10D / 300D gives a wider DOF than 35mm ?!?

[BrettD]

I wanted to know the practical effects of the 10D sized sensors on DOF. While there are many posts on this subject, none have convinced me, so I tried to work it out myself.

Could one of the more mathematically empowered members of the forum see if my logic here is sound.

I took a typical activity (a close up portrait) and simulated what I would do in each situation

35mm Film (or a 1Ds if you are lucky):
- Use an 80mm lens from 1.5m away.
- Use f/4.0

10D:
- Use a 50mm lens from 1.5m away.
- Use f/4.0

For the film, I used the common 1/100" @ 8"x10", giving a film plane CoC of 0.03303 mm:
- Hyperfocal Distance: 48.43 m
- Near focus limit with 1.5m focal plane: 1.46 m
- Far focus limit with 1.5m focal plane: 1.55 m
- DOF with 1.5m focal plane: 9 cm

The 10D's smaller sensor needs a CoC of 0.02142 to give the same results:
- Hyperfocal Distance: 29.18 m
- Near focus limit with 1.5m focal plane: 1.43 m
- Far focus limit with 1.5m focal plane: 1.58 m
- DOF with 1.5m focal plane: 15 cm

As you can see, the full frame sensor looks like it will have a narrower DOF than the 10D, when trying to take the same picture.

Brett D

[hmhm]

BrettD wrote:
Could one of the more mathematically empowered members of the forum see if my logic here is sound.
Brett D

Yeah, that's the idea, to adjust the focal length to preserve angle of view, and adjust the CoC to account for increased enlargement to a given print size. The same calculation applies to differences between film formats as well (e.g. 35mm vs 645 vs 4x5).

I'm curious how you calculated your CoC for the 10D, though, as it seems to be about 1/1.54 times your 35mm CoC, where it should be 1/1.59 (or 1/1.6 for "close enough" ) .
-harry
p.s. I now await the obligatory response of a representative of the "no, DoF is the same" faction.

[BrettD]

Well Spotted !!!

I was using diagonal measurements of the sensors, and could easily have got one wrong, can you spot the wrong number ?

film: 35 x 23.3
10D: 22.7 x 15.1

But it only makes about 1cm of difference to the 10D DOF so my question remains.

Brett D

[teddynet]

The DOF characteristics of a lens are relative to it's focal length not the size of the medium recording the image.

If you took an image on a 35mm camera with a 50mm lens and then cropped the resulting image to the size that a 10D sensor would have recorded the DOF wouldn't change.

The smaller sensor size simply records less of the same sized image at the focal plane, there is no magnification of focal length with a DSLR, it's only an apparent magnification due to a reduced field of view.

[hmhm]

BrettD wrote:
I was using diagonal measurements of the sensors, and could easily have got one wrong, can you spot the wrong number ?

film: 35 x 23.3


I've always seen 35mm film stated as being 36mm x 24mm.

The proper comparison is the ratio in one dimension, not the ratio of the diagonals, so it's 36mm / 22.7mm = 1.586 (or 1.6 rounded off).
-harry

[teddynet]

hmhm wrote:
p.s. I now await the obligatory response of a representative of the "no, DoF is the same" faction.


This depends on how you're looking at it, if you use the same focal length lens on each camera then the DOF will be the same but the FOV will differ, if you use different focal length lenses but match the FOV so the images appear the same in the viewfinder then the DOF will be different.

So for a 'standard' lens on each camera the DOF will be different but this is due to the different focal lengths not the 1.6 crop/magnification factor.

[hmhm]

teddynet wrote:
So for a 'standard' lens on each camera the DOF will be different but this is due to the different focal lengths not the 1.6 crop/magnification factor.


The case where you hold the focal lengths the same is fairly silly, as it's purely academic. The "answer", though, is that the 10D will demonstrate shallower depth of field due to the increased enlargement.

In the case where you adjust focal length to keep the FOV the same, the shorter focal length on the 10D will indeed increase depth of field. The added enlargement needed for the 10D to achieve a print of a given size will actually reduce depth of field, but this effect is linear whereas the effect due to focal length is exponential, so the net result is wider depth of field.

Depth of field depends on focal length, subject distance, aperture, and "circle of confusion". If you want to say that the size of the sensor has nothing to do with it, then answer the question "how does one arrive at the value for CoC?" The enlargement factor from the film/sensor to the print is an integral part of calculating a CoC.

