Most photographers believe that larger sensor sizes produce images with shallower depth of field, but that's not exactly true.
Before we can fully explain depth of field, let's talk about how a lens works. Light rays reflect off of an object, and a lens can focus those light rays onto a digital sensor. Focusing a lens will allow a single-point source of light at a precise distance to be focused as a single point on the sensor. Everything else in your scene that is closer or further away from your focus distance will create blur circles on the sensor rather than sharp points because those light rays converge before or after the camera sensor rather than directly on it. These blurry circles are called the circle of confusion and the circle of confusion "limit" is the largest the circle can be while still being perceived as a single point by a human viewer. The further away these rays converge from the sensor, the larger blurry areas of light or "bokeh" will be produced.
What Exactly Is Depth of Field?
Technically speaking, depth of field is determined by what is acceptably sharp by a human viewer, which means that things like resolution, image size, and viewing distance can change the depth of field.
To understand this, imagine that you have a 100-megapixel image file. If you had a 4 x 6 print of this image and you were viewing it at arm's length, you would have a hard time determining exactly what was in focus and what wasn't because the human eye would only be able to perceive about 2 MP worth of detail at this size from this distance. Now imagine you printed the same image the size of a movie screen and could get as close as you wanted to it. From this perspective, you would easily be able to determine what was in focus and what wasn't, which would technically make the depth of field shallower. Camera manufacturers have come up with a standard that assumes that you are going to print the image at an 8x10 inches and view it at 25cm. With these parameters, the circle of confusion limit is .029mm on a 35mm sensor. Anything larger than that will appear blurry.
Remember that a lens is only able to focus on something at one precise distance at a time. Anything closer or farther away from this exact point isn't technically in focus, it just may appear to be in focus to a human based on how much detail they can perceive. If you had a photograph with unlimited resolution and clarity and you could infinitely zoom in without losing any detail, the depth of field would become shallower as you zoom in because you would easily be able to see what was sharp and what wasn't.
Smaller Sensors Usually Produce Shallower Depth of Field
Most photographers assume that smaller sensors will produce a deeper depth of field but technically speaking, smaller sensor cameras usually produce a shallower depth of field because they tend to have higher pixel density/smaller pixels on the sensor. The circles of confusion being projected by the same lens will be the same physical size on both a 35mm and micro four-thirds sensor but when you blow up both images for print, the smaller sensor image will need to be enlarged more to produce the same size print because it came from a smaller source. When you blow it up more than the 35mm image, you are also blowing up everything including the circles of confusion and a human would now be able to more easily see what is in and out of focus.
Remeber how a circle of confusion needed to be less than .029mm to appear "in focus" on a 35mm sensor? On a micro four-thirds sensor a circle of confusion must be smaller than .015mm.
Imagine if you had a full frame 35mm sensor that was 20 MP and a micro four-thirds sensor (which is 1/4 the size) which also had 20 MP. If you attached both cameras to the same 35mm lens, the full frame sensor would capture the entire scene projected by the lens but the micro four-thirds sensor would capture only the center of the scene. Both images have the same resolution but the image taken with the smaller sensor would be zoomed or cropped in and it would give the viewer an even closer look at all of the details. This would allow the viewer to notice details like precise focus more easily meaning that the smaller sensor actually produced shallower depth of field. Check this out for yourself on any depth of field calculator.
In the video above I didn't get too deeply into this because it can get confusing and this phenomenon is very difficult to see unless you have cameras with wildly different sensor sizes and resolutions. The more important bit of information is what exactly is causing changes in depth of field.
The Only Three Things That Affect Depth of Field
1. Changing the focus distance
The only way to change your focus distance is to move your subject or move your camera. As you move your camera further away from your subject, your focusing distance will increase and your depth of field will increase. This occurs because the light rays that are bouncing off of your subject and entering your lens are converging more slowly the further you move the camera away.
2. Changing your focal length
Your lens' focal length is the physical measurement of the distance between where the light rays converge to your camera's sensor. As the lens moves further away from the sensor, the light rays will converge more slowly onto the sensor, which means that light rays will have a tendency to focus further in front of and behind the sensor which creates larger circles of confusion (bokeh) and a shallower depth of field.
