DOF comparisons

[kallisti]

Did a test series comparing depth of field at various apertures on a DSLR and a point-and-shoot. DOF varies by focal length, aperture, and distance. Subject distance was held relatively constant throughout and I tried to give equivalent fields of view for the shots. I was handholding though, so they aren't perfect.

The DSLR was a Nikon D700 with a 24-70 f/2.8 zoom lens. The point-and-shoot was a Leica D-Lux 4. All shots taken in aperture priority mode (obviously). Since the purpose of this test relates to DOF, no effort was made to hold white balance constant or even exposure constant. Whatever the camera meters said was right is what I shot. All shots are JPEGs right off the memory cards without any tweaking done in Aperture.

First the DSLR shots:

Nikon D700 f/2.8 60mm

Nikon D700 f/4 58mm (got a little zoom creep in this series)

Nikon D700 f/5.6 58mm

Nikon D700 f/8 58mm

Nikon D700 f/11 56mm (more zoom creep)

Nikon D700 f/16 56mm

Nikon D700 f/22 56mm

Now the point-and-shoot:

Leica D-Lux 4 f/2.8 12.8mm (but similar field of view to 60mm in 35mm terms)

Leica D-Lux 4 f/4 12.8mm

Leica D-Lux 4 f/5.6 12.8mm

Leica D-Lux 4 f/8 12.8mm

Even though the field of views are roughly similar throughout, the focal lengths needed to produce them aren't. 60mm on a full frame body is close to 12.8mm on a point-and-shoot. So even wide open at f/2.8 the Leica has a fairly large DOF (to my eyes equal to about f/11 on the Nikon). Because of their tiny focal lengths, it is impossible to "blur out the background" with point-and-shoots even at large apertures. The only way point-and-shoots can isolate a subject is in macro mode with VERY close subjects. At very short subject distances, DOF shrinks rapidly even with small apertures.

 

[MattSepeta]

Interesting!

I like this... What inspired you to do this study?

 

[kallisti]

I like this... What inspired you to do this study?

There was a thread the other day (here) by someone new to photography who was frustrated that he couldn't "blur the background" with his new point-and-shoot despite the fact that the manual evidently showed an example shot with an isolated subject.

So thought I would do a series with an actual comparison to visually show why it isn't possible.

What I didn't include above is the changes to exposure that happened at the various apertures. With the DSLR @ f/22 the ISO was 2500 and exposure time was 1/60 sec. The point-and-shoot had a larger DOF @ f/8, as well as having an ISO of 400 and exposure time of 1/125 sec.

 

[NeGRit0]

Wow, this was an interesting read. Thanks for sharing!

 

[kallisti]

Wow, this was an interesting read. Thanks for sharing!

Thanks. Had other things going on and forgot to follow up on this thread. Nice of you to respond.

Just realized that I shot the Leica in 4:3 aspect compared to the 3:2 aspect of the Nikon. This explains some of the difference in field of view between the two cameras (the Leica shots include more vertical while the Nikon shots have a bit more horizontal). Doesn't impact on the "teaching point" though.

When comparing the DOF of a full frame DSLR to a P&S, the P&S has a fairly generous DOF even at a "fast" aperture secondary to the extreme wide angle focal length these cameras employ. While the combination of a small sensor and extremely wide focal length may result in an equal field of view compared to a "normal" focal length in 35mm terms, the DOF is related to actual focal length and not apparent focal length. P&S cameras are incapable of isolating subjects when compared to a DSLR. That's one of the reasons people are willing to put up with the added bulk and cost of a DSLR.

 

[FX120]

People never seem to get that DOF has little to do with sensor size, only the optics up front.

 

[kallisti]

People never seem to get that DOF has little to do with sensor size, only the optics up front.

You are agreeing with me I think? DOF relates to the actual focal length of the lens on the camera, not the "apparent" focal length used in camera marketing. It is also related to subject distance and aperture, but with the extremely wide focal lengths of P&S lenses the "actual" focal length of the lens trumps the other two variables in the vast majority of cases, resulting in crazy-large DOF compared to what can be achieved full-frame either digital or film. Or are you not agreeing with me?

 

[GoCubsGo]

People never seem to get that DOF has little to do with sensor size, only the optics up front.

I must not be around much, I've never actually heard someone say the DOF directly correlated to the sensor size.

 

[TH3D4RKKN1GH7]

What makes people believe that is when you do a simple test, Take shot of subject here with this camera and again here with camera B, the one with the larger sensor will always have the narrower depth of field. People don't take into account the other bits that you have to compensate for when shooting with the smaller sensor camera.

 

[GoCubsGo]

^ I guess when you put it that way I have heard/read that before. I'm pretty sure I gloss over such statement because to me, DOF is largely controlled by the optics.

