Full-frame vs Crop-sensor comparison : Depth-of-field & Perspective
When the differences between full-frame and crop-sensor cameras are discussed, there is an inevitable question about whether the crop sensor multiplies the focal length. Whether a 50mm lens on a crop-sensor acts like a 75mm lens (on a 1.5x crop sensor) or 80mm lens (on a 1.6x crop sensor).
The answers given on the photography forums are confusing – yes, the focal length effectively increases. No, it doesn’t. Two answers that are polar opposites. The discussion (which tend to devolve into arguments) are convincingly made for both sides. And that’s because the topic is a complex one … and therefore the answer is (kinda) complex as well.
One argument goes along the lines that the crop sensor is just that, a crop. An enlargement. That nothing changes. The focal length doesn’t change – you just get less of the scene. And that there is no “equivalent focal length” when you go to a crop sensor camera.
But what really happens is more complex than that.
With this article, I want to help analyze what happens when you change lenses between a full-frame camera and a crop-sensor camera. And we’ll analyze whether there is actually an equivalency between certain focal lengths, when using a crop-sensor camera vs a full-frame camera. In other words, whether your 50mm lens becomes “equivalent to” a 75mm or 80mm lens when used on a crop-sensor camera.
Since this article ended up being a long meandering discussion, I thought it best that we start with the final summary. Just to save the impatient people some work.
Summary:
Yes, a 50mm lens does indeed behave like an equivalent focal length of a 75mm lens (on a 1.5x crop sensor), or an 80mm lens (on a 1.6x crop sensor) for the same composition … however, the depth-of-field increases by about a stop.
Yes, a 100mm lens on a crop-sensor camera will give you the same perspective as a 150mm / 160mm lens (on a full-frame camera), if you don’t change position … however, the DoF increases.
But let’s discuss this with some images:
Notes on depth-of-field / DoF
Before we can go much further, we need to recap on Depth-of-Field:
- Shallow depth of field is NOT the same as bokeh.
The image above certainly has nice, smooth bokeh. But it also has shallow-depth-of-field. Two things which seemingly are the same, but aren’t. So on that note, if you are one of those who say things like “give it some bokeh”, then you need to stop. It is meaningless.
- You can not “zoom with your feet”, because if you change your position, your perspective changes.
With a zoom, the perspective does not change – you are merely enlarging the image. This distinction becomes an important point. If you put a 100mm lens on a full-frame body, and a 100mm lens on a crop body, and you want the same size for your subject in the frame, then you are going to have to move much further back to get the same image size with the crop sensor camera. And then your perspective has changed. And yes, then it will give you the same perspective as a 150mm / 160mm lens on a full-frame body when shooting at that new further distance.
- A tutorial on Depth-of-Field
Our friends at Cambridge In Color have written one of the clearest explanations on DoF that I’ve ever read. Spend some time there. The concept of “circles of confusion” is an important one. It is on this that everything that you need to know about DoF, hinges upon. An important note there – DoF is defined via circles of confusion, which is specified for a certain print size, at specific distance.
- The way that DoF is defined (via circles of confusion), means that viewing distance, and the size of the final image, affects how DoF is perceived. This then implies if you are comparing a 12 megapixel image and a 36 megapixel image, then you can’t judge the DoF of the image at 100% resolution. You are better off going to “full screen” on both those images, and comparing at an equal size. You have to “equalize” this for comparison, because if you view a 36 megapixel image at 100%, the DoF will seemingly be shallower than a sensor with lesser resolution.
- DoF changes incrementally. In other words, the only point at which something is most definitely *in* focus, is the actual plane of focus. From there it is an incremental change in sharpness to the foreground and background. By the time you are using a small aperture then, the depth-of-field extends deeper into the background and foreground. Even then, at the extremities of that small-aperture DoF, there is no exact point at which we can immediately say the image flips from sharp to un-sharp.
Direction & Quality of Light
I tried to distill the essence of what we, as photographers, work with – light! Before we can truly grasp on-camera flash and off-camera flash, and really, any kind of photography, we have to be aware of the direction and quality of light. We need to observe the light that we have, and then decide how best to use it, or enhance it.
With this book, I try my best to share those “aha!” moments with you, and I do believe this book can make a difference to your photography.
The book is available on Amazon USA and Amazon UK, or can be ordered through Barnes & Nobles and other bookstores. The book is also available on the Apple iBook Store, as well as Amazon Kindle.
Figuring out a reasonable comparison method:
To take this tutorial away from the armchair to an actual shoot with comparison examples, the ideal would’ve been to have a full-frame camera that has exactly the same resolution as the crop-sensor camera. Just to keep things simple. And I’d have to have two lenses which are exact 1.5x versions of each other. For example 50mm and 75mm. Or 100mm and 150mm. Just to keep things even.
This doesn’t exist (that I know of), I tried a few ways of reasonably comparing the full-frame sensor and crop-sensor effect.
Then I had this brainwave – I could use these two Nikon lenses:
– Nikon 85mm f/1.4 lens (affiliate) on a full-frame camera,
– Nikon 58mm f/1.4 lens (affiliate) on a 1.5x crop sensor.
The 58mm focal length effectively acts like an 87mm lens on a 1.5x crop sensor camera. Close enough!
One obstacle here is that the two lenses display different optical qualities. For example, a harsh bokeh might make the background look sharper or crisper than a lens with smooth bokeh. And this might affect our perception of the DoF!
Then, if I use the 85mm lens on the Nikon D4 (affiliate) which has 16,16 megapixels,
and the 58mm lens on the Nikon D810 (affiliate) in crop mode (15,36 megapixels),
then I have two cameras which give me very similar resolution.
Two things which are affected when switching a lens between a full-frame camera body and a crop-sensor body, are:
– angle of view, and
– depth-of-field.
And if you move your position, your perspective changes too. That makes a direct comparison difficult too.
I used these lenses and cameras in two different ways to visually try and explain:
- Nikon D810 at full-frame (FX) and then set to 1.5x crop (DX crop), which effectively enlarges the image, but at a huge reduction in resolution.
- Using a Nikon D4 (at 16 megapxiels), against a D810 (36 megapixels) at 1.5x crop, i.e., 15,36 megapixels, with “equivalent” lenses.
And then, finally, we’ll have a look at what a Depth-of-Field Calculator says, and whether this corroborates what we see.
1. Full-frame vs 1.5x crop, using a single camera and “equivalent” lenses
Both these images were shot with the Nikon D810 (affiliate), but with the two different lenses. I had the camera on a tripod, and didn’t move it at all. I just changed lenses, and the crop in the camera.
Click on the image to bring up larger versions.
On the left, the full-frame (36 megapixel) image with the Nikon 85mm f/1.4 lens.
On the right, the 1.5x crop image (15,4 megapixels) with the Nikon 58mm f/1.4 lens.
Both lenses were used at f/2.8
Notice how the FF image on the left has shallower depth-of-field in the background. The city appears more blurred than the image on the right.
Keep in mind that the 58mm lens’ bokeh might have had a slight effect here, if the lens’ bokeh itself is less smooth. However, the background does look sharper.
I did have to really reduce the 36 megapixel image on the left to bring it to the same size as the lower resolution image on the right.
Just to give you an idea, here is how the actual sizes would’ve compared, if I made this composite’s size the same as the height of the full-resolution image:
The different sizes here are largely not relevant. Remember, the circles of confusion is defined for a specific print size when viewed at a certain distance. So if we bring the two images to the same size, then we see the way the DoF actually appears in the photo.
Now, if this does not seem all that convincing since we had to play with the image size, let’s look at how the DoF appears when used with two cameras with similar resolution:
2. Full-frame camera vs 1.5x crop camera, using “equivalent” lenses
This is where I used two cameras with nearly the same resolution, and two lenses who were very nearly “equivalent” to each other for the sensor size.
Top image: the Nikon 85mm f/1.4 lens on the Nikon D4. The D4 has 16,2 megapixels.
Second image: the Nikon 58mm f/1.4 lens which gives an equivalent of 87mm on the Nikon D810 when it is set to 1.5x crop in the camera. The D810 then has 15,4 megapixels.
I wanted to use the cameras on a tripod, but the cameras are at different height, so this made the tripod very clumsy when I tried to match the exact framing and composition. In the end, I hand-held the cameras, and tried for compositions as similar as I could. I tried to keep our model, Olive, the same size in the frame. So there is some difference to the images in that sense – the perspective is slightly different – yet, the DoF is noticeably shallower in the full-frame image.
Now, after all this, you may still not be entirely convinced that there is an “equivalency” in the focal lengths if you simply multiply the focal lengths by the crop factor. There were uncertainties perhaps in both the ways I set these comparisons. You might still wonder what happened to the argument that a crop sensor just gives you an “enlargement” of the full-frame image.
So let’s go back to the “armchair method”, and look at what a Depth-of-Field calculator says.
3. Depth-of-field calculator
For these, I used the Simple DoF Calculator app on my iPhone. I like it. But there are others available too. In any case, you should have a DoF calculator on your phone. It just makes sense.
This is an exercise that you can play with on your own, using DoF tables, or a DoF calculator.
