DOF FX vs. DX?

Paperman

All of you are right - just expressions are different ...

DaveyJ

I have been shooting fish UW for 55 years. We have used a lot of different approaches to get good UW images, both stills and video. Currently Go Pro Black is used mush of the time when a dive or free dive starts until finished. The best Nikon set up we have used is with the 12-24DX Nikkor lens.

I have used everything from Nikonos IV to Nikonos RS (a bad leaker which ended Nikon's UW 35mm cameras) to Living Stream units with still large format to movie cameras and the whole rear of the units were custom built to keep from reflecting back into the rig.

The recently introduced Ikelite for the Nikon D5200 is about the most single practical unit I know of in the DSLR world. I see now reason to use say a D800 UW as the cost of the housing and lens, and camera added to the very real hazard of the rig flooding would rule out that approach. I stay 33 ft from the surface or less as I despise "color correction" of deeper shots. The deeper you get the more the pressures get very great and everything is mostly shades of black and blue. Artificial lighting will swing it back to looking "normal" if that is what you call it.

I always look for the side view as the most important shots of a fish and depth of field is pretty important when you are close and a fish is long. Somewhere around 10mm to 12mm lens on DX cameras is quite nice for still getting the whole fish in focus. Caution here is I have lost 8 cameras to this kind of exposure and I have never had similar loses with land cameras.

Sato

Because of this:

Use the actual focal length of the lens for depth of field calculations. The calculator will automatically adjust for any "focal length multiplier" or "field of view crop" for the selected camera.

When you change the sensor size the calculator changes the (calculated) focal length based on the new sensor size.
Therefore the depth of field changes.

------------------------

As far as I understand it: FX gives you shallower depth of field when using your lens at its widest aperture (f/1.8), But as the diffraction limit is further away (Compared to DX) you can get about the same max depth of field as you get on DX before the diffraction limit kicks in.
Is that correct or am I still missing something here?

ChromiumPrime

+1 @Paperman

@spraynpray

It's a great link no doubt and very good resources but I fail to see how it's different from what I or roombarobot are saying... it's just more in depth. Frankly (and this is just my personal choice) when I was learning all of that as a beginner, I found those types of explanations very cumbersome and it made it much harder for me to understand the technicalities behind photography... mind you I do have a background in engineering.

Where I'm from we say it's scratching one's left ear with one's right hand. For some this is the only way it works but for some a more simpler and to the point explanation is needed... just a matter of how each one of us processes things (not that there is anything wrong with either ways as long as they get the job done!).

And while I'm at it, maybe a more precise definition of the terms we're throwing around is needed. From the article you linked to:

Although print size and viewing distance influence how large the circle of confusion appears to our eyes, aperture and focal distance are the two main factors that determine how big the circle of confusion will be on your camera's sensor.

Note the emphasis on the word "appears" (emphasis is from the article not mine). So then what exactly are we talking about here? DoF or apparent DoF?

And here is another quote from the article which also says almost exactly what I said in my reply to OP:

Depth of field also appears shallower for SLR cameras than for compact digital cameras, because SLR cameras require a longer focal length to achieve the same field of view.

@TaoTeJared

First of all, was name calling really necessary? Really?

Second, CoC is just the technical way to explain the phenomena that is DoF... not going to debate which one is more correct as they both have their uses and, quite frankly, it could be acceptable to use either term interchangeably as long as all parties know the intended meaning of those terms.

Third and last, DoF is not bokeh and bokeh is not DoF... unless you meant something different by "bokeh effect."

spraynpray

OK, my last word on the point I am trying to make (although I have said it before): Who the heck would ever maintain everything the same except change the sensor size? If you changed only the sensor, your composition would go awry so you would have to change position or focal length to regain your composition and in that video he didn't do that. That is why I said it is not reality. Now some of you all in this thread may understand DoF, but you haven't taken the time to read what I originally wrote and let it sink in before you replied which is why we are where we are.

Oh, and I also have an engineering background from which I made enough to retire at 49 so what? I could read before I went to school and that is the only qualification you need here!

TaoTeJared

Exactly, if the only thing you change is the sensor, then the DOF does not change.

Nooooooo! (Pounding head against a wall.) Roombarobot - do your own trials. Even though you keep writing, quoting everything that proves that statement wrong - you still are missing what it is saying.

