Depth of field Full frame vs Crop?

[snerkler]

I thought that I understood the basic principles of DOF and the different sensor sizes, and how for a given focal length lens you are not changing the lens parameters per se, just cropping the middle of the image circle on DX sized sensors compared to FX. However, I've recently been trying to decide between swapping my D750 for a D810 for cropability for shooting wildlife, or getting a used crop body such as the D7200 and it's 'raised a question' I just can't get my noodle around

If you look at these two calculations using a 200mm f2.8 lens with a 10m subject distance the crop body has a smaller DOF, which is what I'd expect as you have a longer effective focal length.

In the case of the FX you have a near limit of 9.8m and a far limit of 10.2m. If I use the FX camera in DX mode, essentially cropping the centre of the image circle like the DX camera I would expect that the near and far distances would remain at 9.8 and 10.2m because as far as I'm aware there is no difference between using an FX camera in DX mode and cropping in post to give the same framing. Obviously if I cropped in post I wouldn't expect the near and far distances to magically change.

So what else is going on with a DX body that changes the effective FOV and DOF, obviously it's not purely down to the cropping of the image circle otherwise I would expect the shot above to have the same DOF using a DX body and an FX body in DX mode.

Clearly I'm missing something obvious, but I can't think what it is?

 

[realspeed]

Spectacles ?

 

[Nawty]

Magnification.

CoF is changed with magnification.

 

[snerkler]

Magnification.

CoF is changed with magnification.

Ahh yes, I keep forgetting that DX 'magnifies/enlarges' the image

Thanks for this. So if I want more shallow DOF I'm obviously better going down the DX body route rather than the DX mode of the D810, assuming same subject distance

 

[Phil V]

You can choose to study circle of confusion and airy discs and magnification vs viewing distance but it'll make most people's head spin

The easy thing to remember is that your phone has no control of DoF because of the tiny sensor and that a 10x8 plate regularly uses f45 to get enough DoF for a 'normal' shot.

 

[HoppyUK]

Ahh yes, I keep forgetting that DX 'magnifies/enlarges' the image

Thanks for this. So if I want more shallow DOF I'm obviously better going down the DX body route rather than the DX mode of the D810, assuming same subject distance

You mean FX body (full-frame) When you crop an image, you are effectively changing the sensor format in DoF terms.

DoF is all about magnification (including the diameter of the lens aperture). Any change to magnification in the imaging chain affects DoF - focal length, lens aperture, sensor format (note how the Circle of Confusion value changes with format in the screen grabs above), shooting distance, print size and viewing distance.

A lot of people forget the last bit - print/output size and viewing distance - but that's actually where all DoF calcs start. The universal standard is based on viewing a 10in print from a distance equal to the diagonal, ie approx 12in. At that size and distance, the human eye is said to be unable to distinguish anything smaller than 0.2mm on the print. That is the starting point for all DoF calculations, though very usefully, the same standard applies to all output sizes, so long as the correct viewing distance is maintained. That's why a huge street poster looks blurred close up, but looks sharp from the other side of the street, and by the same token, if you blow things up on-screen to 100% without also moving back to compensate, perceived DoF reduces.

Edit: the difference in DoF terms between full-frame and crop-format APS-C sensor is about 1.2 stops, when the subject is framed the same. The conversion is f/number x crop factor, eg Nikon full-frame FX at f/6 = f/4 on Nikon DX.

You can choose to study circle of confusion and airy discs and magnification vs viewing distance but it'll make most people's head spin

The easy thing to remember is that your phone has no control of DoF because of the tiny sensor and that a 10x8 plate regularly uses f45 to get enough DoF for a 'normal' shot.

 

[snerkler]

You mean FX body (full-frame) When you crop an image, you are effectively changing the sensor format in DoF terms.

I don't think so. Looking at the example above the DX body (D7000) has a more shallow DOF than the FX body (D700), so even if I cropped the FX image in post to match the DX framing the actual DOF won't change and will still be larger than that shot on the DX body. At least that's what I'd assume. eg if I shot this example above the depth of field on the FX is from 9.8-10.2m. If I then crop the image in post the DOF/zone of focus will still be the same, by cropping the image in post it won't suddenly make less of the subject in focus. It was my understanding that this is all that is happening when using DX mode with an FX body, ie it crops the image 'in post' rather than magnifying the image like using a DX body?