It works this way every time – the full-frame camera gives us shallower DoF, for the same field of view, than the crop-sensor camera does.
Summary
We bring it right back to where we started:
- Yes, a 50mm lens does indeed behaves like an equivalent focal length of a 75mm lens (on a 1.5x crop sensor), or an 80mm lens (on a 1.6x crop sensor) … but, the depth-of-field increases by about a stop.
- Yes, a 100mm lens will give you the same perspective as a 150mm / 160mm lens, if you don’t change position … but, the DoF increases.
So, yes, a 50mm does (kinda) act like a 75mm / 80mm when used on a crop-sensor camera. The focal length is effectively increased on a crop-sensor camera. But at the loss of the shallow DoF that a larger sensor gives you.
How important that change in depth-of-field is to you, is something you need to decide for yourself. For me, I like that extra control that shallow DoF gives me in creating a specific look to photographs. While the change in DoF might seem incremental when compared in a final image, it is an element that I do want to be available to me – the shallower depth-of-field which full-frame cameras allows.
Related links
- Using tele-converters: extra lens compression for tighter portraits
- Making your images pop – through choice of lens
- Composition for full-length portraits – step back!
- 85mm – the best lens to change your portrait photography
- Available light portraits – composition, light and style
A little bit of homework #1 – The armchair version
Let’s change it around now. What would the effect be on:
– the field of view,
– the composition of your image,
– the depth-of-field,
if you used the same lens on a FF camera and a crop-sensor camera, but you moved your position to keep your subject the same size in your viewfinder?
A little bit of homework # 2 – Practical exercise
Better yet, do it as a practical exercise:
Shoot the same portrait of someone (who doesn’t move),
using a FF sensor camera, and then a crop-sensor camera,
while keeping to the same framing of your subject.
and also keeping the same perspective. (i.e., You can’t move.)
Now compare the DoF between the two images, for the same resolution.
Let us know what you find.
Addendum & clarifiction – as a reply to many of the comments
I do agree about the focal length remaining the same. As mentioned several times here, a 100mm lens remains a 100mm lens. It’s physics.
Anyone who follows this website will know I’m also a big supporter of using correct terminology. It helps with clarity if the same things are named the same things, and not obscured by everyone making up their own phrases.
The counter-point to this – and the main reason I wrote this article – is that it is easier to understand what happens with the FF / crop switcheroo, if you think in terms of the effective focal length changing. It really is, even if this makes you grind your teeth.
So while we could remain sticklers for the mathematics of it all – e.g., a 100mm lens remaining a 100mm lens – it is easier to understand what happens if we think in terms of the field of view changing = effective focal length being different.
If we remain in one spot, then using the same focal length lens (say a 50mm lens), on the FF camera and on a crop-sensor camera, makes little sense to a portrait photographer. You just get a crop version. This of course makes sport photographers and wildlife photographers very happy.
Now if we think of the difference in using a FF camera and a crop-sensor camera, and that this effectively brings in that 50mm / 75mm “change”, and we frame our subjects the same way on the FF camera, and with a crop-sensor camera … then there is this effect as described in the article. Call it a change in field of view … call it a change in effective focal length … there is something happening here that is more than just a crop of the image.
I honestly think that considering this as the focal length effectively changing on you, is more meaningful in a practical shooting sense, than thinking purely in terms of the mathematics of the optics.
I hope this makes my intent with this article more clear.
The DoF increases exactly the same crop factor in fstops.
nice write up and you got it right :)
This is a topic that, as you mentioned, is often discussed and argued in various online forums.
I shoot both crop and full frame and keep my crop for specific reason-the loss of shallow dof. In dark venues (weddings, etc) I often use my 50 1.2 or other large f/stop prime. With full frame, the dof can be razor thin but place that same lens on a crop sensor, and you get more dof at the same f stop. There is also the great Sigma 1.8 zoom lens that gives the benefit of taking in light like a 1.8 but with dof like a 2.8 on full frame. This benefit will only improve over time as noise gets better on crop sensors.
Great read!
Thanks for this article, message quite simple and clear !
Great article. I second rudy’s opinion about being able to gather more light and keep reasonable DoF at the same time on a crop body.
Very interesting perspective. it seems to me (correct me if I am wrong)-if I take a photo of a subject at distance x feet using one lens and both a DX or FX camera, the image will take up exactly the same area in cm^2 on either sensor. It will however take up a larger portion of the total area on the DX sensor. Since DOF is a function of aperture opening and magnification, under the scenario I just described, I would expect the DOF to be unchanged. A confusing subject, to be sure.
I have always known this but have wondered how the telephoto shortening or compression effect is different between crop and full.
I agree with EVERYTHING in this article, besides for the line where he says perspective doesn’t change when you zoom……That is just not true. Perspective does change in many ways when you zoom.
Other than that, this is really good information.
Zach … do a few test shots and show us how the perspective changed when you zoomed, while you remained stationary.
“One argument goes along the lines that the crop sensor is just that, a crop. An enlargement. That nothing changes – you just get less of the scene. And that there is no “equivalent focal length” when you go to a crop sensor camera. But what really happens is more complex than that.”
Yes, there is no equivalent focal length!
Obviously if you put the same lens on a FF and a smaller or bigger sensor, the only difference is…well one is cropped. Then if you want the same Frame of View while keeping the same perspective you’ll need different focal lengths on the different sensors. But if you’re keeping the same distance then the DoF will be different from the different focal lengths. Keep in mind that different focal lengths used at different distances to obtain the same magnification of a subject have the same DoF. Since you’re instead at the same distance, you’ll have different DoF characteristics. Different DoF = Not Equivalent. There is no equivalent focal length, only crop. If you want shallower DoF at the same distance and perspective, use a bigger sensor.
(Assuming you’re at the same f-stop of course)
Matthew … what you describe there about the DoF not changing, doesn’t match what we’re seeing here in the photos.
It also doesn’t match what the DoF calculator tells us.
DOF is similar for say 85mm f2.8 (on FX) vs 58mm f1.8 (on DX), however the background blur will be greater on FX, blur circles will be bigger on the FX due to the longer focal length used.
There is more ‘separation’ between subject and background with a larger sensor with same DOF, and equivalent focal lengths and framing.
I have been through this with FX vs MFD (36×48) trying to justify an upgrade (or not), and found that yes a larger sensor does give more background blur even at the same DOF, but that for the cost it was not worth it.
A couple of minor but important things – could you change equivalent focal length to equivalent field of view? a 100mm lens is a 100mm lens regardless of sensor/film size unless you have a magical means to violate the laws of physics.
Second, when comparing DoF you need to alter the focal ratio (f/#) to account for the change in focal length. A 10mm lens at f/1 has the same aperture as a 100mm at f/10. f/number = focal length divided by diameter of aperture.
Aha! Thank you for the explanation. Nothing is as simple as it seems.
It is errant to use the same f/stop when comparing the full frame and crop sensors, because a lot of light is being wasted using a full frame lens on a crop body without a focal reducer. The concept of “equivalent focal length” applies to aperture as well, not just focal length.
It would be more accurate to compare a full frame body and lens to a crop sensor body with a focal reducer and full frame lens. The real reason that crop sensors typically have a larger DOF is because manufactures market crop sensor cameras as lightweight and travel friendly, and do not make huge lenses for them.
It would be more accurate to use a crop
Niel…
Great article on DOF and I commend you for the effort you put into putting it together. Of course we all know that
first we have to have a great photograph and the we can develop our skills on DOF.
Nice article, and a subject that has intrigued me for a long time.
In photography everything is a trade off, and another subject worth discussion is pixel density, ie, when you crop heavily into an FX image you lose print size, so best you always compose/ fill frame as much as possible.
All the best.
Leon
If we take a shot with, say, a 50mm lens on FF and then another shot with the same lens and aperture from the same spot, focussing on the same subject using a crop sensor camera (with the same pixel to sensor area density) then we crop the FF image (in print or on screen) to match the crop sensor image would we expect the DoF to be the same? According to a DoF calculator a D4 50mm f8 @ 3m has a DoF of 1.85m. D7000 (crop sensor) 50mm f8 @ 3m has a DoF of 1.17m. This suggests that the answer to my question would be that the crop sensor in that scenario would have less DoF. I don’t have a FF camera to test this and I may well be misunderstanding something so apologies if this is irrelevant nonsense.
No Ben you’re right. Only the field of view changes. It’s not confusing or complicated. A 100mm lens is a 100mm lens no matter the body you put it on. The minimum focal distance is still the same no matter the body. If you’re photographing and want to achieve the same field of view as the crop sensor with a FF, you will have to step in and therefor you will change the perspective, and DOF while trying to cut the edges of the frame or shrink (for lack of better words) your field of view. This changes the depth of field and essencaily gives you a completely different photo. There is no math needed here. Only the field of view changes between crop vs full. That’s it, that’s all. A few years back I rented a ff so I could get a better understanding. All the misinformation out there really had my head spinning. After experimenting I got it. It’s simple, and there is no “equivalent” or math involved.
I wonder what kind of math equation I would need to figure out a medium format vs crop sensor? None! The focal length of a lens is the focal length of a lens.