 

[TH3D4RKKN1GH7]

In other news, DAMN THAT D700 SMOKES THAT P&S LOL.

 

[Edge100]

People never seem to get that DOF has little to do with sensor size, only the optics up front.

Not quite. Bigger sensors require that you get closer to the subject to get the same field of view. Closer to the subject means smaller DOF, even with identical f-stops and field of view.

So sensor size doesnt change DOF per se, but it does make you get closer, and that most certainly does change DOF.

 

[kallisti]

Not quite. Bigger sensors require that you get closer to the subject to get the same field of view. Closer to the subject means smaller DOF, even with identical f-stops and field of view.

So sensor size doesnt change DOF per se, but it does make you get closer, and that most certainly does change DOF.

Actually, I'm not sure I agree with this statement. In the above test I was equally far away from the subject with both cameras. I didn't need to get closer with one camera to preserve the same field of view. It is just that for an equivalent field of view with the P&S the focal length was much, much smaller because the sensor is much, much smaller. Subject distance for both was approximately 3 feet. In both cases the field of view was near the upper limit of the focal lengths the lens supports (12.8mm in the case of the Leica 5.1-12.8mm lens, 60mm in the case of the Nikon 24-70mm lens). I chose these camera/lens combos because their zoom range (with regard to field of view) is almost identical (24-60 for the Leica, 24-70 for the Nikon when both are compared in 35mm full frame terms) and the maximal aperture of the lenses is identical (the Leica can go to f/2 when shooting a wide-angle field of view, but when zoomed out like here it's maximal aperture drops to f/2.8, exactly the same as the Nikon.

Again, the key point is that even though both lenses have roughly the same zoom range with regards to field of view and both lenses are equally fast, the lens in the Leica is really a 5.1-12.8mm lens. Can you even imagine shooting with a 5.1mm lens on a DSLR? You almost wouldn't need a focus ring--the hyperfocal distance on that lens at f/2.8 would be 12". Meaning that if you focused the camera at 12", everything would be in focus from 6" to infinity. How's that for broad DOF? That's exactly what the Leica delivers though when shot at it's widest angle of view. Too bad the overall image quality can't compare to the Nikon since from a creative standpoint there might be times where having that broad of a DOF at f/2.8 would be useful.

 

[kallisti]

....Too bad the overall image quality can't compare to the Nikon since from a creative standpoint there might be times where having that broad of a DOF at f/2.8 would be useful.

I will try to give a concrete example as to why a P&S might actually be better than a DSLR in some circumstances.

Let's say you are traveling on vacation and find yourself in an interesting locale. After dinner you find yourself walking around at night and see an interesting low-light scene that you wish to shoot handheld. You aren't carrying around a tripod because you are on vacation with friends/family. Before you went out that night you had two options: you could carry your DSLR or a P&S.

The shot you want to capture requires a "storytelling" aperture to use Bryan Peterson's term (i.e. you want maximal DOF, everything in focus). Your P&S has a fast aperture (f/2.8) and produces acceptable images up to ISO 400. Your DSLR is a Nikon D700 coupled with a 24-70 f/2.8 zoom and produces acceptable images with an ISO of 2400. Assume the lighting requires an exposure on the P&S of f/2.8 1/60 sec ISO 400. On the D700 to capture the same DOF, you will need to shoot at f/11. This is 2 stops lower which you will need to make up somewhere. Assuming you choose to bump the ISO to 1600 you can keep the same shutter speed of 1/60 sec. You could bump the ISO even further to achieve a shorter shutter speed and still walk away with an equal or better image. Because of it's full-frame sensor, the noise on the D700 at a higher ISO will be negligible. So you can still make this shot hand-held and may very well end up with a higher quality image than the P&S produces.

Now let's change things a bit and assume that your DSLR is a Nikon D40. This is an older generation DSLR with a crop sensor. It has much worse noise characteristics at high ISOs. Let's assume that the highest ISO where you can obtain acceptable noise is ISO 800. For the same shot where you need to shoot at f/11 on a DSLR for the DOF you require, your shutter speed will now need to be 1/30 sec at ISO 800. You are much more likely to get blur from camera shake at this shutter speed. Let's say that you actually require an ISO of 400 on the D40 to reach acceptable noise levels. Now your shutter speed is 1/15 sec (which is going to introduce significant blur when handheld). God forbid that the DOF you want requires a smaller aperture than f/11. In that case there is zero chance you will get a sharp image hand-held. Remember that the P&S is shooting at 1/60 sec for the same exposure and DOF. It's also lighter and has built-in image stabilization which will further reduce the impact of camera shake.