3. Changing the lens' aperture
The final way that we can change our DOF is with the lens' aperture. By stopping down the aperture, you are physically blocking the light rays that are coming from the edges of the lens that would produce the most blurry circles of light on the sensor. Closing down the aperture will create a darker overall image, but will also increase the depth of field.
If you'd like an illustrated example of how each of these changes affects depth of field, this video does a great job of explaining it.
Conclusion
The sensor size itself does not produce shallower depth of field, but bigger sensors will force photographers to move closer to their subjects or to use longer lenses to produce similar fields of view of a smaller-sensor camera. Moving forward and increasing your focal length will both decrease depth of field.
If you enjoyed this, you may also enjoy my recent video/post debunking lens compression.
I guess generally when people talk about this, they are talking about equivalency, and it's much easier to say that a larger sensor gives a shallower depth of field than all of this.
For equivalency, a 50mm 1.8 FF lens at X distance gives a shallower depth of field than a 25mm 1.8 MFT lens at X distance.
That's just much easier to convey.
And is exactly what you say in your conclusion. Still, this is a valuable article and video for an understanding of how focal length, distance, and aperture work. And the lens compression thing is a one of the most propagated myths out there - that was a great article post of yours. Wish more people understood that lens compression is not a thing.
I will be disagree on the title of this article and agree with you
Yes, The sensor size is impacting depth of field on similar aperture lens, period
It's funny to see how all brand talk about 35mm equivalent only on the focal range, but never on the aperture!
My f2.8 Sony RX100 Camera will never have the same depth of field than any L lenses F2.8 on my Canon 5D, but still, both are described at f2.8!
Yes, it's cheaper to get a large sensor camera and an average fast lens than a crop sensor camera with super fast lens to produce the same effect.
APS-C or more cropped camera sensor are smaller in size, but can't produce similar look without heavy lenses. This is just physics.
You always need fast lens with large aperture as sharpness goes when stopped down. So a 1.4 lens will be very good at f4.0. A F4.0 native lens will gives similar sharpness at F8 (in the same price range) . So, when people says they don't use F1.8 or less, so it do not matters... Well, think twice.
You wrote "Yes, it's cheaper to get a large sensor camera and an average fast lens than a crop sensor camera with super fast lens to produce the same effect."
Interesting point, but... how many FX owners standing in front of the lens shelf in the camera store say "Oh, I won't bother buying this nice fast lens because this slower lens can already do what a fast DX lens can do."?
(Also, I'm a DX owner, and I do not feel the need to try and achieve FX-style DoF in my photos.)
I don't really agree about talking about aperture equivalency, unless you are specifically talking about DOF. But arguably more importantly, in regards to aperture, is exposure.
Your RX100 at f/2.8 will get the same exposure as your Canon 5D at f/2.8.
So when you say "both are described at f2.8" - what else should they be described as? Not only is aperture a mathematical ratio that does not have anything to do with sensor size, but is critical for exposure. If you said, well this is a 2.8 lens but I'm on M4/3 so it's really an f5.6.... good luck getting proper exposure.
The equivalency there is only in regards to DOF, and that - as this article points out - is only due to the shorter focal length with smaller sensors when trying to achieve the same field-of-view.
Yes, I'm talking about DOF specifically as this article is about.
F2.8 on a smaller sensor create a DOF different than a F2.8 on large sensor.
PERIOD. (even if, technically, aperture is the same)
But where a lot of advertising on lenses are displaying "equivalent 35mm" about focal length, this DOP never to be found their equivalent 35mm" on DOP.
Why ?
Marketing BS
We should be comparing apples-to-apples. Here's an experiment that you can try. Set up a DoF test scene. Put a Nikon D7000 and a FX (Full-Frame) lens (your choice of lens focal length) on a tripod. Focus as appropriate and take an image. Now take a Nikon D800E and use the same lens and take a picture the same scene. Note: Do not move the tripod and do not refocus and do not change the f-stop. Yes, with the D800E your test scene will fill only 45% of the frame. In the photo editor of your choice, carefully crop the image to be exactly that taken with the D7000. Print the D7000 image and the cropped D800E image as an 8x10 print. Compare the DoF between the two images.