I think you have to calculate the DoF for the FF at the crop-equivalent 75mm. That will give you a DoF that is 0.76065m at F/8 which is equivalent to F/5.3 on a crop-sensor.
It’s a great article!
However, Neil, I would like you to mention in your article the following:
1. You’re are saying that the DOF on FF is about one stop less than on crop. It means that the same DOF on FF will be at crop’s f-stop+1stop. For examle, f2 on crop = f2.8 on FF. It is well known that optical characteristics of lenses increase when they’re stopped a bit. So it is said that f2.8 on FF makes a better overall image than f2.0 on crop. It would be great if you check it out.
2. You were very clear when you wrote that you like extra control that shallow DoF gives you. I don’t have FF, I have a 1.5 crop camera. On wide angle lenses (up to 55mm) sometimes I feel the lack of shallow DOF but on 70mm and higher I tend to stop-down my lenses. Recently I acquired a fabolous FA* 85/1.4 (Pentax). Even on crop @f1.4-2.0 it just melts away all the background. Sometimes I have to shoot even at f4.5 in order a viewerer can find some hints what was in the background. What are your thoughts on this? I would like to know how often do you really use f1.4-2.5 in your work?
I’m still confused here. From what I understood, a cropped sensor causes a lens to stay the same, just with an equivalent focal length. I don’t know if it makes sense, but about the depth of field I always saw it like this: if you use a 25 on a MFT camera, the equivalent focal length is 50, but the DOF is that inherent of a 25, which is greater (I don’t know if this is the cause or the effect here).
Anyway, what about compression? We say that a longer focal length compresses distances and vice versa. Longer is generally more flattering for portraits. Does is mean that it I use my 45 1.8 on a MFT I will have the compression similar to a 50 on a ff body? Or does it multiply?
Same thing on the other end: does a 12 mm lens exaggerate distances like a 12 or a 24 on a ff body?
Nice article. But does it really matter? With the new blur filter in Photoshop CC 2014, you can blur the background, if you would like it more blurred than the lens does. What I would be more concerned with would be the over all quality of the image. Is full frame better than the crop frame? Less noise? Sharper image? I would guess that for what most photographers do, no. For portrait and wedding photographers I would say, it makes no difference at all.
Of course you can do a lot in post. But you need the skills to make it look natural and, in case you have a lot of pictures, it takes additional time.
Principally, a FF lens needs to resolve less line pairs/mm to get an equally sharp image. Since for the same FOV the resolution of the whole picture is multiplied by the dimension (e.g. lens: 50lp/mm -> 50lp/mm*36mm line pairs/picture height). As you have larger pixels, too, you get less noise for the same AOV. For us it makes a difference, as clients want large pictures, and most are taken during the golden hour at larger ISO.
Joe … I would agree that for the majority of purposes, this difference doesn’t matter that much.
For example, this wouldn’t turn me off the Fuji XT1 and 56mm f/1.2 lens, knowing the DoF is now equivalent to something else. It just doesn’t matter that much in terms of the overall artistry of any particular photograph.
Still, this article has value in that it helps us (including myself), understand what is happening.
Nice comparison but I use both Fuji crop and Canon 5D and the dof/bokeh is much more pronounced on FF. Try this same test with some nice trees in the back of the subject which are close to the subject not far like your city skyline and you will see a difference. Plus you are using the same camera for both examples so I’m sure they will be very close as you are not really using a crop sensor.
Sebby, I used the Nikon D810 in the 1.5x crop mode. So it is in effect a crop sensor camera.
This is especially so when you compare the 1.5x crop of the D810 which is 15 megapixels, to the full-frame 16 megapixels of the Nikon D4. (I really tried to make things as equal as I could in these comparisons.)
Thanks Neil. Ths is a great comparison and meshes with what I have felt for years. I own both full frame and crop sensor cameras.
This is making me think that I should use my crop sensor cameras for my macro work and my full frame for portraits and weddings. Am I off here?
Thanks again!
It makes sense, and you have a point there.
I have no idea how diffraction would kick in for FF and crop-sensor cameras at the respective small apertures. That might be another consideration when working with small apertures with FF / crop. If someone else has real experience and knowledge of this, it might be interesting.
Neil,
Diffraction comes on earlier for crop sensors.
https://www.cambridgeincolour.com/tutorials/digital-camera-sensor-size.htm
You have to scroll down a bit to find the calculator.
Another way to test this is to use the same lens on two different size sensors (at the same distance to subject and same f/ stop). Then crop the larger sensor’s image to match the smaller sensor’s image. Using two different focal lengths (i.e. an 85mm vs. a 58mm) on two different size sensors at the same f/ stop, only equates to the same field of view (approximately) and not the same DOF. The difference is more in the focal lenght then in the sensor size, unless of course one is trying to match field of view. An 85mm f/1.4 is a 85mm f/1.4 on any size sensor at the same camera to subject distance.
You and some other have said the same thing and may have missed it where Neil says so in the opening under the summary section. Of course an 85 is an 85 regardless of camera but the field of view changes. Another way of saying that is equivalent focal length. If you want the same photo from a crop and full frame, you have to use equivalent focal length and not the same lens. Equivalent focal length on a crop is a gain of dof with a crop sensor.
So from the same distance, an 85 on a crop will give the equivalent of a 130ish mm lens on a full frame. You end up with a different photo just like you would if you back up with the 85 on a crop to match the full frame field of view.
As a reply to many of the comments above:
I do agree about the focal length remaining the same. As mentioned several times here, a 100mm lens remains a 100mm lens. It’s physics.
Anyone who follows this website will know I’m also a big supporter of using correct terminology. It helps with clarity if the same things are named the same things, and not obscured by everyone making up their own phrases.
The counter-point to this – and the main reason I wrote this article – is that it is easier to understand what happens with the FF / crop switcheroo, if you think in terms of the effective focal length changing. It really is, even if this makes you grind your teeth.
So while we could remain sticklers for the mathematics of it all – e.g., a 100mm lens remaining a 100mm lens – it is easier to understand what happens if we think in terms of the field of view changing = effective focal length being different.
If we remain in one spot, then using the same focal length lens (say a 50mm lens), on the FF camera and on a crop-sensor camera, makes little sense to a portrait photographer. You just get a crop version. This of course makes sport photographers and wildlife photographers very happy.
Now if we think of the difference in using a FF camera and a crop-sensor camera, and that this effectively brings in that 50mm / 75mm “change”, and we frame our subjects the same way on the FF camera, and with a crop-sensor camera … then there is this effect as described in the article. Call it a change in field of view … call it a change in effective focal length … there is something happening here. And this is what happens then.
I honestly think that thinking of this as the focal length effectively changing on you, is more meaningful in a practical shooting sense, than thinking purely in terms of the mathematics of the optics.
I hope this makes my intent more clear.
I had somehow figured that x1.5 crop factor should be applied to both focal length and aperture. However, Rudy makes a good point – the aperture value continues to play its role at the stated value even on a crop sensor camera, while the dof is at an equivalent or “apparent” value. That is, a 50mm f/1.4 lens @ f/1.4 on a crop sensor will allow light of 1.4 while dof will be that of f/1.8 (0.5 stops more on a x1.5 crop sensor) and focal length will be 75mm.
Neil, was the shutter speed the same for both pics in 1 above with f2.8?
The exposures were exactly the same. i.e., same shutter speed.
Thanks for the interesting article and all the work you’ve put into demonstrating the results.
If I understand you correctly, you still seem to be saying that an equivalent lens will produce pretty much the same result if you also adjust the aperture for the crop factor.
So in the last example with the two different sensors, the 75mm and 50mm lenses would have achieved the same DOF if the 75 was at f/2.4 and the 50mm at f/3.6?
My understanding is that the apertures are equivalent in terms of exposure but must be adjusted for the crop to determine DOF. Is that so?
Right conclusion but the numbers are the other way around I think Richard. 75/3.6 and 50/2.4 both give an aperture of 20.8mm diameter, which will generate the same DOF (all other bits of the capture and viewing chain being equal).
Richard … that’s the gist of it. But as mentioned elsewhere, it is probably not important enough in the grand scheme of things, to chase after.
Just simple question, is it possible to choose between FX and DX format at FX camera like Nikon D810 using FX lens? I always wondering it and can”t read it anywhere.
Anyway very informative article.
Yes, that’s exactly what I did here with the D810. I used it as an 1.5x crop camera.
Thanks Neil for another outstanding post, always a pleasure to catch up with what you are doing. At client level, I don’t think many would ever spot the difference, but of course both FX and DX formats have their merits. Yes I own both formats, using Nikon’s incredible D4 with a D7000 as a back up. I do find the DX model useful for some aspects of my work.
Neil, would the test have turned out if yoiu had actually used a crop Nikon, say the D7000, or D7100, and not a FF camera set to Crop Mode? Just curious
In what way could it be different? (Assuming same resolution.)
I’ve seen this topic come up a lot over the past year with the same questions from various people. Neil you wrote a great article to convey what happens when comparing a full frame sensor to a crop sensor with an equivalent “Field of View” lens.
So to assist here is my example:
A 50mm f/1.4 is still a 50 f/1.4 whether it is mounted to a full frame or a crop sensor camera. The lens doesn’t magically change to a different lens. What changes is the “Field of View” and the “DOF” between the two sensor sizes.