Think of the sensor as being the same as the field of view. The DOF changes because you have changed the apparent focal length because it changed the filed of view.

TaoTeJared

@TaoTeJared

First of all, was name calling really necessary? Really?

Calling authors of blogs idiots for writing stuff that is not true - yes it is necessary. These are not opinions - they are facts. If someone wants to pass themselves off as an expert, is smart enough to create a blog, then they are smart enough to learn. Choosing not to learn, ignoring the correct science and writing a fairy tail, passing it off as true - that is an idiot. Actually I would call them something much worse - but idiot was profoundly nicer.

Second, CoC is just the technical way to explain the phenomena that is DoF... not going to debate which one is more correct as they both have their uses and, quite frankly, it could be acceptable to use either term interchangeably as long as all parties know the intended meaning of those terms.

CoC is not DOF - they are not interchangeable. Coc is a calculation, DOF is a measurement.

Third and last, DoF is not bokeh and bokeh is not DoF... unless you meant something different by "bokeh effect."

You did misunderstood. That was the point - Bokeh is not DOF.

You guys are really messed up on this one. We can get you straightened out.

heartyfisher

"If you do not change the lens, then DOF does not change" This is not correct.. the smaller sensor has a thinner DOF because the image is larger and the circle of confusion is larger for any point that is not in focus.

heartyfisher

This is so simple...DOF is dependent on focal length, aperture, and subject distance. When we change the area on the sensor, there is no difference in DOF, just like cropping in post processing does not change DOF.

Now, apparent DOF, when we take a 4"x6" print and enlarge to 20"x30" what we see is a degradation in the image, and the apparent DOF may change.

But, all of this is obvious only to folks who ave been hanging around behind a camera for some time. Beginners may get confused as spraynpray has suggested.

And, after the canon cameras were pulled out, I lost interest... )

LOL (about the lost interest)

You said "DOF is dependent on focal length, aperture, and subject distance. When we change the area on the sensor, there is no difference in DOF" .. actually there is no difference in the size of CoC but because the image is "magnified" on the smaller sensor the CoC looks larger. therefore the DOF is thinner.

roombarobot

@heartyfisher

The circle of confusion depends on the optics, not the sensor.
Also, the image is not "magnified" on a smaller sensor. Less of the image is captured, it is just cropped, not magnified.


Do this thought experiment... If changing only the sensor size would change depth of field, then I should be able to crop out the center of my image from an FX camera and change the depth of field.

Or, watch this video and tell me why, when thatnikonguy changes only the sensor size that the depth of field stays the same =

ChromiumPrime

Because of this:

Use the actual focal length of the lens for depth of field calculations. The calculator will automatically adjust for any "focal length multiplier" or "field of view crop" for the selected camera.

When you change the sensor size the calculator changes the (calculated) focal length based on the new sensor size.
Therefore the depth of field changes.

------------------------

As far as I understand it: FX gives you shallower depth of field when using your lens at its widest aperture (f/1.8), But as the diffraction limit is further away (Compared to DX) you can get about the same max depth of field as you get on DX before the diffraction limit kicks in.
Is that correct or am I still missing something here?

Thank you! That little paragraph completely eluded me when I was looking at that page and so I too was stumped. But I also think the calculator is doing even more compensations behind the scenes than we're aware of.

Another interesting peculiarity is if you hold all variables constant and change the body only (sensor size), you'll notice that DoF not only changes but also gets shallower with smaller bodies. :-?

ChromiumPrime

Calling authors of blogs idiots for writing stuff that is not true - yes it is necessary. These are not opinions - they are facts. If someone wants to pass themselves off as an expert, is smart enough to create a blog, then they are smart enough to learn. Choosing not to learn, ignoring the correct science and writing a fairy tail, passing it off as true - that is an idiot. Actually I would call them something much worse - but idiot was profoundly nicer.

Just a piece of friendly advice but drop the name calling and you'll go much further and win more people than you ever thought possible.

CoC is not DOF - they are not interchangeable. Coc is a calculation, DOF is a measurement.

Please reread the first sentence of my second point... pretty much what you're saying now, don't you think?

"CoC is just the technical way to explain the phenomena that is DoF"
vs.
"Coc is a calculation, DOF is a measurement."

Phenomena are often first measured and then explained by calculations

You did misunderstood. That was the point - Bokeh is not DOF.

Glad we agree on something then

You guys are really messed up on this one. We can get you straightened out.