 

[HoppyUK]

I don't think so. Looking at the example above the DX body (D7000) has a more shallow DOF than the FX body (D700), so even if I cropped the FX image in post to match the DX framing the actual DOF won't change and will still be larger than that shot on the DX body. At least that's what I'd assume. eg if I shot this example above the depth of field on the FX is from 9.8-10.2m. If I then crop the image in post the DOF/zone of focus will still be the same, by cropping the image in post it won't suddenly make less of the subject in focus. It was my understanding that this is all that is happening when using DX mode with an FX body, ie it crops the image 'in post' rather than magnifying the image like using a DX body?

I was correcting what I assumed was a typo in post #4, but that aside, you're forgetting the effect of cropping that is effectively a format change. See my post above. DoF is not set in stone at the moment of capture, but only when the image is finally printed and viewed.

Edit: If you use a Nikon FX camera in DX mode, it is behaving exactly like a DX camera (apart from the fact that the number of effective pixels is substantially reduced). Framing changes by the crop factor, and so does DoF, also by the crop factor.

 

[Faldrax]

...rather than magnifying the image like using a DX body?

Using a smaller sensor does not magnify the image

To go back to the FX / DX comparison

Shoot a scene on FX, then change to DX mode, then shoot the same scene on a DX camera which has the same number of pixels as the DX mode of the FX camera (use the same prime lens for all 3 shots, the FX will have a wider FOV)
Then (assuming you have a suitable large format printer), print all 3 images at the same DPI.
The 2 DX images will be smaller, as they have less pixels - the magnification is the same.

What normally folks do is print (or view) the FX and DX images at the same physical size - the DX image will be at a lower DPI and will APPEAR to be magnified as a result.

 

[snerkler]

Using a smaller sensor does not magnify the image

To go back to the FX / DX comparison

Shoot a scene on FX, then change to DX mode, then shoot the same scene on a DX camera which has the same number of pixels as the DX mode of the FX camera (use the same prime lens for all 3 shots, the FX will have a wider FOV)
Then (assuming you have a suitable large format printer), print all 3 images at the same DPI.
The 2 DX images will be smaller, as they have less pixels - the magnification is the same.

What normally folks do is print (or view) the FX and DX images at the same physical size - the DX image will be at a lower DPI and will APPEAR to be magnified as a result.

Magnifying was probably the wrong word tbh, enlarging was probably better. From discussions elsewhere regarding the extra demands DX cameras put on lenses I was informed that this was because the DX image has to be enlarged more when viewing on the same medium.

 

[snerkler]

I was correcting what I assumed was a typo in post #4, but that aside, you're forgetting the effect of cropping that is effectively a format change. See my post above. DoF is not set in stone at the moment of capture, but only when the image is finally printed and viewed.

Edit: If you use a Nikon FX camera in DX mode, it is behaving exactly like a DX camera (apart from the fact that the number of effective pixels is substantially reduced). Framing changes by the crop factor, and so does DoF, also by the crop factor.

This is where my noodle's getting confused, and I'm probably thinking about it in the wrong manner. But this is what I mean. Take this image, for arguments sake lets say the letters in focus are 0.5cm, and the zone of focus is 1cm

If I crop the image in post, the letters are still 0.5cm and the zone of focus has not altered and so still 1cm, all we've done is zoomed in.

However, if I was to shoot the same shot using a DX with the same lens in the same position I would expect framing like the second pic, but a narrower zone of focus. Does this make sense, and is my thinking correct? If so, my belief was that when using the DX mode on FX camera it essentially does what I did, ie take the FF image and then crop it afterwards. Or am I wrong in that it takes a DX image from the outset? I'm sure I tried this myself and couldn't see any difference in DOF between the same shot on my D750 in DX mode, and then FX mode which I then cropped in lightroom.

Sorry, I know it's not a massive deal, but you must know what it's like when you can't get your head around something, it bugs the hell out of you

 

[Faldrax]

Magnifying was probably the wrong word tbh, enlarging was probably better. From discussions elsewhere regarding the extra demands DX cameras put on lenses I was informed that this was because the DX image has to be enlarged more when viewing on the same medium.

There is nothing inherent in using a smaller physical sensor than the image circle that will put an extra demand on a lens.