On a full frame the “FOV” and “DOF” of the this lens will be just what it states: 50mm f/1.4. On a crop sensor, I’ll use the Canon 1.6x here since that’s what I shoot, that same 50 f/1.4 has a “FOV” equivalent of a 80mm on a full frame and the “DOF” of f2.24 equivalent (f/1.4 x 1.6x) on a full frame. Yet you are still getting the light input of the lens at f/1.4, it is just the aesthetics of the “DOF” changes slightly based on sensor size. Hence why a 1.6x crop has less shallow DOF.
Hope this helps :)
Thanks for this great summary. Up until now I’ve only had this other site on Equivalence to point people to, which of course was too long for people to read. I don’t know why people argue about this so much on forums, it’s not like we go to a shoot and think do we want to use this lens on the crop or the other on the FF.
What’s not being addressed here is perspective foreshortening as you change from FF to crop.
Tatsu … foreshortening of your perspective … as you move and change position? or don’t move?
Expand a bit on your observation. (Photo examples would be especially nice.)
Hi Neil. Another term is telephoto compression. We know that portraits are used with long lenses to flatter our subjects, because of the “slimming” effect telephoto yields. What I am wondering is, do you lose some of this perspective distortion? See: Perspective Distortion
There are 4 photos of yellow bars on the page taking at different focal lengths and correspondingly different distances to compensate. However, the sensor size is the same on all four photos, so it’s not a good example of what I mean. The compression is obvious though. In your photos, it is not.
Another way of asking: Does a crop sensor at 58mm has the same telephoto compression effect as a full frame at 85mm at the same distance? I’m going to guess yes.
That’s the difference between what they illustrate – Perspective Distortion,
and what I tried to illustrate – that the field of view remains the same between “equivalent” focal lengths.
They changed position … i.e., the perspective changed.
I didn’t.
So back to your original comment – there is no “perspective foreshortening” when I changed between FF and crop, because *I* did not change position.
Ok, thanks for confirming that FF and crop has the same perspective distortion at the same distance with FOV equivalent lenses.
This also exposes another related myth: that wide angle lenses distort faces. They don’t, it’s just that to get the face to fill the frame, when using a wide angle, you will need to move closer.
Thanks for pointing me to the Cambridge In Color DoF tutorial. I’ll have to make a note of this to send to anyone who wants to start a discussion – they’re better off to begin there than with measly me. As for bokeh changing the PERCEPTION of DoF this is an important point. Many people should look up the discussions down the centuries about how framing an oil painting made it look different even. There is no such thing as an objective “look” from a human. Seeing is not with out eyes, but 95% (at least) of our “visual” perception is actually a computational process in the afferent nerves and then the visual cortex.
I don’t think that “equivalent” lens is a fair term and may even contribute to the confusion in discussions. I think a better statement would be something like, “a 50mm focal length on a 1.5x crop sensor would provide a similar angle of view to a 75mm focal length on a full-frame sensor.” The key term is “angle of view” because on a 1.5x crop, the physical distance is 50mm (the “actual” focal length).
In terms of dependence:
DoF is dependent upon the focal length and the distance of the subject to the focal plane.
Angle of view is dependent upon the sensor size and the focal length.
So, with the above dependencies in mind, for different sensor sizes to have the same angle of view, the focal length must be changed (or the distance of the focal plane to the subject). If the focal length is changed (or the distance of the subject to the focal plane), then the DoF is changed.
Thanks for the article!
Can you get the same result with a crop sensor simply by opening the aperture up one stop? E.g. would your city background shots have looked the same if you’d opened the 58mm up to f/2.0? I.e. is it only at wide open that FF wins?
Tom … the change in DoF appears to be in the order of one stop. (Perhaps 1.3 stops.) So in theory you could gain the same shallow DoF with a wider aperture.
However, you run out of choices when you are shooting wide open already.
Ultimately though, I don’t think the change in DoF is substantial enough. In other words, if you are happy with the DoF at a certain aperture, then a stop less DoF isn’t a dramatic change. It is noticeable in comparison, but it isn’t huge.
Nicely written and demonstrated Neil. I abandoned FF cameras and adopted the m4/3 platform a couple of years ago for various reasons. This subject is discussed/argued Ad nauseam on various equipment forums. The 2X crop of the m4/3 sensor dramatizes this phenomenon even more than the APS sensors found in the common DSLR models. As has been pointed out by others, one thing that can be pointed to as an advantage to a crop sensor is the extra sensitivity gained by faster glass with still usable, or at least not so razor thin, DOF. I love using my Voigtlaner 42.5mm f/0.95 wide open – understanding that for the sake of these kinds of discussions I’m using the “equivilent” of an 85mm f/1.8 on a FF camera. Having the ultra fast aperture while still getting both eyes in focus is very useful. The common reply is that on the FF camera you simply stop down and raise the ISO to get the same results – true enough I’m sure. Anyway, thanks for the article.
Between Fx and Dx … To achieve same perpective, we need same distance. Now the problem is framing .. .
With Fx, to framing like Dx, they need longer focal length for same distance perpective.. Longer focal lenght = more DoF Bokeh …
Otherwise Dx to framing like Fx, need shorter focal length for same distance perpective .. Shorter focal length = less DoF bokeh …
——————————-
So … With same distance same framing .. because of diferent focal length, Fx winning in DoF Bokeh.
With same focal length and same framing .. Fx winning in Perpective
——————————-
same focal length = same DoF Bokeh
——————————-
correct me if i am wrong
Dani .. you’re going to have to work through the article step-by-step, and check for yourself.
Start here: shallow depth of field is NOT the same as bokeh.
Also, I’m not sure what “winning in perspective” would mean.
This was one of the reasons I switched from APS-C sensors to full-frame back in 2008. The other reason is one you don’t go into in this article and it is that on an APS-C sensor camera, lenses show the distortion of their actual focal lengths and not their equivalent focal lengths. For example, on a camera with an APS-C sensor, a lens with a FF equivalent focal length of 24mm displays as much distortion as you would get from a 16mm lens on FF (cropped to match, of course).
So I’m wondering, as I am one of those people chasing that equivalent backround blur on a full frame, but with my crop sensor…
I am wondering, if you shot at say, 200mm on a full frame at 2.8, and you wanted to emulate, not duplicate exactly, but get a similar creamy, out of focus background on a crop sensor camera with a slower lens, how much longer of a lens would you have to use on that crop sensor camera to compress the background more, in order to get a similar style / look to the background blur?
I know that it wouldn’t be the same framing, but at 200mm on a full frame at 2.8, the background, if it is far enough away, turns into a creamy, dreamy, buttery blend of specular highlights and colors. How would I get a similar effect with a crop sensor and a slower lens? Since I know I would have to use about a 135mm lens at 1.8 to get a similar effect. But if I were to use a longer telephoto lens, say a 300mm lens on my crop sensor at f/5.6 or f/8, would that be a similar blur?
Todd .. it’s not that easy a comparison. Even if the DoF was similar, the perspective and angle of view would be different. It would be a different photograph.
The best comparison would be between a lens like the full-frame 24-70mm f/2.8 and the Sigma 18-35mm f/1.8 DC HSM.
Then the f/1.8 on the crop-sensor roughly translates into the same DoF as the 24-70mm f/2.8 would show at f/2.8 for full-frame.
Why is it wrong to have the whole picture in focus…I don’t understand why a camera that makes IT more difficult to achieve a good DOF would cost more and is thought to be better much better .
… because selective focusing / selective DoF gives a specific look that is very attractive, and can help draw attention specifically where you need it.
In other words, with photos like this, an aperture f/22 would give a cluttered image …
… as the pull-back shot at f/5.6 already implies.
Also, for the same level of sensor technology, the larger sensor will have better high-ISO performance.
Anyway, for portraits and similar niches of photography, shallow DoF is very useful technique.
That said, the difference between full-frame and crop-sensor is arguably not that huge, and is incremental. Still, having a larger sensor gives you options with regards to how you want to use DoF and selective focus.
My question is related to your “homework” assignment, and even after reading this in detail I am a bit unsure of the answer.
I cannot directly compare myself without serious effort because one of my cameras is film (I have a 35mm Canon AE-1, and a Fuji X-A1), and I would have to burn the rest of the film and have it developed.
I happen to have a 50mm lens I use for portraits on the X-A1, and I have a 50mm lens that I am starting to use for portraits on the 35mm film.
My question is, if I shoot a portrait from the same spot with both 50mm lenses, will the faces look the exact same? I realize that I may see more overall in the frame, but will they have the same proportions and be just as flattering as each other?
I realize that I may have to step in closer with the 35mm to get the face to fill the frame the same way, changing my perspective, but I am asking what if I don’t do that… In the end, I want to know if my 35mm camera will produce portraits that are less flattering than my Fuji?
Update to my post.
I carried out the homework assignment to try and answer my question.