There's an article on photo.net that explains all this.
-harry

[BrettD]

hmhm wrote:
I've always seen 35mm film stated as being 36mm x 24mm

Thanks.

Any idea where the 35mm comes from then???

Brett D

[robertwgross]

Isn't 35mm the overall width of the film?

Then 24mm is the usable width of the frame on the film, or height, if you look at it that way.

---Bob Gross---

[BrettD]

Thanks Bob, That sounds somewhat familiar, all is now clear.

Brett D

[leony]

The 10D uses the same lenses as 35mm EOS cameras. What is different is the size of the sensing medium. DOF is not dependent on the size of the chip, as any investigation will show.

A 50mm lens will produce the same "compression" or DoF on 10D as on 35mm. 10D just "crops" the image. Think of it this way: if you take a picture with 35mm scan it in PS and then crop the central part of it, the compression (and hence DoF) will remain the same while the "angle of view" is smaller.

To get the same "angle of view" on 10D as 35mm you need to use shorter lenses:

80mm lens on 35mm AND 50mm on 10D will produce the same "angle of view" - 50 * 1.6 = 80
but the compression of 50mm lens is that of 50mm lens, which is less than that of 80mm lens, hence greater DoF.

It's easier to understand this if you start thinking in terms of "CROPPING" not "FOCAL LENGTH".

The reason Digital P&S cameras have such great DoF is that their lenses are shorter (G2 - 7mm-21mm focal length) since the chip is smaller and that is required to maintain a reasonable angle of view. Think of how much DoF an 8mm full-frame fish-eye has on 35mm...

Hope this explains it.

[JamieAU]

Is it just me or does all this mathematical talk turn others on?

[hmhm]

leony wrote:
Think of it this way: if you take a picture with 35mm scan it in PS and then crop the central part of it, the compression (and hence DoF) will remain the same while the "angle of view" is smaller.


And then when you take that smaller image, and enlarge it to a print of a given size, say an 8x10, the extra enlargement needed to make that smaller "crop" into an 8x10 will enlarge the out-of-focus "blur", making it more apparent on the print, thus reducing depth of field. In effect, you need a smaller circle of confusion on the 10d sensor, since that CoC is enlarged to a greater extent in forming the print.

It is correct that the smaller focal length has a greater effect than the greater enlargement, but both are legitimate factors, and those who say "but the lens makes the same image at the film/sensor plane" are overlooking the fact that this image is enlarged to a greater extent in the smaller sensor.
-harry

[EXA1a]

jamieau wrote:
Is it just me or does all this mathematical talk turn others on?
This thread is about understanding the principle and you can't do that without some science/mathematics behind it.
There are many people out there (and in this forum, too) who don't understand the principle of "depth of field". If you do, you don't have to work through this kind of threads, just ignore it. If you don't and if you want to understand, you could read some textbooks or try to get the essence out of these threads.
Without a basic understanding (of DOF and other principle) you would not be able to utilize your full potential in photography.

--Jens--

[RichardtheSane]

I am not even going to go down the line of questioning your logic, but I do have one question.
You say the depth of field changed because of increased enlargment. You use 10x8 as an example.
Are you saying then that the DOF changes if I crop my 35mm slide down to the same size as the 10D sensor and print the cropped slide at 10x8?

[EXA1a]

yep. that's the point.

[BrettD]

RichardtheSane wrote:
I am not even going to go down the line of questioning your logic, but I do have one question.
You say the depth of field changed because of increased enlargment. You use 10x8 as an example.
Are you saying then that the DOF changes if I crop my 35mm slide down to the same size as the 10D sensor and print the cropped slide at 10x8?

Yes, absolutely correct. There is only ever 1 plane in focus, the rest of the DOF area is simply within what we deem to be acceptable focus.

To measure DOF, you need to define exactly how big a single point is allowed to be in a real world print. The commonly accepted size is 1/100" on a 8"x10" print.

So as you need to enlarge a crop more than a full frame, you are enlarging the focus error in the parts of the photo that were 'just inside acceptable focus' to outside acceptable focus.

Brett D

[BrettD]

jamieau wrote:
Is it just me or does all this mathematical talk turn others on?

I was hoping this thread could avoid the complex mathematics (that is why I left out the contents of the huge spreadsheet that helped me get to my results).