In this scenario, the P&S may actually produce a better image than the DSLR. Subject sharpness is often the most important characteristic of an image. If your subject is blurry from camera shake due to slow shutter speeds then it doesn't matter how much money you spent on that "pro" lens. If it's blurry then it's a crappy shot. P&S cameras excel in situations where you need maximal DOF in low-light situations where a tripod isn't feasible. They are also less bulky and weighs less, which means there is a higher chance you will actually be carrying a camera to catch "that shot" when you see it. High-end rangefinder cameras (like the Leica M9) find a happy medium between bulk and image quality. They will also run you $10k (or more) to achieve this "happy medium." Pricey, to say the least. DSLRs aren't always better than P&Ss.

 

[FX120]

Not quite. Bigger sensors require that you get closer to the subject to get the same field of view. Closer to the subject means smaller DOF, even with identical f-stops and field of view.

So sensor size doesnt change DOF per se, but it does make you get closer, and that most certainly does change DOF.

If you put a true 50mm lens on a camera with a 4/3'rds sensor, opened it up to say f/2, and focused it on an object 10 ft away you would get the exact same amount of blurring on out of focus objects as you would have with a FF DSLR with the same 50mm lens at f/2 from 10ft away. Your field of view would have changed, but assuming that the optical properties of the lenses were the same the exact same amount of out of focus blur would be present.

However, if you put a 25mm lens on the 4/3'rds camera to get the same FOV as your DSLR, you would now have less out of focus blur because you've changed the focal length of the optical system, and hence your effective depth of field.

It's not the sensor, it's the optics.

 

[OreoCookie]

It's not the sensor, it's the optics.

Not quite. Photographers don't choose focal lengths, they choose viewing angles (we just got accustomed to use mm). Physically, you're correct, focal length, distance to the subject and aperture determine the depth of field. However, photographers choose viewing angle, distance to subject and aperture. If I use a 4/3rds camera for my photography, then I'd use a 25 mm lens instead of a 50 mm lens to get the same photographical effect. On P&S cameras 50 mm is already a `very, very long focal length' (i. e. it has a very small viewing angle).

From the creative point of view, you need to compare equivalent focal lengths rather than lenses of the same focal length -- and there, the size of the sensor does come into play.

 

[sangosimo]

sensor size can control depth of field and field of view.

Aperture and focal length are optical constants

most people understand this "crop factor"

If I shot an image with my 5d (full frame body) @ 50mm and 2.8, and I wanted to recreate the image on a 7d; I would need to use the 75 @ 31.25mm and ~f1.8. The 7d has a 1.6 crop.

 

[mtbdudex]

OP;
Thx for this thread, I sent link to my wife, brother inlaw & a few friends, great visual for them (and I ) DOF.

 

[Edge100]

If you put a true 50mm lens on a camera with a 4/3'rds sensor, opened it up to say f/2, and focused it on an object 10 ft away you would get the exact same amount of blurring on out of focus objects as you would have with a FF DSLR with the same 50mm lens at f/2 from 10ft away. Your field of view would have changed, but assuming that the optical properties of the lenses were the same the exact same amount of out of focus blur would be present.

However, if you put a 25mm lens on the 4/3'rds camera to get the same FOV as your DSLR, you would now have less out of focus blur because you've changed the focal length of the optical system, and hence your effective depth of field.

It's not the sensor, it's the optics.

Your example is right, but it misses the point. If I want a certain shot, I frame it based on focal length AND distance to the subject.

On a FF body with a 50mm lens, I have to be half the distance to the subject to get the same framing as the 4/3s sensor body with the same lens. Ergo, the DoF will be much less with the FF body since I'm closer to the subject.

You're right; the sensor does not change the DoF per se; but the sensor does require you to change the camera-subject distance based on the crop factor, and this most certainly does change DoF.

 

[ChrisA]

People never seem to get that DOF has little to do with sensor size, only the optics up front.

That's not quite true. The sensor size will determine what focal length lens you will need and the DOF is a function of focal length as well as aperture. With the small sensor you will be using a short lens and the DOF will be less. But then there is one more important factor the "circle of confusion" the larger camera will have larger pixels. You criteria for what is in focus depends a lot on the pixel size.

You don't have to do a series of tests you can calculate the DOF

So except the acceptable "circle of confusion" may depend on the sensor size, you are right DOF depends only on the focal lenght and aperture but in any reasonable real-world case the focal length used will depend strongly on the frame size.

"Fixed focus" cell phone cameras take advantage of this effect. they use very tiny sensor so the DOF goes from right in front of the lens to infinity.

Still you CAN blur the background with a P&S camera. You just need to set up the shot differently and change the composition a little. Basically you'd move up closer and use a wider lens shooting with the subject very close to the camera. Buit this is a different perspective with the subject looking larger relative to background objects