I shot some of my daughter’s toys on our piano, and placed a book farther in the background. Then I shot with my Fuji AX-1 and set up a shot on a tripod as if I were shooting a portrait of the toys. I set the aperture to 5.6 (which is the widest on the 50mm kit lens) and the ISO to 400 to match the film settings as closely as possible. The camera then determined the shutter speed. I focused on a spot I could remember, and centered the camera on a spot I could also remember. I used the timer and took a frame.
Then I left the tripod in the same location and replaced the Fuji AX-1 / 50mm lens with my Canon AE-1 (35mm film) and it’s 50mm f1.8 lens. set up the focus and center on the same locations without having moved the tripod. Then using the built in meter I adjusted the shutter speed manually to get metering of f5.6. Then I set the f stop to 5.6 manually to ensure the program wouldn’t change it’s mind if the condition slightly changed by the time I took the picture. I used the timer and took a frame.
I also then adjusted the tripod to “reframe” the scene so that the field of view of the Canon matched what the Fuji had shown (which of course required that I get closer) and changed my perspective. I took a third shot this way.
I haven’t burned the rest of the roll from the analog to have the prints yet. But I will tell you what I think the answer is based on what I was seeing in my 35mm viewfinder in comparison to the Fuji. I believe the proportions and whatnot between the two cameras/lenses to be exactly the same. It literally only affected “Crop”. Like I could crop the analog shot post production and end up with the exact same flattering image the 50mm had on the fuji.
I also noticed that for the third shot, my depth of field got smaller as I moved in, and the perspective changed so that the objects appeared what I would call less flattering.
So yes, the 50mm lens produces the same flattering shots regardless of sensor or film, and with the same proportions. So the terms “equivalent” or “effective” focal length are really misleading and not true. A 50mm focal length on one sensor is not “equivalent” to 85 on another in any way other than a literal crop factor.
Does that seem right? I will try and update this if I think otherwise when the film is developed.
You changed your perspective.
You changed your perspective, as you again mention … and you noticed other differences as well.
Now, as you noticed, between your first two shots, everything looked the same … except the crop was tighter. So yes, a 50mm lens remains a 50mm lens. As stated in the article.
What your test is missing out on, and this is the entire motif in this article … you need to compare a 50mm lens on the 1.5x crop sensor, against a 75mm lens on the full-frame camera. Then you might see for yourself that the words “equivalent” and “effective” are necessary short-hand to describe what is happening.
Soooo … you have more homework to complete still.
Just so I understand this correctly , A crop camera will never match a full frame camera for shallow DOF. It will only match the frame size. For example 85mm 1.4 on FF has same FOV as the crop camera with a 58mm 1.4 lens from the same distance however the DOF is reduced by about a stop. it is all to do with distance from the focal plane and nothing else, a FF camera will always allow you to get closer to your subject when composing for FOV. Therefore a FF camera will always yield more shallow
DOF. Proof of this is put the same 58mm 1.4 lens from the crop camera on the FF camera and you will be able to move a considerable distance nearer your subject whilst maintaining a similar FOV and as you are closer to your subject, the DOF will be narrower.
Not that this forms part of the debate, but it seems to me whatever lens you put on a crop camera you will always yield a narrower DOF when using the same lens on a FF camera. or I may be mistaken :-)
Will a crop camera ever match a full-frame camera for shallow DoF? Perhaps.
Let’s consider the Sigma 18-35mm f/1.8 lens Canon mount / Nikon mount:
If we consider that f/1.8 on a crop sensor gives us about the same DoF as f/2.8 on a full-frame sensor (at equivalent focal lengths), then we’re on the same playing field.
And if we do the 1.5x crop multiplication to get to an equivalent focal length, then we see the 18-35mm lens gives us the same field of view as the 27-52mm lens on a full-frame camera. It covers most of the range of the 24-70mm lens.
NOW it gets interesting and complicated:
The wider aperture gives us a faster shutter speed … but it could validly be argued that you need that higher shutter speed to compensate for the higher risk of camera shake due to the cropped sensor’s narrower field of view. Think of the angle you’d have to move the camera for the same “size” of blur from the shake.
Alternately, you can also consider that gain of 1.3 stops in aperture, to be eaten up by the poorer high-ISO performance of the same level of technology to the sensor (and the same resolution for the sensors) … so you gain nothing there by the wider aperture. It just compensates you for the crop sensor by trading off other advantages / disadvantages.
So to answer your question in short – yes, it is possible for a crop camera to match a full-frame camera for DoF, if you choose a lens that has a wider aperture.
I know … this feels like a spiraling journey down the rabbit hole where you can get lost in all of this.
I really like your article. But in the the summary you have mentioned that a 100mm on crop sensor will give the same perspective as 150mm on a full frame. I think perspective is determined by the distance to the subject. So it is better to state that a 100mm on crop will have same field of view as a 150mm on full frame. Then again you will be at the same distance to achieve the same framing on both crop using 100mm and full frame using 150mm. So you aren’t wrong either. Still, not associating focal length with perspective will create less doubts among readers. Keep up your good work.
Thanks for (I think) a very correct and informative post. As I have too little time to really digest this info, BUT as I do trust your judgment, may I ask for an executive summary on the following dilemma of mine? For creative shots (shot wide open), should I go for an FF 85mm F/1.4 or an APS/C 56mm F/1.2 (Fuji’s)?
Background: I have both a FF DSLR (Nikon D700 that I will upgrade to a D750) and a Fuji X-M1 (that I am thinking of upgrading to an X-T1 /T10, currently used as easy all around camera with the 18-55 OIS 2.8-4.0), also an old Canon 5D (and 85mm F/1.8 that I should sell..). I do like the small form factor of the Fuji X, especially as it enables small wide angle lenses given that there is no mirror. If money were no object I’d spurge for a Leica and shoot a Nokton .95…
Neil -thanks , appreciate your concise answer. Somewhat related, then: how much better is the Nikon 85mm F/1.4 vs the Nikon F/ 1.8 bokeh ? Both are highly rated optically, but the price difference is quite significant, about $1,000.
Andre – the Fuji 56mm f/1.2 (which is an awesome optic, btw), will give you about the same DoF as an 85mm lens used at f/1.8 (for the same subject size in the frame).
That wouldn’t put me off using the Fuji though – the difference in how the background blur appears between those two is such that most people won’t notice the difference. You’d have to see side-by-side comparisons to be aware of it. So if the smaller size of the Fuji cameras appeal to you, let that help sway it for you.
Neil – thanks for your concise answer, much appreciated!
The earlier version – the Nikon 85mm f/1.8D, had harsh bokeh. I didn’t much like the lens. The new lens, the G version has surprisingly good bokeh. So if you’re on a budget the Nikon 85mm f/1.8G is good value for money.
Here is my review of the Nikon 85mm f/1.8G
Great – thanks for added info
“To take this tutorial away from the armchair to an actual shoot with comparison examples, the ideal would’ve been to have a full-frame camera that has exactly the same resolution as the crop-sensor camera. Just to keep things simple. And I’d have to have two lenses which are exact 1.5x versions of each other.”
Here is what I suggest: D300s, 12MP, DX (70/2.8) v. D700, 12MP, FX (105/2.8). Set a tripod and a stool for the subject then just swap the cameras. I’d recommend a camera to subject distance of 2m and subject to background distance of 10m.
The image won’t be cropped. The angle of view is almost identical. The image scale in a final print say 8×10 will not change. What will change is the background blurring which can be an important esthetic difference.
How about an article on DoF v. Background Blur as they are not quite the same and DX v. FX v. 645 as the format matters.
Dave
Hi Neil,
the following:
is incorrect.
It should read:
How did you come to that conclusion – that DoF doesn’t change as you change camera bodies?
(You’re not factoring in how DoF is determined.)
Well written and with lovely examples.
As has been pointed out, though, rather than saying that depth-of-field increases by ‘around a stop’ you can be more specific and say that the equivalent aperture increases by the crop factor (something that will be backed-up by a D-o-F calculator).
Quite simply, this ‘equivalent aperture’ is the f-number given by dividing the equivalent focal length by the aperture diameter.
Obviously equivalent apertures can’t be used in exposure calculations because the dominant way of describing exposure is based around light-per-area, not light-per-whole-image.
As soon as you look at it from a whole-image perspective, you’ll see there’s no such thing as a ‘depth-of-field’ advantage to one system over another: shoot equivalent apertures at the same shutter speed and whatever ISO is necessary and, on most modern cameras, you’ll get the same depth-of-field and the same noise performance.
Richard … a good point.
With this article, I tried to keep it as non-confusing as possible, and cut through some of the constricting pure descriptions, such as “the focal length of the lens doesn’t change.”
Similar with what you describe about the aperture. I decided to keep it to the mention of the DoF changing, rather than the aperture effectively changing, since that would imply a change in exposure.
This is a very valid point:
“Shoot equivalent apertures at the same shutter speed and whatever ISO is necessary and, on most modern cameras, you’ll get the same depth-of-field and the same noise performance.”
No I don’t agree with the conclusions in this test. 58 lens on APS-C is not equivalent 85lens on FF, it is still exactly 58 lens and it still retains the full aperture 1.4 aperture (not 2.1). The DOF will only be shallower on FF, if the APS-C is unnecessarily forced to shoot the scene from the same distance from subject (which is what this test did).