What does this experiment do? It holds the f-number (aperture), focal length, Subject distance, and magnification the same. The purpose of cropping the FX image is to hold the magnification the same. The magnification being the same means the Circle of Confusion (CoC) will be the same for each of the two images. I chose the D7000 and the D800E because their pixel areas are within about 5% of each other. The CoC won't quite be the same between the two, so you might see a slight DoF difference between the two.
Finally, go back and take another image using the D7000 and this time use a DX lens of the same focal length as original DX lens (do not use the "35mm equivalent" lens). Print this image at 8x10 and compare to the D7000 with the FX lens. You should find that the DoF is the same.
What does all this mean? It means that sensor size doesn't have any effect on DoF, which is the subject of the original article.
"35mm equivalent" means that it has the equivalent Field of View (FoV), not that it is "optically the same as". A 50mm lens on my D7200 is going to have the FoV of a 75mm lens on a FX camera, but it's going to deliver all of the optical properties of a 50mm lens, including DoF. When we compare a lens based on a "35mm equivalent", we're comparing apples-to-oranges.
The question is also how blown out you want the background to be.
I don't want FF frame for several reasons:
1: I am an amateur, I don't need it.
2: Weight is an issue because of artrosis. I need to carry less, not more.
3: Full frame and especially the lenses are too expensive for my income (but spending money hurts once, dragging weight hurts every single day).
4: Aps-c is good enough for my needs.
5: I can blow out the background enough for my needs.
This picture was taken at f4 with an 1,8 lens.
What about blowing out the background? You mean making the background blurry?
Although he may not be entirely clear in his statements, I don't understand why people are voting him down, when he's obviously happy with what he's got? Perhaps they're not happy with what they may have. ;-)
I can't find "arthrosis". Is there another name for it or is that the name in a particular language?
It's probably translated from Dutch (Artrose).
https://www.healthhype.com/arthrosis-define-causes-symptoms-treatment-di...
Thanks. I'd never heard of that before but then, there're lots of things I've never heard of before.
Be glad that you didn't discover you have it. It hurts and I have more and more trouble walking longer distances.
This is probably the most straight forward, easy to understand explanation of this I've read. It took me a long time to understand this stuff and I had to hobble my understanding together and make the connections myself from reading multiple sources over the years (I'm a slow learner). This is def getting bookmarked for sending to friends.
Thanks! I'm ashamed to admit how long it took me to research this and explain it quickly and simply.
Nah man, I get it. My day job is as a graphic designer for medical courses. Sixty percent of the job our team does is to simply try and figure out how to explain (for me, visually) concepts that are often either complex or abstract. It is a challenge, and always impressive when effective.
As an aside, if you want to, go look for my responses to your video on Lens Compression from last week. I had never really considered the validity of that term before that video (which was incredibly enlightening to me), but once I did, the contrarian in me felt like there was something missing. I have three posted responses, and by the time I get to the end of the last one, I think I finally hit on what bothered me about your premise. Tell me what you think (as if you don't have anything else to do). Either way, I love and respect this site a lot. So take anything I type with that as a context. =)
I agree with you. The title is definitely clickbait ;)
Ha, fair enough my man. =)
In regard to the conclusion: Stopping down to an equivalent aperture (i.e. same iris size) on an equivalent FoV, will give you the same DoF.
The last factor that affects depth of field is the degree of enlargement to the final display size. You did describe the effect, but you forgot to list it.
Also, the "circle of confusion" is defined as the largest blur circle that will be accepted by the average viewer as a sharp point at a given degree of magnification and a given viewing distance.
Canon, for instance, calculates their "circle of confusion" for the 24x36mm frame with the givens being a 5x7-inch print viewed at 10 inches (according to their book Lens Work III).
^^This. Part of the equation that is often omitted from any discussion about depth of field. Admittedly this is much more important when determining the amount that will appear in sharp focus, not ability of a lens to separate a subject from the background.
Depth of field is an illusion... the only part of the image that is truly in focus is the focal plane. Anything else that “appears” in sharp focus due to all the factors discussed is the illusion.
The best part is when you take the same exact photo (with the GH5) and crop the D850 shot, this is the easiest way to explain it to people. Im not sure why this is a tough concept for people to understand, they call it a crop sensor for that very reason, treat it like a full frame (24x36mm) image plane and crop it. Thats it, not magic.