I just wanted to point out (and have it confirmed) that for a typical activity ie. I am here, subject is over there, and I use my zoom lense to frame it just how I like it, then at the same apeture, the 10D will in fact give a slightly wider depth of field when I print the picture than my old film SLR would have.

Brett D

[PaulB]

DEPTH OF FIELD is dependant upon three factors:

1. Focal length of lens.

2. Aperture

3. Subject distance

End.

DEPTH OF FOCUS is measured at the film/sensor plane and can be defined as:

the distance away from the focal plane that an image may be and still meet a given resolution specification.

This resolution specification is defined by the CIRCLE OF CONFUSION adopted for use in the measurement.

No amount of spreadsheets and measurement will alter the laws of optics and physics I'm afraid.

[EXA1a]

PaulB wrote:
DEPTH OF FIELD is dependant upon three factors:

1. Focal length of lens.

2. Aperture

3. Subject distance

End.



NOOOOOOOOOOOOOOOOO!

I thought some people start to understand the principle of Depth of Field and you tear it down again.
Depth of Field is dependent of 4 factors:

1. Focal length of lens.
2. Aperture
3. Subject distance
4. Circle of Confusion

Please take a look at DOF calculations here:
http://dfleming.ameranet.com/equations.html

In a strict physical sense, without including the image capturing system (film, sensor, retina) with the equation, "depth of field" is not existing because only the focal plane is in focus, everything further away or closer is out of focus.

--Jens--

[PaulB]

Jens,
You are assuming that the perceived sharpnes on a print that it acceptable to you is the same as is acceptable to me. If our standards for sharpness differ then we have to adopt different circles of confusion to make our DoF calculations.

However the fact remains that when focusing upon a subject the DoF of any lens is still only dependant upon the three factors I quoted.
When making a print we have to accept that subjective rather than objective measurements have to be made to determine what is or is not within the depth of field apparent ON THE PRINT (or monitor or whatever).

There remains the question of a smaller area sensor and the DoF as compared to a full frame 35mm area sensor/film.
Using the same lens, same aperture, same subject distance then the DoF appearing on the surface of the sensor of a 10D will be exactly the same as on the full frame of a 35mm film camera.

I would imagine you have read the following:
http://www.photo.net/learn/optics/dofdigital/

[EXA1a]

PaulB wrote:
Jens,
You are assuming that the perceived sharpnes on a print that it acceptable to you is the same as is acceptable to me. If our standards for sharpness differ then we have to adopt different circles of confusion to make our DoF calculations.

However the fact remains that when focusing upon a subject the DoF of any lens is still only dependant upon the three factors I quoted.
When making a print we have to accept that subjective rather than objective measurements have to be made to determine what is or is not within the depth of field apparent ON THE PRINT (or monitor or whatever).

There remains the question of a smaller area sensor and the DoF as compared to a full frame 35mm area sensor/film.
Using the same lens, same aperture, same subject distance then the DoF appearing on the surface of the sensor of a 10D will be exactly the same as on the full frame of a 35mm film camera.

I would imagine you have read the following:
http://www.photo.net/learn/optics/dofdigital/



DOF is a subjective measure because it is related to the objective factors (focal length, aperture, distance) AND the subjectice COF.
Without COF (or with point-sahed COF) there's zero DOF.
With a given COF the DOF only relates to your given factors.
COF is nothing fixed. Generally "standard COF's" are being used (around 0.03mm for 24x36mm and around 0.02mm for 15x22mm)
If your personal COF has the value x for a 35mm film camera, then your personal COF is to be simply calculated by x/1.6 for 10D sized sensor.
On the surface of the sensor there is no DOF defined. A DOF comes in play after you define a COF.

I give up.
--Jens--

[PaulB]

Jens wrote:
If your personal COF has the value x for a 35mm film camera, then your personal COF is to be simply calculated by x/1.6 for 10D sized sensor.

I always assumed that the critera to choose the CoF for a digital sensor was only for it to be bigger than the pixel size and therefore would vary with the pixel size on different sensors - ie. the CoF for a 10D would be calculated differently to a1D1Ds etc?

And if the DoF is simply a subjective measurement - made so by the choice of CoF then why try to give a 'one size fits all' answer?