He handicaps the crop sensors by forcing it to use a 58 lens versus the 85 lens on the FF and then makes the APS-C shoot from the same length from the subject! If the test was fair, both sensors would use the same lens.
And with same lens, for the same size printed crop of the subject scene, the FF camera will be forced to shoot from a closer distance to achieve the same scene, which then results in exactly the same DOF.
Dustin – check the Addendum to this article. I directly address your observation there. The focal length doesn’t change, but HOW people use the camera and lens, does change when shooting comparable subjects. That is where the crucial difference comes in.
Now, about this comment:
That’s an odd description – handicapping the crop sensor. As for your “fair test” comment – that would be an irrelevant comparison in this regard for you’d just have a cropped version of the same image.
Where this article comes in, and explains things, is with this typical scenario – shooting a loosely-framed portrait of someone:
Using FF camera, an 85mm lens makes a lot of sense for that kind of portrait.
Now let’s switch to a crop-sensor camera, and see what we can come up with:
Approach #1:
If you used the 85mm lens on the crop-sensor camera, and:
1.1.) You stayed in the same position, you would have a much tighter framing – hence a different photo. Nothing really there that we can compare.
1.2.) You moved position to get the same framing of your subject – then your perspective changes. Nice, but not really a comparable photo, since the size of the background elements changed. Their relative position would also have changed.
Approach #2:
If you wanted the same framing and same perspective (i.e., you stayed in exactly the same spot) with a crop-sensor camera, then you’d have to entertain the idea that you’d have to use an “equivalent” lens. And then you’d put the 58mm (or thereabouts) lens on your camera … and tadaaaa! Same framing and same perspective. A comparable photo.
And then this is where this article kicks in …. and you scroll back to the top, and sit back and read it, and actually consider what is happening here. No handicap.
Neil,
Thanks for the reply and wish you Happy Holidays. And I think you are somewhat balanced in viewing both arguments impartially, and though your Summary conclusions are technical correct the way they are phrased I still do not concede the argument.
Your conclusion stated:
“Yes, a 50mm lens does indeed behave like an equivalent focal length of a 75mm lens (on a 1.5x crop sensor), or an 80mm lens (on a 1.6x crop sensor) … however, the depth-of-field increases by about a stop.”
Answer: Yes your statement is technically correct, but I still think it is misleading because it assumes the crop sensor photographer must use a shorter focal length when in practice we may choose to slightly increase the distance from the subject instead and use exactly the same lens. A 50mm lens on crop sensor is still exactly a 50mm lens, it will have same DOF, same background blur, the only difference is the smaller sensor results in a crop of the scene, so you have a more narrow field of view.
The only reason the DOF increased on the crop sensor in your test, is because you used a 50mm lens versus the 75mm lens. If you used the same lens, from the same distance to subject, the DOF and perspective would have been the same. Again, the only thing that changes is the FF has a wider FOV. Now to achieve the same “framing” the crop sensor merely increases the distance to subject rather than being forced to use a shorter focal length except in some instances where photography can not vary the shooting distance.
1. Crop sensor versus FF, with same lens from same distance = Exactly same DOF and same background blur, narrower field of view.
2. Crop sensor versus FF, with same lens, with FF shot at shorter distance from subject so it can achieve same framing as crop sensor. = Crop sensor has less DOF than FF and crop sensor also has more background blur than FF.
3. Crop sensor versus FF,with different lenses from same distance from subject to achieve same framing, FF sensor will have less dof and more background blur.
The slight change in perspective in crop versus FF in achieving same framing to me is too insignificant for me to care about so I will not argue that point. But in the real world, a reasonable sized crop sensor like (aps-c) you can nearly always achieve same framing, background blur, and DOF as Full frame.
4. Slightly off-topic but interesting. APS-C can nearly restore equivalent FF field of view by using a FF lens with a .71 focal reducer and gain a 1 stop advantage in light.
For example.. here is my setup using only 2 lenses Nex 6 (APS-C).
Koinca Hexanon AR 35mm/f2.0
Konica Hexanon AR 85/f1.8
Zhongyi Lens Turbo II Konica AR to Sony Emount .71 focal reducer.
with 2 lenses I have the following Full Frame “equivalents” depending is I choose to use the focal reducer:
35mm*.71*1.5 = 37mm/f1.4
35mm*1.5 = 52mm/f2.0
85mm*.71*1.5 = 90mm/f1.3
85mm*1.5 = 127mm/f1.8
In Summary, I think it would be interesting for a part 2 of your article if the test was done using the Same Lens on crop versus full frame sensor with something like a 35mm or 50mm lens and see what the real world use and result is. I am claiming that the photographer will instinctively vary the shooting distance to achieve the desired framing, and would only reach for a shorter focal length if for some reason cannot increase distance from subject, and so in practice there will be very minor differences in the resulting photo. Crop sensor would actually have more background blur than Full frame sensor because shooting from a further distance. For reasonable sized crop sensor like APS-C, I think this statement will hold up.
Dustin, that would be a meaningless test because the Perspective would change. Different photographs with no real equivalency.
My test is just as relevant as your test in my opinion. In your test you use a shorter focal length on the crop sensor and compare it to a longer focal length on the full frame sensor so you are comparing difference in focal lengths not differences in the size of sensor but then claim the differences are due to the sensor size when they are not.
In my proposed test you use the same lens with same framing, Which then means crop sensor is shot from increased distance to subject and will have same or shallower dof and same blurred background due to the with only slight change in perspective. This is closer to real world, where the photographer will frame the shot anyway and can vary the distance to subject.
Your argument that it is not relevant because the perspective changes is wrong because there are certainly a large percentage of shots that a relatively small perspective change is not important or would be in the out of focus part of the image or part of a blurred background. In many cases, the change in perspective will usually be beneficial (portraits). Shooting APS-C instead of FF with same lens to achieve same framing will result in the same focal length, same aperture, same to slightly less dof, and same to slightly more background blur due to shooting from increased subject distance and the only notable difference is small change in perspective to background objects in the scene.
You are right that both sides of the argument can make valid views, because it is complex, and I happen to be on the side of the argument that a 50mm lens is a 50mm lens regardless of sensor size because all the properties of the lens stay the same. The only thing that changes in the final image is field of view if shot from same distance to subject or perspective if crop sensor shoots from further distance to subject. And IMO the minor differences are easily worked around by the photographer.
I need to change my original comment from WAY above. Using a crop sensor with the same lens, from the same spot and with the same settings as a full frame will NOT change the depth of field. The reason it can’t is because of physics/lens characteristics-they are static. I was wrong and have to suck it up :)
Scenario: If we crop a full frame photo to match the crop sensor photo (same settings and position), we will have identical depth of field. In other words, the sensor can’t change the outcome..
I still agree with Neil’s original findings but have to agree with Dustin as well:
Dustin wrote:
“My test is just as relevant as your test in my opinion. In your test you use a shorter focal length on the crop sensor and compare it to a longer focal length on the full frame sensor so you are comparing difference in focal lengths not differences in the size of sensor but then claim the differences are due to the sensor size when they are not.”
I have to agree that the loss in depth of field is not due to the sensor but due to a change in focal length to achieve equivalent framing.
In return, Neil wrote:
“I am comparing the differences in sensor size, when using the SAME FRAMING between different cameras (with different sensors). Then the differences in DoF are actually due to the difference in sensor size.”
While it may be semantics, this can’t be true. It has to be the change in focal length or we would HAVE to see a difference in depth of field between crop and full frame as described in my scenario above where we crop a full frame photos to match the framing achieved by the crop.
The disagreement here is how the data is interpreted even if the results can be agreed on. Since the sensors and lens in my scenario above was not changed and we achieved identical depth of field in the photos and Neil’s change in focal length resulted in less depth of field for the crop, it has to be the lens. We know this because it was the independent variable in the test. The change in independent variable lead to a change in the dependent, the depth of field.
Is the crop forcing you to use a wider focal length? Yes, but you could have shot a pano to achieve the same framing (using the same lens as the full frame) to achieve the full frame depth of field.
Bottom line, the sensors does not change anything except it is forcing you to use a different focal length/different lens. We know that changing lenses or focal length can have an impact on depth of field.
Rudy, check my comment #72 below.
I’ve also added the Little Bit Of Homework #2 for you … and everyone else.
Doing it as a practical exercise will be very useful in understanding this.
Dustin wrote:
“In my proposed test you use the same lens with same framing, Which then means crop sensor is shot from increased distance to subject and will have same or shallower dof and same blurred background due to the with only slight change in perspective. This is closer to real world, where the photographer will frame the shot anyway and can vary the distance to subject.”
How come that the crop sensor “will have same or shallower dof and same blurred background” with increased distance to subject when using same focal length and same aperture? Isn’t it the opposite?
Dustin, to answer you again on this:
I am comparing the differences in sensor size, when using the SAME FRAMING between different cameras (with different sensors). Then the differences in DoF are actually due to the difference in sensor size.
I think you need to step back a bit, and quietly read through this article again without such a hard mind-set.
My consideration that I want to compare what happens when we have the SAME FRAMING / SAME FIELD OF VIEW … would imply no change in perspective. This makes it 100% relevant, because this is what this article is considering.