However, it is often the case that a DX (or APS-C, or 'crop', however you want to refer to it) sensor will have smaller pixels (so closer together) than a FX (or FF) camera - hence the more densely packed sensor can reveal weaknesses in a lens that a less densely packed sensor would not.

This is the same as comparing a 20 MP D5 can be regarded as less demanding on lenses than a 36 MP D810 - but they are the same sensor size.

 

[sk66]

Phil and Hoppy gave you the answer... COC.
DOF does not exist as a fixed characteristic in an image, it varies with the amount of enlargement required (print/display size) and the viewing distance (apparent size). Because a DX crop and a DX sensor have the same physical size, they will require the same amount of enlargement for a given display size. That gives them the same relative COC and the same apparent DOF when viewed from the same distance (the only remaining variable).

Remember that DOF is only "apparent acceptable sharpness."
Take an image and display it on your screen at a smaller size, then display it at a much larger size. The larger size will have less apparent sharpness (DOF). Then walk away from your computer and view the image at a distance where it is the original size in your FOV and the original sharpness will return... *that* is DOF/COC at work.

And DOF/COC only relate to what is in "acceptable sharpness," it says nothing about the characteristics of what is not...

 

[woof woof]

DoF is not set in stone at the moment of capture, but only when the image is finally printed and viewed.

I do know what you mean but I fundamentally disagree with this thinking and this way of looking at it. The DoF is set in stone when you press the shutter button and all you can then do is make it more or less obvious by printing smaller or larger or viewing from nearer or further away.

Once you've picked up the kit, set it up and pressed the shutter button the result you will see at any print size and viewing circumstance you choose is decided and is set so for me it all starts with my intent... What size image do I want to produce and how am I going to view it and these decisions should decide the kit and the settings.

It's no good thinking "The DoF isn't set until I print it out and decide where I'm going to stand to look at it" because you should have thought of that first and chosen your kit and settings to match your final image size and viewing requirements.

 

[sk66]

If I crop the image in post, the letters are still 0.5cm and the zone of focus has not altered and so still 1cm, all we've done is zoomed in.

On my screen, the letters are not the same size, and the area of blur is not the same size... i.e. the letters in the enlarged image are less sharp (a bit harder to read) and the DOF is less.

 

[sk66]

all you can then do is make it more or less obvious by printing smaller or larger or viewing from nearer or further away.

It is the "more or less obvious" that *defines* DOF... DOF is only "apparent/acceptable sharpness."
You're thinking of it more in terms of image resolution...

Edit: But I do agree... if you know the output size/viewing conditions required, then DOF needs to be considered before the shot is taken.

 

[The23rdman]

I do know what you mean but I fundamentally disagree with this thinking and this way of looking at it. The DoF is set in stone when you press the shutter button and all you can then do is make it more or less obvious by printing smaller or larger or viewing from nearer or further away.

Once you've picked up the kit, set it up and pressed the shutter button the result you will see at any print size and viewing circumstance you choose is decided and is set so for me it all starts with my intent... What size image do I want to produce and how am I going to view it and these decisions should decide the kit and the settings.

It's no good thinking "The DoF isn't set until I print it out and decide where I'm going to stand to look at it" because you should have thought of that first and chosen your kit and settings to match your final image size and viewing requirements.

Who thinks "I'll use f/2.8 because I'm going to print this at 30x20 and view it from 6 feet away" before they press the shutter?

 

[snerkler]

On my screen, the letters are not the same size, and the area of blur is not the same size... i.e. the letters in the enlarged image are less sharp (a bit harder to read) and the DOF is less.

Yes they are physically bigger on my screen too as they're zoomed in, but this wasn't my point tbh.

When you use DOF calculators for example, you take the subject distance, lens, sensor size and aperture and it tells you what the near and far distances are, regardless of how much you may or may not zoom into an image afterwards. So going back to the lettering, hypothetically speaking if the printed letter are 0.5cm high in the real world this never changes. Yes we can zoom in later, but the original letters are only 0.5cm high.

To go one step further and getting some real figures involved. If I shot those words on a FF camera from 1.5m away at 200mm f2.8 this would give me a DOF of 1cm, with a near point of 0.5cm and a far point of 1.5cm. These distances are set regardless of what I then do with the image later on, all I'm doing in post by cropping is zooming in making the letters appear larger but the original DOF would be the same, you still see the middle letters in focus and the outer ones just out of focus. If I were to shoot the same setup but with a DX body I would expect a smaller DOF so the outer letters would probably be totally blurred. Therefore the two resulting images would look different. ie the FX image cropped would show the outer letters slightly in focus, the DX image would show the outer letters out of focus.