Just a slight niggle with the use of the term 'bokeh'. It has little to do with the sizes of Circles of Confusion, which is more of a measurable threshold than anything else. Bokeh is a made-up word (dubiously adopted from Japanese), first popularised by Mike Johnson in the mid 90s to describe the 'quality' of OOF detail. Different lens designs render OOF details in different ways, some good, some bad. Bokeh is a useful term to describe this, but has nothing to do directly with the amount of DoF.
Fair point. I've elaborated in the post
This is very helpful. Does this mean that if I take a shot with, say, my nifty fifty on my 5Ds but crop it to 20 MP size it will have the *exact same* DoF as shooting it on my 20 MP 7D2, assuming distance and aperture are consistent? (I use these two camera examples because the pixel pitch is identical and also because I have both and I've always wondered this.)
Essentially yes dof will be pretty much identical if not the same.
Paul Scharff---yes, but pixels don't have anything to do with depth of field. The same physics applied with film, and they even apply to a groundglass image.
Under changing your focal length, you wrote:
“When you zoom a 70-200mm lens, you are physically moving the center element from 70mm away from the sensor to 200mm.”
I don’t think that’s correct. If you take let’s say a Canon 70-200 F/4L (physical length of the lens is 172mm) the center element won’t be 200mm away from the sensor even if you add the 44mm flange distance. With longer focal lengths this is even more apparent. An EF 800mm F/5.6L is only 461mm long.
What we call focal length these days is a translation from the angle of view of the lens. A single lens element with a focal length of 200mm has about a 12° diagonal view angle on a full frame sensor. A lens system which produces the same diagonal view angle is what we call a 200mm lens.
Ok this has led me down a long path but I think I finally figured it out. I have changed to the article to say that focal length is the calculation of an optical distance from the point where light rays converge to form a sharp image of an object to the digital sensor. It's not necessarily the center of a lens (though it might be).
But how can a lens be physically shorter than it's focal length? Wikipedia says "A telephoto lens works by having the outermost (i.e. light gathering) element of a much shorter focal length than the equivalent long-focus lens and then incorporating a second set of elements close to the film or sensor plane that extend the cone of light so that it appears to have come from a lens of much greater focal length. The basic construction of a telephoto lens consists of front lens elements that, as a group, have a positive focus. The focal length of this group is shorter than the effective focal length of the lens. The converging rays from this group are intercepted by the rear lens group, sometimes called the "telephoto group," which has a negative focus."
Read more here: https://en.wikipedia.org/wiki/Telephoto_lens
Yes. Originally, the word "telephoto" defined a lens that was designed to be physically shorter than its focal length length. A lens that was physically the same length as its optical length was a "long focus" lens.
By the same token, a wider-angle lens that was physically longer than its focal length was called "retrofocus."
https://m.dpreview.com/articles/2666934640/what-is-equivalence-and-why-s...
The question is what else you keep equal. For the parameters that photographers care about (subject distance, angle of view, f number) then using a bigger sensor does lead to shallower depth of field, as the dpreview link posted by Trevor demonstrates empirically. Sure, if you look at the physical size of the aperture in mm that's a different story, but all photographers I know reason in terms of f number, not physical aperture diameter.
Also see this Stanford applet to explore the geometry of depth of field in the case of thin-lens approximation. https://graphics.stanford.edu/courses/cs178-10/applets/dof.html
Well, yes and no.
To say that larger sensors produce shallower depth of field is pretty much on the same accuracy level as saying that higher ISO is producing a more exposed image. If shutter speed and aperture remain the same then increasing ISO will definitely result with more exposed image. In the same fashion if focal length and distance from the photographed subject remain the same, use of bigger sensor will result in shallower dept of field. But if we start stipulating that the other parameters can be adjusted then none of those statements may be true.
So indeed larger sensor itself will not do the magic and author explains very well why but this does not change the fact that given availability of the lenses and other practical aspects, a portrait photographer will grab a FF DSLR over M43 mirrorless.