Don’t change the goalposts to this being what you want it to be. We are considering specific things here, with a consistency. That’s it. Arguing about it doesn’t make it different.
You can’t have “the same framing” and have a change in perspective. You’re changing too many things to make any comparison, or come to any specific USEFUL observation.
Bam! And here we are. This is what we are considering. I don’t understand why you are still arguing this point when you agree with me on this.
Again, as I mentioned to you in an earlier reply – check the Addendum to this article. I directly address your observation there that the focal length doesn’t change, but HOW people use the camera and lens, does change when shooting comparable subjects. That is where the crucial difference comes in.
At this point I am repeating myself, and you are changing the goalposts. There’s not much more to be discussed here.
Rudy … it’s not so much the lens. Of course, the specific optical characteristics will affect the look of the image and the bokeh … but DoF isn’t a function of lens design (as far as I can figure out.)
DoF is defined in a specific way, in relation to the Circles of Confusion, when viewing a print of a certain size, at a specific distance. (The exact details are mentioned above in the linked article.)
This is where most people get unhinged around the idea that keeping the framing consistent, while using “equivalent focal lengths” on different size sensors, will give a different DoF.
Just to reinforce this, check your DoF Calculator on your smart phone.
I know it is counter-intuitive, but that is how it works out.
A similar question comes up every so often: for example, comment #63,
but the answer remains the same: my comment #64.
I’ve added this now as a practical exercise (via A Little Bit of Homework) in the main article.
Shoot the same portrait of someone (who doesn’t move),
using a FF sensor camera, and then a crop-sensor camera,
while keeping to the same framing of your subject,
and also keeping the same perspective. (i.e., You can’t move.)
Now compare the DoF between the two images, for the same resolution.
Let us know what you find.
I think in terms of DOF, you have to go more than one F stop to get the same DOF. Using your method (DOF calculator) Two shots having same Field of View:
Focus Distance 15ft 75mm F/2.8 Total DOF 2.04 feet Full Frame
Focus Distance 15ft 50mm F/2.8 Total DOF 4.71 feet 1.5 Crop Equivalent
Focus Distance 15ft 50mm F/2.0 Total DOF 3.29 feet 1.5 Crop Equivalent One Stop
Focus Distance 15ft 50mm F/1.4 Total DOF 2.31 feet 1.5 Crop Equivalent Two Stops
Focus Distance 15ft 50mm F/1.2 Total DOF 2.06 feet 1.5 Crop Equivalent Two and 1/3 Stop
This is a 2 and 1/3 stop difference. That’s huge if your trying to blur your background! Different calculators may give slightly different distances, but the 2 1/3 stops remains. If your trying to isolate your subject, Full Frame makes it a lot easier.
Thank so much Neil.
1. When I go to DXO mark to see the lens testing I see the lens always performs better in full frame camera than in crop sensor camera. I want best quality picture so can crop sensor give me same quality picture than full frame? Let say I want a picture with 85mm f.2.2 on full frame. To get similar DOF and picture frame, do we have same quality picture with 58mm f/1.4 on crop sensor?
2. About picture quality (sharpness) can crop sensor camera (with $1000 lens) reach the top sharpness of full frame camera (with $1000 lens)?
Thanks
Different lenses will have different quality. So a direct comparison in terms of image quality, while skipping from FF to crop, using different lenses, would be difficult.
Can you send a link to where the DXO Mark website shows that the same lens, performs better on FF than on a crop camera? I would like to see what parameters they are considering.
I know some have complained about using the “equivalent” focal length versus physical focal length stuff, but I agree with you. Practically speaking it makes better sense to me.
However, at the bottom it’s not equivalent focal length; it’s equivalent angle of view. A 35mm lens on a Nikon crop frame camera has a very close angle of view (44 degrees diagonal) to a 50mm on a full frame camera (46 degrees diagonal). The only problem is that I have years of experience knowing what I’ll see in the viewfinder of a FF camera with a 50 mm lens, and no memory what a “46 degree” angle of view looks like.
Too bad that what you write won’t stop the confused posts in various forums, but keep fighting the good fight. Maybe it’ll slow down a bit.
I’m late to the party but with that said, I do appreciate the write up. Generally, I don’t care crop vs FF as the term FF could mean 8×10 or 5×7 or traditional 35mm, I shoot both a Nikon FF but more so now with my alleged cropped Fuji X-T1. The article is interesting and a good read. Was thinking that some of the DOF difference you present could be due in part to the 58mm lens producing small circles of confusions compared to the 85mm and those smaller CofCs projected onto smaller pixel pitch of the D810 vs the larger CofCs projecting onto larger pixel size of the D4 (4.87UM D810 vs 7.28UM D4). DOF difference is undeniably there in your pictures but was thinking how much pixel pitch and differences in CofC could affect the final outcome. It’s a fun mental exercise but I love the outcome of my Fuji especially paired with the Fujinon 23mm 1.4 to stop shooting with it. Again, thanks.
There is some confusion here.
To keep the same framing (percentage of object in image), zooming with the feet does not change the perspective (angle of view) , it is zooming by focal length that changes the perspective. The angle of view is fixed by the sensor/focal length combination.
Unless of course you have a different meaning of perspective.
tan(aov) = o/d where d = distance from object
To maintain the same framing -“o”, you either change “d” or “aov”.
Zooming with feet changes “d” while zooming in/out changes the “aov”.
Lou … you appear to understand Perspective differently than the rest of us.
If you MOVE, you change the perspective … and with ‘perspective’, I / we / everyone means the positioning and balance between objects in the foreground and background.
If you just zoom, and stay static, you are essentially just “enlarging” the scene … no change to the perspective. (The opposite of what you’re saying.) However, if you move … the perspective does change.
Also, you absolutely can not “zoom with your feet”. it is contradictory.
Where you complicates thing is by bringing in “percentage of object in the frame” … and relating that to zooming.
If you shoot at 35mm and 85mm, and move your POSITION to keep the same percentage of the object in the frame, your PERSPECTIVE most definitely did change.
I totally agree with your arguments. Bokeh is an important thing in photography esp when it revolves around portraiture but sadly a huge amount of photographers out there worry more about blurring the background rather than creating an shot with impact. After shooting for quite some time (i dont regard myself as a pro photog) i would not buy an expensive camera and lens just for the sake of having a blurry background. If i was a newbie though, i would have been tempted to opt for a full frame than a crop with the comparison!
Good post, hence why I picked the 85mm 1.4 to match my favorite real lens for portraits, the 135 f2. They are more or less equivalent. I don’t care so much about dof on wide end, so just use a 24 1.8. That way with 2 bodies I get 24/35/85/135… my favorite focal lengths in just 2 primes.
So, if I like to shoot birds and need “reach” and currently use a 7D (approx. 19 MP), I would need at least 38 MP in a full frame to achieve the same reach/resolution — by cropping the full frame image to match the cropped sensor. Anything less than 38 MP would reduce my “reach”. Is this a correct interpretation?
Essentially yes, for the same resolution.
But let me get this straight. F stop ONLY increases IF you use a full frame lens on a crop frame body right? Not if you use a crop frame lens on crop frame body. Right? Because a 50mm AFS DX f1.8 will remain f1.8 on a DX body, but only the field of view would be equivalent to that of a 75mm(obviously FX lens) lens put on a FX camera. Right?
I can’t really answer you, because you are confusing some terminology I think.
But to answer your question as it is stated – no, the F-Stop doesn’t increase. The aperture remains the same value.
I ran across this article on DoF and crop vs. FF equivalences, while looking for an answer to a related question. Even though this article didn’t relate directly to my question I read through it just in case, and found it lined up with how I had always understood these issues. As you state and as we all understand, the resolution drops off the farther we get from the plane of the focus. There’s no “falling off a cliff” where the image goes radically from sharp to unsharp. It’s a gradual thing.
So the question I was researching is this: Is there an industry accepted optical formula for deciding on how to measure out the DoF scale on a lens, or is this a judgment call made by the lens maker? After all, the decision on where to draw the line is an arbitrary one, based on what is considered sharp vs. unsharp.
I ask the question as I have two radically different DoF scales, on two different lenses, for the same focal length and aperture. My FF Voigtlander 21mm f4 has a much broader DoF scale than my crop sensor 18-55mm Fujifilm zoom when set to 21mm and the same aperture. The X series has the option of displaying a distance scale with a DoF slider on screen or in the viewfinder. At f11 and 21mm, the DoF shown on the viewfinder scale on the crop sensor lens is much narrower than the scale marked on the FF lens. Obviously this should be the other way around.
Any insights on how the lens makers decide on these scales?
Hi there Peter
They are most likely different because of the other thing you mention – there are no absolute yes/no points where the image is soft / sharp. It’s a gradient … and this is open for interpretation.
There is another aspect here …. traditionally, DoF is defined by what appears sharp enough when printed to 8×10 and viewed from a specific distance.
Now with digital, we scrutinize the images differently, and I think our perception of what is sharp and what is ‘sharp enough’ has shifted to where we are more critical of sharpness.