 

[sk66]

These distances are set regardless of what I then do with the image later on

No... and this is the problem w/ DOF calculators. The results the calculator gives are based upon a given COC. And that COC assumes a fixed display/print size (of the entire sensor area) *and* a fixed viewing distance.

In reality, if you are not going to be displaying the (entire) image at a size and viewing distance where it occupies ~ 45* diagonal viewer FOV, then the "default COC" is *wrong* for your intended use... and the default results from the calculator are *wrong.*

Edit: For example, say you are going to take a D810 image and make a large fine art print for display in a gallery where it will be viewed closely. It would be more appropriate to set the calculator to use a more demanding COC of .02 or .01 instead of the FF default .03mm... because the viewing conditions are going to be much more demanding than the assumed defaults...

BTW, it is easily arguable that the default COC's are outdated and "wrong" anyway... the only thing they really do is give you a standard "starting point" to compare against.

 

[sk66]

Who thinks "I'll use f/2.8 because I'm going to print this at 30x20 and view it from 6 feet away" before they press the shutter?

For me it's more like "this image will be viewed more critically so I need to stop down"....

 

[snerkler]

No... and this is the problem w/ DOF calculators. The results the calculator gives are based upon a given COC. And that COC assumes a fixed display/print size (of the entire sensor area) *and* a fixed viewing distance.

In reality, if you are not going to be displaying the (entire) image at a size and viewing distance where it occupies ~ 45* diagonal viewer FOV, then the "default COC" is *wrong* for your intended use... and the default results from the calculator are *wrong.*

Edit: For example, say you are going to take a D810 image and make a large fine art print for display in a gallery where it will be viewed closely. It would be more appropriate to set the calculator to use a more demanding COC of .02 or .01 instead of the FF default .03mm... because the viewing conditions are going to be much more demanding than the assumed defaults...

BTW, it is easily arguable that the default COC's are outdated and "wrong" anyway... the only thing they really do is give you a standard "starting point" to compare against.

OK, thanks. I think I'm going to have to try this myself and see as I just cannot my head around how a DX shot and an FX shot cropped in post can look the same using my example.

So from my example above the FX cropped in post would look like this, as already posted
1.

Whereas from the DX body I would expect it to look more like this due to the smaller DOF.
2.

From my understanding of how DX mode on FF works I would expect the image to look like 1, and not 2. However, this could be wrong if the crop is done at image capture rather than after the 'event'.


As I said though I will have to test it myself to get my head around it, I've got use of my mate's D7200 this weekend

 

[The23rdman]

For me it's more like "this image will be viewed more critically so I need to stop down"....

Yes, that makes more sense.

 

[sk66]

As I said though I will have to test it myself to get my head around it, I've got use of my mate's D7200 this weekend

Ok, remember that DOF is "apparent." That also means it's subjective and even variably dependent on the individual's visual acuity. And remember that DOF only relates to what is acceptably sharp, not what is OOF.

In the D810 example I gave, a COC of .02mm is the "DX default." The calculator you used allows you to select that instead of a camera/sensor at the bottom of the list. But most do not know what COC to use.
The DX .02mm would be (more) appropriate for the D810 image if the "DX area" will occupy the default ~45* FOV... regardless of whether the image is cropped or not. And the more demanding COC will generate results of correspondingly less DOF (the same as if a DX sensor was selected).

 

[snerkler]

Ok, remember that DOF is "apparent." That also means it's subjective and even variably dependent on the individual's visual acuity. And remember that DOF only relates to what is acceptably sharp, not what is OOF.

In the D810 example I gave, a COC of .02mm is the "DX default." The calculator you used allows you to select that instead of a camera/sensor at the bottom of the list. But most do not know what COC to use.
The DX .02mm would be (more) appropriate for the D810 image if the "DX area" will occupy ~45* FOV... regardless of whether the image is cropped or not. And the more demanding COC will generate results of correspondingly less DOF (the same as if a DX sensor was selected).

Thanks for taking that time for your explanations (that goes for everyone). I'm not sure if it's helped or confused me more But I appreciate it nonetheless.