No thank you! I'll keep my ways )

heartyfisher

@heartyfisher

The circle of confusion depends on the optics, not the sensor.
Also, the image is not "magnified" on a smaller sensor. Less of the image is captured, it is just cropped, not magnified.

You are kind of correct .. but also kind of not :-)
The CoC depends on the Focal length and the aperture( and the distance to the point of focus). so if the focal length is say X and the aperture is say Y then the CoC is Z and these values do not change whatever optics you use ( however many elements or ED glass etc)

But note that the CoC (the value Z) is a physical value that increases the further away from the point of focus. DOF is the distance limit from the point of focus where the CoC can be discerned as blur/sharp.

So although the CoC is the same for distance D from the point of focus. On the larger sensor it may fall within the "sharp" region of the DoF whereas for the smaller sensor the same distance D falls within the "blur" region because the CoC looks larger on the smaller sensor. So for the same Focal length, aperture and subject distance, DOF on the smaller sensor is thinner.

ChromiumPrime

You are kind of correct .. but also kind of not :-)
The CoC depends on the Focal length and the aperture( and the distance to the point of focus). so if the focal length is say X and the aperture is say Y then the CoC is Z and these values do not change whatever optics you use ( however many elements or ED glass etc)

But note that the CoC (the value Z) is a physical value that increases the further away from the point of focus. DOF is the distance limit from the point of focus where the CoC can be discerned as blur/sharp.

Hearty, but in this case aren't we talking more about flange distance than the size of the sensor? I agree that flange distance is another variable on DoF but the assumption was on using cameras with the same mount.

So although the CoC is the same for distance D from the point of focus. On the larger sensor it may fall within the "sharp" region of the DoF whereas for the smaller sensor the same distance D falls within the "blur" region because the CoC looks larger on the smaller sensor. So for the same Focal length, aperture and subject distance, DOF on the smaller sensor is thinner.

Let me ask the question in another way: if you frame a scene that spans 100 feet and snap a picture with a DX camera and an FX camera from the same distance, same FL and same f/stop, then you take the FX shot and crop it to the same portions of the DX shot, would there be a difference in DoF?

heartyfisher

Let me ask the question in another way: if you frame a scene that spans 100 feet and snap a picture with a DX camera and an FX camera from the same distance, same FL and same f/stop, then you take the FX shot and crop it to the same portions of the DX shot, would there be a difference in DoF?

No and Yes ! :-) (should it be yes and no ? ;-) )

If you view the Fx image on a 23" monitor and the Dx image on the same monitor. then the DoF of the Dx image will be thinner. If you view the FX image, cropped to DX and the Dx image on the same monitor then they will look the same.

Ironheart

Lets approach this from a practical perspective.

I used the calculators on this page: http://www.tawbaware.com/maxlyons/calc.htm

Assumed 0.020mm for CoC and f/2 for the aperture.

So what matters is Field Of View (FOV). This means that regardless of what sensor size, lens length, distance to the subject, you wind up with the exact same FOV in your end photo. I've calculated some equivalents here:

A:
50mm lens
FX sensor
5 meters from subject

B:
35mm lens
DX sensor
5.25 meters from subject

C:
50mm lens
DX sensor
7.5 meters from subject

All of these produce exactly 3.6 meters of horizontal Field of View, or in other words, the same framed/composed picture.

When you calculate the Depth of Field (DOF) for each of these equivalent scenarios, you get these results:

A: 0.8m DOF
B: 1.84m DOF
C: 1.8m DOF

My only point being is that if you try and take the same photo with different sensor sizes regardless of what you do (move back, use a shorter lens) you will change your DOF. Also, you will notice that this DOF calculator doesn't take into consideration the sensor size, only lens, f/stop and distance to subject.

The sensor size is only used to calculate the FOV. Hope this helps.

heartyfisher

"Assumed 0.020mm for CoC and f/2 for the aperture."

Interesting calculations .. but are you equating the CoC to the F2 ?

edit: just looked at the calculator and you enter the CoC as an input parameter ok.. but you cant use the same CoC for FX and DX the CoC for DX has to be smaller than the CoC for FX due to the difference in sensor size. eg if you included a tiny p&s 1/3.2" sensor then you obviously cant use the same CoC for your calculations for that sensor. .. i dont know if it would make a difference to the conclusion but i think you may need to use 0.013 CoC for the calculation of with DX sensors.

edit2: How did you calculate the "3.6 meters of horizontal Field of View"?