I know that even if I would print to the same size, that the same photo shot with a 12 megapixel camera, or with a 36 megapixel camera, will appear the same in terms of DoF when viewed on the print. But should you look closer to the actual digital file of the 36 megapixel camera, you’ll notice that the apparent DoF is much smaller. So I tend to shoot even more conservatively now with the Nikon D810 – being careful that I have enough sharpness, despite what the classic way of thinking about DoF would tell me.
Thanks, adjusting our DoF scale depending on resolving power of the sensor makes perfect sense. The DoF scale will always be an arbitrary judgment call on what is sharp vs. not sharp, so we’re free to make our own judgments as to what works best for our photography.
Reading through the comments here, we should all keep in mind that having differing DoFs over different formats for the same focal length is an old issue. In the film days we had 35mm, medium format and large format negatives. A 210mm Schneider Super-Symmar f5.6 lens in front of a 4X5 negative produced a vastly different perspective and DoF compared to a 35mm Vivitar Series 1 70-210mm f3.5 zoom used at the 210mm end. The 210mm Schneider was closer to a 60mm lens in 35mm terms.
Those of us who used multiple film formats, knew this instinctively.
In the second image using the Nikon 58mm f/1.4 lens ( 87mm equivalent) on the Nikon D810 why does it appear that the model’s face/ image is sharper and brighter.
I think that mild difference is due to the cameras having different sensors, and the RAW software interpreting the RAW files slightly differently.
Hi Neil! Great article on theese matters.
Now, if you might think it viable to update this topic on some current things that are happening on the market — how does the FF camera behaves comparing it to an APSC mirorless (if it makes any diff.) camera _with_ speedbooster and probably the same lens?
Good article.
I just did some experiments with a 70-200 f2.8 on APS and FF canon.
I shot from the same position, and set the lens to 200mm on FF, and to 125mm on the APS (200mm equivalent field of view).
At f2.8 the FF obviously looked better than the APS at 2.8, but what surprised me was that even f4.0 on the full frame looked superior to f2.8 on the crop!
I.e. a cheap f4.0 lens on a cheap full frame will give you nicer photos than a 70-200 f/2.8 II on APS body!
In theory this all makes sense with how a full frame lens will behave on a crop sensor camera.
Where I get confused is how Nikon calls their 18-70 mm dx lens the equivalent to a 28-105 mm full frame lens. Which cannot be the same. I have a telephoto lens that is full frame that starts at 70 mm and it does not look the same as a telephoto end of their mid-range Zoom 18 – 70 mm. In fact I have to back up in order to get the equivalency on the full frame lens to look the same. Can anyone explain why they would call that mid-range Zoom lens a 70 mm when it isn’t.
Dan … the lens manufacturer says there is an equivalency between the 70mm (on the crop lens), which effectively makes it behave like a 105mm lens.
So if you shot with the full-frame camera and the 70-20mm lens at the 70mm setting, then the composition and perspective would look different. In theory you’d have to zoom to 105mm setting on your 70-200mm lens … if you wanted the same composition in the final image, comparing the DX and FX cameras and lenses.
Hi Neil
Can I add another consideration to the mix? Particularly relating to architectural photography or interiors. I find when I use a wide lens, like my 14-24 Nikon lens and shoot at 14 to 18mm I get a bucket load of distortion on the edges of the frame, so if there was an object on the corner it looks very bent into an odd shape. So I appreciate the 5×4 cameras are maybe better for interiors due to the camera movements, and so I got the 24 PC-E lens from nikon. I find if I stitch 5 frames together it must be equivalent to around 15mm on the zoom lens, but the overall perspective seems to me to much more natural. So by doing this I am effectively increasing the combined size of the sensor plane to something larger than a full frame Phase One IQ100. Do you think there could be a difference in the ratios between object perspective, focal length and sensor size, so the longer the focal length and larger the sensor size, the less squishing of the edges of an image?
Lets take one lens, at different distances DOF differs.
Lets take a given frame, the so called crop factor implies that the shooting distance must be multiplied by the crop factor (geometry) in order to cover the same frame.
In order to obtain the same DOF you must to reduce the aperture by the crop factor (because you are further away).
If i am not mistaken the model is Glass Olive. :) :)
Yup, that’s her.
It’s very good analysis. However in terms of Bokeh Effect, I think there’s something missing here. Your analysis will come to the conclusion that a Fuji 56mm 1.2 will have similar Bokeh effect as a 85mm 1.8 on a full frame. I think the 56mm 1.2 only has similar bokeh with a 85mm 1.8 on a crop sensor (considering this, a Fuji 90mm f2 have similar bokeh effect with the 56mm 1.2). The 85mm 1.8 on a full frame will have “one more stop” of bokeh, so my conclusion is that a 56mm 1.2 on a Fuji is similar to a 85mm 2.8 on a full frame.
First of all, I need to accentuate again that “bokeh” and “depth of field” aren’t interchangeable.
When you say:
— “I think the 56mm 1.2 only has similar bokeh with a 85mm 1.8 on a crop sensor”
What makes you think that?
— “The 85mm 1.8 on a full frame will have “one more stop” of bokeh”
What do you mean by this?
— “so my conclusion is that a 56mm 1.2 on a Fuji is similar to a 85mm 2.8 on a full frame.”
I don’t understand this jump in logic. How do you get to this from the rest of what you mention?
Excellent article. Thank you! That the DOF increases with cropped sensor camera everything being equal is new info to me that I just acquired from a bird photography book. And this article is more followup on that subject. However, this one discussed the effective focal length subject once again and I completely agree. 100mm is 100mm – its physics is a lame way of seeing it IMO. Since they bring different effects and the photographers need to adjust them and be prepared for this effect it is very effective to use different term to capture this. So ‘effective focal length’ definitely is a better way of saying it.
I would like to add that sensor resolution should be another parameter that needs to be considered when we talk about this effective focal length. If the FX and DX are not the same resolution (# of pixels) the frame may look the same but resolution won’t match and that may turn into non-equivalence when printed or projected to screen.
Out of curiosity, I did some quick tests today with a Tamron 100-400mm lens on a Nikon D610 and Nikon D90. It appears based on my tests the background blur effect has more to do with the focal length and less to do with sensor size (everything else being equal).
When I shot at 400mm the background blur was very comparable for both DX and FX frames. Of course the DX frame was more cropped (so contained more of the subject).
But when I used equivalent focal length to get about the same frame then the DX frames blurriness was clearly less. I used 400mm for the FX and 270mm for the DX.
My conclusion from this – the higher the focal length the higher the blurriness. Sensor size is not a parameter for blurriness.
Did you change your position to reframe the subject to the same size in the viewfinder?
Or did you just crop the image? And when you cropped the image, did you make an attempt to look at the cropped / uncropped versions side-by-side (not at 100% )?
You comment is missing some important factors to be able to come to the conclusion you did.
I provided two test scenarios. In both scenarios I shot the subjects (in my front yard, stationary leaves or objects at different distance. Each one of them has a background of a forest farther from them) from the same position.
In the first scenario, I shot all frames with 400mm. These are not equivalent frames. DX ones are more zoomed (cropped). I took the FX frames and cropped them to look equivalent to the DX ones. Now I compare them side by side and the background blurriness look very similar. All parameters were the same except the sensor size (of course camera bodies). Of course my crop was not mathematically done. I did it visually. The width/height were +/- 10% difference. I doubt that would make a big difference to blurriness.
The 2nd scenario is pretty clear. Matches your observation. “Yes, a 100mm lens on a crop-sensor camera will give you the same perspective as a 150mm / 160mm lens (on a full-frame camera), if you don’t change position … however, the DoF increases.”
Yes – these were also the same subjects and shot from the same position. No cropping. The frames were equivalent because they were shot with equivalent focal length. Same aperture, shutter and ISO.
I would like to clarify, my conclusion was done from some quick tests and it is not meant to misguide anybody. Even for myself – it is not a final conclusion but a conclusion that helped me make a decision to use my new zoom lens on my DX body for a trip next week.
Another very important aspect is you can’t change focal length without changing aperture , they go hand and hand. I’ll explain; a 50mm lens on FX acts like a 75mm on DX 50 x 1.5 = 75. Now we have to do the same with the aperture (f-stop) , so f /2.8 on a FX camera becomes f /4.2 on a Dx camera. 2.8 x 1.5 = 4.2 . This will explain why the DX camera will always have greater DOF and never give that soft background. And, as a side note you also have to factor the ISO as squaring it by a factor of two. 1.5 squared is 2.25 therefore ISO 100 x 2.25 on FX = ISO 250 / ISO 250 on FX to reverse it Divide : ISO 250 on FX 250/2.25 becomes ISO 100 on a DX. This explains why people who try to mimic the settings of an FX to a DX get images that are expensive in noise.
Is there a difference in aperture between FF and a crop sensor? This is not in reference to DoF but to the amount of light passing through the lens. Does f/2.8 a function of the reciprocal of the diameter, or is it a function of the amount of light hitting the sensor: is f/2.8 on FF a larger diameter (needs more light) than a 2.8 on a cropped sensor . . . . . or am I missing something?
I think this is what Anthony in Post 98 is trying to explain but I don’t quite follow him.
Hi there Lawrence — in terms of exposure, there is no difference in aperture between a FF sensor and a crop sensor.