?

Ade

It may be a bit of a misnomer. I think @BabaGanoush has it right here; it's more accurate to say that DX sensors uses the center or 'sweet-spot' of the image circle, but even that reduced image circle area is formed by contributions from the entire lens.

TaoTeJared

Nope - the DOF becomes wider (increases/deeper) as the crop factor increases. You can play with the Calculations here

Well, that site says the opposite. For a DX camera, using a 50mm lens set at f/16, the DOF is 9 ft. For an FX camera, same lens and f-stop, the DOF is 17.2 ft. Or do you mean that the crop factor increases as you keep more of the image; going from DX up to FX? But my brain says all that is settled solely with the lens as the image hits the focal plane. The sensor just keeps more or less of the image. Sigh. Serious confusion. I'll have to figure out a way to test this for myself to get it to make sense for me. For now, I'll just take what you are saying on faith and move on. Thanks.

Mashed two thoughts together - I shouldn't answer the phone when posting. You are correct the DOF DECREASES becomes narrower on FX and increases (wider) on DX as the website shows.

spraynpray

It may be a bit of a misnomer. I think @BabaGanoush has it right here; it's more accurate to say that DX sensors uses the center or 'sweet-spot' of the image circle, but even that reduced image circle area is formed by contributions from the entire lens.

Blimey. Idununnerstan all that. @-)

snakebunk

@Ade: There is interest for your diagram. I have been thinking about the DX uses all of the lense theory and I can not make it happen within my limited optical knowledge.

Edit: I have read though that you can not make a tele lense smaller just because it is to be used for DX cameras only (like you can with wide angel or normal lenses). And I have read that the diameter of the first glass of a tele lense must be at least the focal length divided by the maximum aperture, regardless of the sensor size.

TaoTeJared

Maybe some images will help.
Per our favorite Photography Life


A second


50mm lens on:
--FX = 46° FOV
--DX = 31°30'
Nikon calculated

The lens itself always "sees" 46°. On DX though only 31°30' of the full view is recorded - thus it only uses the center part (67%) of the lens.

GreenFlash

I think this is right, if you are talking about a lens designed for an fx camera. And in that situation you can truly call the sensor a crop sensor, because it does crop off the outside part of the image as projected by the lens.
But lenses can be designed to project images on a smaller sensor. eg all of the compact cameras with much smaller sensors. And there are getting to be more lenses specifically designed for dx cameras, i.e. aps-c sized sensors.

Msmoto

@GreenFlash

Image size will always be the same on whatever the sensor is… an example is the 10.5mm f/2.8 Nikkor… on crop sensor this is 180 degrees, corner to corner, on full frame, slightly more than 180 degrees can be seen, but at any given focus, the image size is the same no matter what is behind the lens. The lenses designed for smaller sensors do not cover a full frame, but the image size at a specific focal length is the same no matter what one places behind the lens to capture the image data.

michael66

Maybe some images will help.

The lens itself always "sees" 46°. On DX though only 31°30' of the full view is recorded - thus it only uses the center part (67%) of the lens.

I do appreciate your efforts. I hope there are others on here that are in the same ignorance boat as I am.

Yes, I see what you are saying for a compare between FX on FX and DX on DX. But you have a DX lens on a DX body. What happens when you use the same FX lens on both bodies? Holding my breath before I yell, "Eureka!"

----

I had to come back to edit this after re-reading this thread. The issue I am having is about DOF, not FOV. The difference in FOV, I understand. It is the idea of the DOF changing when using the same FX lens on a DX vs. a FX body.

donaldejose

I have long felt the best solution to the DX vs FX issue is to have one of each in the same control layout. Thus, the D600/610 should be matched to the D7100. The professional D300s replacement should be matched to the D4 or D800. You should also have a collection of complementary lenses. Then the biggest difference in switching between bodies will be the size of the viewfinder. You can use the camera best suited to the task at hand. For example, I use my D7000 rather than my D600 for macros of insects for more working distance because my Tokina 100mm FX macro lens becomes a 150mm macro lens when used with a DX body. Alternatively, one could use a D800 in its DX crop mode or just shoot it in FX and crop dramatically and I have done that but it seems wasteful of captured image. It just seems to me I should fill more of the frame with my subject when composing the shot. My personal preference would be to put 24mp on the subject filling the frame in both FX and DX. Hopefully Nikon will soon produce a D300s replacement which will allow the parings I suggest above.

sevencrossing

It is not surprising you get confused
As one factor effecting DoF is the "Circle of confusion"
http://en.wikipedia.org/wiki/Circle_of_confusion

clskelton

It is not surprising you get confused
As one factor effecting DoF is the "Circle of confusion"
http://en.wikipedia.org/wiki/Circle_of_confusion

I think I'm starting to understand this better. So when the DX image is enlarged to 8x10 size, it requires further magnification than FX because the DX sensor size is smaller. This additional magnification makes out of focus areas more apparent to the human eye and therefore depth of field appears to be shallower than FX at 8x10. This would also mean anytime we crop our images it decreases the (apparent) depth of field when enlarged to 8x10 because further magnification is required.