Ironheart

I used the "Dimensional Field of View Calculator". Changed the Units to "meters", plugged in 50mm, 1.5 focal multiplier, and 7.5 meters Distance to Subject. Press Compute, and viola! FOV 3.6m.

As far as the CoC I must appeal to higher authority but the site itself said to use 0.019 for cropped sensor. DOFMaster uses 0.020 for a Nikon DSLR but 0.030 specifically for the D3. To me it seems that the pixel pitch would mean more than the overal size of the sensor but whadda I know.

The calculators on that site are all really cool, fun for a rainy day :-)

Ade

To understand DoF and why sensor size affects it (and why the CoC of a DX sensor is smaller than an FX sensor), please humor me through this thought experiment:

1. If we define DoF as the depth where the image is perfectly in focus, then in theory the DoF of any picture is always zero.

That's because an image is only perfectly focused at one distance: the focus point. Anything before and beyond the focus point is -- by definition -- not perfectly focused.

2.If instead we define DoF as the depth where the image is still "sufficiently focused to look sharp", then an important question is: "when viewed at what distance?"

A slightly out-of-focused image might still look sharp when viewed from a few yards away. So viewing distance affects our perception of focus.

3. To put it another way: DoF effectively depends on the magnification of the image.

We all intuitively know this. A picture might look razor sharp on the camera's LCD, yet when we view it at 100% zoom on a big monitor it becomes painfully out of focus. This is not a digital phenomenon; In the old days, we always check slides & contact sheets using a magnifying loupe.

So DoF depends on magnification: DoF is the area we perceive to be in focus, and our perception of focus changes with magnification.

4. Sensor size affects magnification

Suppose you take an picture using a DX camera and an FX camera, each time filling the frame with the subject. When you print the picture (say at 8x10"), the DX image must be magnified 1.5x more than the FX image to fit into that 8x10" frame. This additional magnification is why sensor size affects DoF.

5. The so-called "circle-of-confusion" is arbitrary

This value is used to estimate when a point is "still sufficiently focused to look sharp" (see #2 above). But we already know this value actually depends on the magnification, the viewing distance, and has a subjective factor as well (how sharp are your eyes?)

So we (science & industry) have defined the circle-of-confusion using a standard distance of 10" when viewing an image magnified to 8x10". At that distance and size, we assume a human being with good vision can distinguish a point about 1/100th of an inch wide (.254mm). This is an arbitrary definition based on simplifications to make the math easy.

6. That is to say, the CoC of an 8x10" image is .254mm (by our definition above)

Since an FX sensor must be magnified about 8.333x to fit fill an 8x10" frame, the CoC of an FX sensor is simply .254mm/8.333 = 0.03 mm. A DX sensor has a 1.5x further magnification, so the CoC of a DX sensor is simply 0.03mm/1.5 = 0.02mm. A CX sensor has a 2.7x magnification compared to FX so its CoC is 0.03/2.7 = 0.011mm.

Ironheart

Thanks @Ade. Makes perfect sense. The only calc I had to redo was scenario A which used the FX sensor. Pluggging in 0.030 for the CoC gives 1.2m instead of 0.8m:

A: 1.2m DOF
B: 1.84m DOF
C: 1.8m DOF

heartyfisher

Thanks @Ade. Makes perfect sense. The only calc I had to redo was scenario A which used the FX sensor. Pluggging in 0.030 for the CoC gives 1.2m instead of 0.8m:

A: 1.2m DOF
B: 1.84m DOF
C: 1.8m DOF

So that would mean that the previous A = 0.8m DOF was for a DX Sensor but with reduced(cropped) FOV.

Ironheart

Yes, 2.4m to be precise @heartyfisher The DOF was smaller, because the subject is closer.

SkintBrit

Just read this whole thread for the first time from start to here. I'm now going for a lie down, as my head hasn't felt this filled since school Physics classes! :-t

Msmoto

I am going with Ade. In theory DOF is zero. Viewing distance, magnification, sensor resolution, are factors which will change the apparent DOF.

spraynpray

Sorry Msmoto, I think that is an incorrect statement. DoF is not zero, it is the focal plane that is infinitely thin, the DoF by its very name and definition has depth as it is a field the depth of which is dependent on the factors you mentioned.