I can mostly follow this, a little over my head... But asking myself while reading - WHY'S would this matter to me as an amateur? Then watching the video illustrate the points so well, I actually realized in the side by side examples... there are some images that to me were way more preferable - brighter, sharper, more interesting. Now to watch again and figure out which settings or considerations (distance, lens) got those results.
I’ve never seen Fstoppers go so low.
This article should be title « how to compare apples and oranges to get a good clickbait ». This is excessively disappointing from Fstoppers.
For the same perspective and FOV, at identical f-stop and print size, a larger sensor will produce shallower DOF. That’s what matters to photographers.
So if with smaller sensors you have to move farther away to keep the same field of view which incresses the distance to the focus plane which increases the dof, then isn't the sensor size what triggered everything after all? You had to move because of the sensor size... :)
If your shooting with a medium format camera it will have a longer focal length lens as its standed lens ie Pentax 67 105mm 2.4 were a 35mm digital or film camera has a standard lens of 50mm.
So if yow were to shoot the same shot with each camera with there standard lens at the same distance and f stop. The Pentax would have a shallower depth of field.
This is due to having a longer focal length lens as it’s standers lens. So by using a larger format with there respective standard lenses you would naturally get a shallow depth of field.
So I am saying a larger sensor has a shallower depth of field or am I missing something
No, you're 100% right. Shoot 8x10 and the DOF is still super shallow at F16.
If you would like to have wider angle with shallow DOF you need to get bigger sensor.
You have to change your focus distance because of sensor size.
I don't really have to know why since it does not matter really. The conclusion is still the same.
Next topic idea :D
-perspective distortion and sensor size
I just use a full frame DSLR - other sensor sizes dont affect me as I really don’t understand this stuff but thank you for giving me increased understanding of this.
Rick McEvoy
http://rickmcevoyphotography.com
Thank you for helping me understand this issue much better than I did before. Being a simple chap I just use full-frame so dont need to worry!
Regards
Rick McEvoy - http://rickmcevoyphotography.com/
Just No!
Allow a Physicist specialized in optics clarify a few points.
1st Your definition of DoF is used as a measure of the zone of Pixel level (100%) sharpness. So when you want to shoot a Landscape and you want the Flower in the foreground and the Mountain in the background to be pixel perfekt sharp then tis measure applies.
However this is not the DOF perceived by viewer. How come? take a 100MP image and downscale it to 2MP the DoF perceived does not change. Why because we perceive the DoF over the image scale not the pixel size. So you can just ignore different resolutions.
2nd. while Your explanation of how DoF changes with Distance F-Stop and Focal length changes is right. Your mistake is to assume that they don't change with Sensor size. What does this mean?
When you shoot a Portrait with a MFT and a Phase one you will not shoot both with the same Aperture and focal length at different distances the reason for that you explained your self in the Lens compression Video.
You are much more likely to shoot at the same distance while you could still shoot with the same sense and just crop the Phase One to MFT that would defeat the purpose of shooting with the Phase one in the first place. So you are going to use different lenses. You then have to apply crop factors to the focal lengths to keep your framing. (e.g. a 42.5mm and 132mm as 85mm äq.) Now different rules apply to get the same DOF namely multiply crop with F-Stop and you are fine. But when the 130mm is f2 you need a 42mmf0.64 to get the same image! good luck finding that lens.
So does Sensor size change DoF well not directly but it very much forces you to do so!
Best regards
So, you essentially agree with the article then lol.
It's not the sensor directly causing the change but other factors that are applied.
Just Yes!?
Thats just like saying more passenger seats don't make a plane bigger. Its the designer who has to make room for the seats. Utterly useless semantics.
That's correct though, more passenger seats do not make the plane bigger there is an upper limit. Also, how exactly does your analogy fit there? A very strange analogy to use.
Sensor size does not impact DOF in any way (COC aside) if you're discussing the actual physics of how things work. What you're doing is mixing up two discussions, the physics of something and the creative decisions applied, They are different points, however, they work together in many situations.
Understanding what causes something properly is a much more effective way to learn. Saying larger sensors produce shallower DOF is not only incorrect but also a half-measure and a very lazy way of explaining things.