Am I spot on, or way off?

Ade

You are spot on, but remember that the image captured is fixed at the time the shutter button is pressed and can't be changed in post (e.g., by cropping). See the following discussion on FX vs DX PoV:

http://forum.nikonrumors.com/discussion/comment/17606#Comment_17606

michael66

@TaoTeJared, @clskelton, @sevencrossing, @Ade

Thanks guys, the light finally went on.

@msmoto, can we create a read-only ( non mods ) thread where certain items, like Ade's essay on DOF can be kept? Maybe a catch-all FAQ?

sevencrossing

Maybe a catch-all FAQ?

+1

With things like

should I buy a Dxxx or wait for the Dxxy?

Why are the guys at Nikon not listening ?

heartyfisher

@GreenFlash is right though about 1 thing.. every part of the lens effects every pixel.. unless the aperture is closed down.

as i said above.. but the "sweet spot" is in the centre bec the light travels different distances between the left side and the right side of the lens thus its hard to make that focus to the exact same point teh further off axis it comes from.

Maybe some images will help.
Per our favorite Photography Life


A second


50mm lens on:
--FX = 46° FOV
--DX = 31°30'
Nikon calculated

The lens itself always "sees" 46°. On DX though only 31°30' of the full view is recorded - thus it only uses the center part (67%) of the lens.

Actually that top diagram is not correct.. the parallel lines coming in will focus to a single point of light..

sevencrossing

Actually that top diagram is not correct.. the parallel lines coming in will focus to a single point of light..

I think they are meant to represent 8 separated points

Ade

Ok, this is what I think is happening re: @BabaGanoush's theory that "all parts" of an FX lens are used even when mounted on a DX camera.

Keep in mind that I'm not an optical engineer and I made the following illustrations simply to help me think about the process based on my understanding of ray tracing -- which may or may not be 100% correct. (Also I'm sick at home today so have a bit more free time).

The first diagram below illustrates a simplified FX lens imaging a white vertical arrow on the left, with a wide-open aperture (click the image for a larger view).


Wide open aperture -- not to scale (larger)

On the right side of the illustration is the sensor, FX in green and the smaller DX area is indicated by the grey hatching. The image of the white arrow appears as an upside-down "real image" on the sensor.

Let's construct two light rays as follows:

1. Chief Ray. A chief ray is a ray that goes through the center of the optical center. (Technically, through the center of the aperture, not the center of the lens, so my drawing is a bit off). The chief ray is colored blue in the illustration.
2. Marginal Ray. A marginal ray is a ray that emanates from the object at the optical axis, touches the margin (edge) of the aperture, before converging back to the image plane (sensor) at the optical axis. The marginal ray is colored red in the illustration.

You can see that these two rays in effect defines the upside-down image on the sensor -- i.e., the chief ray defines where the top of the arrow will appear on the sensor, while the marginal ray defines where the bottom of the arrow will appear on the sensor.

It should be apparent from this illustration that even though the upside-down image fits within the DX area, the entire FX lens was used to form the image. The Marginal Ray used the very edge of the FX lens to contribute to the center, uncropped part of the sensor. Also, the image formed on the FX sensor must be the same exact image formed on the DX sensor (since they overlap).

The second diagram illustrates what happens when the aperture is stopped down:


Stopped down aperture -- not to scale (larger)

With the lens stopped down, we have the same exact process except the Marginal Ray is shifted to touch the very edge of the aperture. We are now using a smaller part of the FX lens for both the FX and DX sensor.

How does Field of View (FoV) fit into this discussion? For the previous two illustrations, I purposely selected an image that fits within the DX sensor's field of view.


Field of View, FX vs DX -- not to scale (larger)

We can now define FoV by drawing lines from the edges of the FX/DX sensors to most restrictive edge of the lens or barrel (marked "Field stop" in the above illustration). The FX sensor has a larger FoV because geometrically it has less angular restriction vs. the DX sensor.

We can see the image circle or "cone of light" provided by an FX lens is much greater than the size of the DX sensor. So yes, a DX sensor only uses the "center" or "sweet spot" of the image circle. However, as the first two illustrations showed, the image within the DX area is still formed using the full extent of the FX lens.

Clear as mud? Open for discussion.

snakebunk

@Ade: Thank you! Great work, you should write a book I cannot say if it is all correct but it makes sense to me.

I actually miss this kind of information in the photograph books I have read so far. They move pretty fast from the basics to composition and other artistic considerations. Does anyone know of a good book about optics related to photography?

heartyfisher

Your first diagrams is accurate but doesn't illustrate the point .. the red arrows starts from the same point as the blue arrow and ends at the same point.

your second diagram is also not correct.. probably due to the incorrect conclusions derived from the first diagrams. the point of crossover for both FX and DX sensors should be in the center of the lense.