 

[Phil V]

This has gone from being a twice yearly can of worms to a monthly one, the facts never change. Some really clever people will use some really complicated answers, and the reality is that at crop vs full frame it's a tricky point to prove either way.

But compare a cheap compact with a std lens at f2 and a full frame dslr with a std lens at f2 and you'll see a massive difference, not based on pixel density or lens sharpness but on the simple physics that a small sensor (or piece of film) gives a greater DoF than a large sensor (or piece of film).

That's an experiment most of us can do with our phone and our camera.
My iphone has a 4.1mm f2.2 lens (not much use to know) but it has the same field of view as a 33mm lens on full frame, so I can measure it against a FF lens if I want to. And at about 5ft the Dof will be just about from 4ft to infinity - with an aperture of f2.2

Now a 35mm lens on FF at f2.2 has a DoF of less than 1ft, and we can easily see that when comparing shots from a camera on our phone.

What the DoF calculators don't take into account is the actual image you're looking at, I tend to not use crop factors etc. But your DoF calculator will let you do the experiment using a crop camera with an 85mm lens and a FF with a 135mm lens (the 85mm would equal 136 on Canon crop) -
135mm f2 on FF at 6 feet gives a DoF of 0.07ft
85mm f2 on crop at 6 feet gives a DoF of 0.11ft

Now they're not 'exactly' the same shot, but it's close enough for Jazz

 

[HoppyUK]

Too much confusing information here now. I explained the issues in post #6.

Where Snerkler still appears to be going wrong is to overlook the fact that DoF only exists as a measurable commodity when the final image is output/printed and viewed from the correct distance.

All DoF calculators assume this* based on that 10in print mentioned in post #6, viewed from a distance equal to the diagonal, ie 12in. As also mentioned, the same parameters apply to any size print when viewed from a distance equal to the diagonal, regardless of size. But you can't change one without the other or it all goes out of the window.

*DoF calculators also assume that the image has not been cropped! The DoF standards date back decades (long before FX vs DX and 100% viewing on screen had even been dreamed of). You could argue that maybe things should be updated, but it's actually not a bad standard at all and rather more importantly, it is consistent and universally used.

 

[sk66]

Thanks for taking that time for your explanations (that goes for everyone). I'm not sure if it's helped or confused me more But I appreciate it nonetheless.

Yes, that makes more sense.

This has gone from being a twice yearly can of worms to a monthly one, the facts never change. Some really clever people will use some really complicated answers, and the reality is that at crop vs full frame it's a tricky point to prove either way.

It's really quite simple... more enlargement equals less DOF, regardless of the cause.

More enlargement due to a longer FL ='s less DOF (larger sensor/same distance/same image)
More enlargement due to a shorter subject distance ='s less DOF (larger sensor/same FL/same image)
More enlargement due to a smaller sensor (image area) ='s less DOF (same FL/same distance/different image)
* and more enlargement due to larger (non standard) relative display size ='s less DOF (this is the one everyone forgets about)

IMO this stuff is too variable and subjective, it's not worth getting much more concerned about it than that... more critical viewing (enlargement) requires greater sharpness/DOF, so stop down (or use a system/technique capable of greater sharpness/DOF).
I have an app for that, I *never* use it... Heck, the vast majority of my work only gets displayed on the web or it's final use is "undetermined." That makes the whole thing about pointless (I just err for larger sizes/more critical viewing).

 

[sk66]

All DoF calculators assume this* based on that 10in print mentioned in post #6, viewed from a distance equal to the diagonal, ie 12in. As also mentioned, the same parameters apply to any size print when viewed from a distance equal to the diagonal, regardless of size.

Quite some time ago it occurred to me that all of the "standards" relate back to the same (approx) 45* AOV. A "normal" lens for any format will record ~ 45*, which is then displayed at a standard size/distance where it occupies that same ~45*... i.e. FL=diagonal and VL=diagonal

And I find that thinking of it in terms of AOV makes it easier to visualize/understand...

 

[jbw]

With the same physical focal length of lens, the smaller-sensor camera has a tighter field of view.
Why?
In order for the object being photographed by the smaller-sensor camera to fill the same proportion of the frame (as in the larger-sensor camera) - we would have to move further away with it.
Moving further away from the subject increases the distance required to obtain sharp focus, which strongly increases depth of field.
We know from the physics of optics that the depth of field increases significantly with the focus distance ( focusing on something further away always has a greater dof ).
In fact, the dof grows approx.by a 'squared' function as distance to the subject being photographed increases.
We say approx. because what is defined as in focus ( within the dof) is somewhat subjective.