A selected point focused on is either in focus or not - which was the point I was trying to make - the outcome of the finished image is all down to subjective (artisitic?) choices made by each photographer (which is another topic altogether!).

[teddynet]

If you're going to factor in subjective values such as print size then you should also include viewing distance, a slightly out of focus image looks sharper from the other side of the room.

We now come to the conclusion that the depth of field of an image can be altered even after it's been framed and hung on a wall.

If this is the case then the list needs revising:

1. Focal length of lens
2. Aperture
3. Subject distance
4. Image size
5. Print size
6. Print quality
7. Viewing distance
8. Viewer's eyesight
9. Viewer's opinion on acceptable focus

This is clearly nonsense.

Another conclusion we could draw from some of the posts here is that if an image doesn't contain enough depth of field then you can increase depth of field by simply making the image smaller. So thumbnails have more depth of field than the full size image? Is the opposite also true, can we blur the background of an image by printing it bigger?

The bottom line is that sticking a 50mm lens on a 10D doesn't get you the same depth of field as you'd get from a 80mm lens on a 35mm film camera, even if you match the field of view.

The answer to the original poster's question: "10D / 300D gives a wider DOF than 35mm ?!?" is Yes, ceteris paribus.

[hmhm]

teddynet wrote:
If this is the case then the list needs revising:
...
This is clearly nonsense.


Actually, this isn't nonsense at all, it's right on. Depth of field defines the distance fore and aft of the focus point in which the subject is "acceptably" in focus.

That "acceptably" is important, because what may appear to be acceptably in focus in a thumbnail may not be acceptable in a large print. We need to think of depth of field as meaning "what do I have to do to make the image good enough to allow viewing under my intended conditions".

The "rigorous" approach to this is to combine enlargement (and thus print size), viewing distance, and some subjective notion of "good enough" (possibly derived based on some rule of thumb regarding a typical person's eye's ability to resolve fine detail), and calculate a "circle of confusion" from that. This CoC should be thought of as a "requirement", i.e. my camera needs to capture this CoC or better, or else once I make the enlarged print that I intend to make, and view it from the distance I intend to view it, my subject won't meet my required "sharpness".

So, yes, all of those things you've listed do change depth of field. We shouldn't think of depth of field as being an attribute of an image, etched in stone when the shutter button is clicked, but rather a requirement that a captured image needs to meet in order to be "good enough" for our intended manner of viewing that image.
-harry

[hmhm]

Perhaps a more concise way to say this is that an image's "depth of field" is the range where the "requirements" for sharpness are met. Those requirements are based on how you view that image afterwards, i.e. size of print and viewing distance. If, after taking the shot, you change the requirements, e.g. you decide to print larger, or view closer, then the depth of field of the captured image necessarily changes. Even though the shot has already been taken, its depth of field has changed, because the range of distances in the image that meets our requirements for sharpness necessarily changes when we change our requirements.

This has very real world applications. If you're shooting images that will be viewed as little squares on a web page, your requirements for depth of field are _much_ looser than if your shot will be viewed as a giant print in a gallery. For instance, you can easily craft a situation where an architecture shot taken for small web use can be made at f/4 and look just fine, while a gallery print of the same subject would really need to be taken at f/22 or something.

Our depth of field calculators all have assumptions about final use built into them, typically something like an 8x10 viewed at some standard distance.
-harry

[DaveG]

To take a step back a bit about D of F. I have a 300 mm lens for my 8x10. I also have a 300 mm f2.8 lens for my 35 mm Nikon. At f11 the depth of field for these two lenses is identical. But on the 8x10 the 300 is a normal lens. On the Nikon it's a 6X telephoto. If I use the 300/8x10 for say a 3/4 portrait I'll need f22 or f32 to get any depth at all. If I was doing the same shot with a 35 mm camera I would be using a 50 mm lens and obviously at f22 I'd be getting a ton more depth of field than with the 300.

So you could say that the 35 mm shooter will have more depth of field for the given shot. This should hold true for the smaller 10D "film size" as well. You'd use something like a 30 ish mm lens to do the same shot and that lens will have more D of F for a given aperture.

As a complete aside (and some fun research) have a look at the Playboy centrefold. It's shot with an 8x10 and either a 300 or 480 mm lens. The model is ALWAYS on one plane since the D of F has barely the depth to cover what is unusually deep on her, and these would be shot at around f32 or more.