Well If you want do dive in deep the you would have to understand a whole lot about lens design to understand why it is so hard to make a decent lens with an f-stop < f1.0. And you would have to do the whole calculation to show that with a recliner lens you get the same angle of view when you multiply focal length with crop factor. But that is the easy part. Then you need to do the calculation to show that:
1st. In the case of fixed distance fixed Focal length you get the same DoF (expressed in Image size) with different Sensor sizes if their f-stops * crop are equal.
2nd. in the case of fixed reproduction scale (e.g. face fills image) and varying focal length and sensor size, DoF (expressed in Image size) only depends on f-stop * crop.
But this isn't easy even for people with a good mathematical understanding of optics.
So sometimes we have to live with half knowledge.
But some half knowledge is dangerous because it leads to wrong conclusions.
Telling people that sensor size don't affect DoF is in general just wrong.
Why? you can't get a MFT 42mmf0.7 however you can easily get a FF 85mmf1.4
So a different Sensor size enforces different engineering constraints that result in lower DoF...
This is the same with the airplane seats analogy. the number of desired seats constrain the size of the Airplane but they don't change it per se.
E.g. an Lockheed C-5 Galaxy is a huge plane with few seats. (a Phase one with a f64 lens in the analogy.)
So it is best to stick in a simple half knowlage that leads to little wrong conclusion.
Therefore multiply f-stop and focal length with the crop and the ISO times Crop^2 and you won't go wrong.
The only case where you could get issues with this frame of mind is when you use a Speed booster. but than you just replace crop with (camera crop)*(speed booster Factor) and it works again.
if you stick with Lee Morris explanation you get a lot of confusion just read the Debate on what DoF actually is. While everyone intuitively knows that the DoF of a 85mmf1.4 looks different that the DoF of a 15mmf4 no matter the Megapixels. We have Lee tell us that it does. And his definition of DoF isn't wrong however it is not wrong other is good application for this definition in calculating the Hyperfocal distance and Image stacking. As it is based on Circle of Confusion. However it does not measure simply put the size of Bokeh Balls expressed in image size but in Pixel size.
This is the subtlety you have to deal with if you want the deep understanding.
A lot of strawman arguments and red herrings in your points, however, I'll just discuss one of your points which is somewhat relevant.
"Telling people that sensor size don't affect DoF is in general just wrong.
Why? you can't get a MFT 42mmf0.7"
Ok, well a 135mm f2 on full frame produces shallower dof compared to a 150mm f3.2 from Hasselblad. Even if you factor in the larger MF sensor Nikon and Sigma have a 105mm f1.4. How does your point make any sense?
Just because a particular lens doesn't exist doesn't make your point valid there are lots of examples of equivalent lenses available on smaller sensors that produce shallower DOF compared to larger sensors.
I'm sorry but that point you made is complete nonsense.
Well if instead of callings stuff nonsense you would actually read you would know that I gave you the tools:
f2 < f2,048 äq = f3.2*0.64 (Hasselblad crop)
but 150mm*0.64 = 96mm äq < 135mm
So I If you make the subject fill both the frame of the FF and the MF you should get the same DoF however you have to be closer with the Hasselblad.
Ok so what about a 105mm f1.4? that would be 164mm/f2.2 on a Hasselblad however they only have a 100mmf2.2 64mmf1.4 äq
so you need to be closer. But that is youst because they don't build one. There have been wider MF lenses in the past like the Mamya Sector C 80mmf1.9 = 50mmf.12 äq
and If you look at Fuji GFX you already can get the Zhong Yi 85mmf1.2 = 65mmf0,9 and you might see a 80mmf1.4 in the future.
However as available refraction indexes impose a theoretical limit at f0.5 (you can get lower with eg. pure Diamond) the way to shallower DoF is a larger Sensor!
Just read this:
https://en.wikipedia.org/wiki/Depth_of_field#Relationship_of_DOF_to_form...
135mm f2 on full frame is 135mm f2
150mm 3.2 on hassy small sensor is around 135mm 2.8 FF (0.2 crop factor)
150mm on larger hassy sensor is around 96mm f2 FF (0.64 crop factor)
105mm is f1.4
I think you might have the numbers wrong on this one.
If Fuji make an 80mm f1.4 that would be pretty amazing though.
Again your point doesn’t make much sense just because a particular lens doesn’t exist plenty of other examples.