 

[jbw]

Sorry - the sentence that says 'why?' should have said 'why does the tighter field of view change the dof?'

 

[rob-nikon]

This is something I tried to get my head around a few years back. I looked at from a different point of view, looking at the same field of view (ie using the same framing). I tested it using a 70-200 on a d7100 and a 300 f2.8 on a d800 so both giving the same framing of a '300mm' effective field of view.

DSC_9172 by -Rob - Nikon-

D71_8462 by -Rob - Nikon-

If looking at the same EFoV FX gives a shallower DoF than DX does in the examples above. If the same focal length lens was used then I too would expect the DX DoF to be shallower but the framing of the images would be different. I took the above images for a blog post but never tried the same focal length on both DX and FX so never finished it off. I was looking at what I preferred with the same image framing.If I was looking at getting as close as I could with a fixed focal length (using the longest lens I had) and not so worried about subject isolation and backgrounds then DX would the way to go for me. I much prefer the isolation FX gives and would rather try to get closer to the subject if possible first than use a crop camera. I do understand your predicament as I too have picked up a crop camera I'm the hope it would get me closer (and to try remote stuff) but I always go for the FX body because of the look it gives. Is there any chance of using a longer focal length to get 'closer' than use a crop body?

Note the stop difference in ISO was put down to a series minor setting differences (exposure compensation difference by 1/3) and in the cameras (metering types).

 

[sk66]

If looking at the same EFoV FX gives a shallower DoF than DX does in the examples above.

Yes, technically it does... but it's probably not more than a couple of inches (at most) in this example.

What this example is primarily showing is how a longer FL will include less of the BG and throw it more OOF (greater enlargement) when there is notable BG separation distance. In fact, you could stop down the longer FL for even more sharpness and equal/more DOF at the subject, while retaining equal/greater BG blur... it's one of "the secrets."
That's why I kept noting that "DOF" doesn't say anything about what is OOF...

I frequently stop down quite hard with longer FL's (800mm, f/16).

 

[HoppyUK]

Excellent examples of the squirrel from Rob above and great photography too (as usual ). That is the best comparison for this - the same subject, framed the same, from the same distance (so perspective remains the same) and focal length adjusted to compensate. And viewed side by side, at the same distance LOL

Yes, technically it does... but it's probably not more than a couple of inches (at most) in this example.

What this example is primarily showing is how a longer FL will include less of the BG and throw it more OOF (greater enlargement) when there is notable BG separation distance. In fact, you could stop down the longer FL for even more sharpness and equal/more DOF at the subject, while retaining equal/greater BG blur... it's one of "the secrets."
That's why I kept noting that "DOF" doesn't say anything about what is OOF...

I frequently stop down quite hard with longer FL's (800mm, f/16). <snip image>

I can't agree with you on that Steven, not as a bald and unqualified statement (my bold). Are you saying that if Rob had been able to raise the f/number by 1.2 stops in the full-frame image above, the two images would not be effectively identical? Because they would be, both in terms of foreground DoF and background blur, as near as makes no difference.

It is true that there are very small changes, but they are only visible (let alone significant) when you compare extremes, like maybe a compact camera (typically 5x crop factor) vs full-frame, that kind of thing. Or maybe even the 2.7x crop factor on Nikon's 1in format (Nikon1-series) vs full-frame. But here were only talking about a 1.5x crop factor and for all practical purposes the equivalence calculation stands.

 

[LCPete]

Excellent examples of the squirrel from Rob above and great photography too (as usual ). That is the best comparison for this - the same subject, framed the same, from the same distance (so perspective remains the same) and focal length adjusted to compensate. And viewed side by side, at the same distance LOL



I can't agree with you on that Steven, not as a bald and unqualified statement (my bold). Are you saying that if Rob had been able to raise the f/number by 1.2 stops in the full-frame image above, the two images would not be effectively identical? Because they would be, both in terms of foreground DoF and background blur, as near as makes no difference.

It is true that there are very small changes, but they are only visible (let alone significant) when you compare extremes, like maybe a compact camera (typically 5x crop factor) vs full-frame, that kind of thing. Or maybe even the 2.7x crop factor on Nikon's 1in format (Nikon1-series) vs full-frame. But here were only talking about a 1.5x crop factor and for all practical purposes the equivalence calculation stands.

Agree an excellent example from Rob shows what happens very well
Am I right in saying that if Rob had changed the aperture from 5.6 to 4 or more with the crop camera you would get the same background blur as the full frame example?

 

[LCPete]

Yes, technically it does... but it's probably not more than a couple of inches (at most) in this example.

What this example is primarily showing is how a longer FL will include less of the BG and throw it more OOF (greater enlargement) when there is notable BG separation distance. In fact, you could stop down the longer FL for even more sharpness and equal/more DOF at the subject, while retaining equal/greater BG blur... it's one of "the secrets."
That's why I kept noting that "DOF" doesn't say anything about what is OOF...

I frequently stop down quite hard with longer FL's (800mm, f/16).

really nice shot see what you mean about the background
Interesting that you use F16 I have never stopped down that much with long lenses I rarely go past F8
The shutter speed must have been very low though or do you raise the ISO quite high?

 

[HoppyUK]

Agree an excellent example from Rob shows what happens very well
Am I right in saying that if Rob had changed the aperture from 5.6 to 4 or more with the crop camera you would get the same background blur as the full frame example?

Yes, almost exactly. The precise difference is f/number x crop factor, ie 5.6/1.5=f3.7, but f4 is near enough.

 

[LCPete]

Yes, almost exactly. The precise difference is f/number x crop factor, ie 5.6/1.5=f3.7, but f4 is near enough.

Thanks Richard

 

[HoppyUK]

Thanks Richard

In the interests of completeness, there are different opinions on this, because the fact is that when you change one thing, actually everything else changes too. If you're being hyper-critcal/technical, and touching on Steven's last point, there are all sorts of subtle changes happening and the common equivalence conversions that I've used are a broad brushstroke. TBH, I think they are plenty good enough for very high quality working and cover the significant changes to depth-of-field that matter and are always noticeable.

But there are other subtle changes. As I mentioned above, they become increasingly noticeable in the out-of-focus areas (bokeh) when you compare extremes, like big changes in focal length and format rather than the relatively small shift from full-frame to APS-C crop format. When you simply zoom a lens, there can be small changes to bokeh, when you change apertures the shape of the diaphragm will be seen in the rendering of highlights, eg wide open perfectly will be round, or pentagonal when stopped down with a five-bladed diaphragm, and closer to circular with nine blades. If you change lenses, say from a prime to a zoom, as Rob did with the squirrel, there will be changes there too. - not to depth-of-field as such, but to the smoothness of the blurring. It goes on...

Personally, in practical terms I think there are plenty of more important things to worry about If I was ever lucky enough or good enough to get that close to a red squirrel, I'd be mostly concerned about controlling my excitement first, then using a fast enough shutter speed to avoid camera-shake and nailing focus 100%. Once I'd got that right, I'd firing away to capture the right moment. The background would be blurred regardless, and not a deal-breaker either way.

 

[sk66]

really nice shot see what you mean about the background
Interesting that you use F16 I have never stopped down that much with long lenses I rarely go past F8
The shutter speed must have been very low though or do you raise the ISO quite high?

The ISO is rather high (6400, D5) and I used on camera flash (catchlight edited out).

 

[sk66]

Are you saying that if Rob had been able to raise the f/number by 1.2 stops in the full-frame image above, the two images would not be effectively identical? Because they would be, both in terms of foreground DoF and background blur, as near as makes no difference.

Yes, I am saying they would not necessarily be "nearly identical."
The changes to the image are exactly the same as would occur if you simply zoomed in.
Take the 200mm @ f/5.6 from 20ft (guesstimate) on FF and you get 1ft DOF. Change it to 300mm f/8 and you get .6ft DOF with the subject being (potentially) sharper.
Then crop the 200mm image... the characteristics of the remaining image do not change, but the DOF decreases due *only* to the cropping. And there can be a notable difference between sharpness/lack of sharpness as recorded at the sensor when compared to the changes due to enlargement/display (i.e. you take the less sharp 200mm cropped image *and* you enlarge it more).

But typically with shorter FL's and more DOF these differences are very minimal. And yes, the perspective is the same in both images.