Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
04-08-2009, 01:51 PM   #31
Veteran Member




Join Date: Sep 2007
Photos: Gallery
Posts: 975
QuoteOriginally posted by falconeye Quote
@nanok: Thanks for trying to make things clearer.

Because the topic about DoF tends to confuse many people and answers tend to be long, I established my take at it:

My short "FoV and physical aperture is all" statement.

I try to make this a rock in an ocean of incertitude to help people out. A certain level of arrogance helps in making it a rock. Sometimes. One simply cannot argue everything and forever. But I agree to your statement about my arrogance
i know exactly what you mean, and i think it is a good (complementary) aproach (i know it works for me: if i am intrigued enough, i will look it up and figure out if it's true or not), i find it also helps though, for "the rest of the people" if somebody is more patient and tries to shed some light (i guess some would call it similar to the "good-cop -- bad cop" negotiation strategy). i do agree "you cannot argue everything for ever"

QuoteQuote:
@Ron: Nobody called you an idiot (I am not that arrogant). I assumed You have made observations and wanted to learn more about it.

That DoF is shallower for larger formats was never put into doubt by anybody in this thread. As nanok explained.

It may help to realize that the term "aperture" can have two meanings:
- the absolute (or physical) aperture (in mm)
- the relative aperture (or f-stop value, which is the quotient of focal-length / physical aperture -- the reason why you write f/16)
good point, this would need to be clearly stated as well, another case of "obvious but often overlooked"

QuoteQuote:
(...)
@nanak:

About the circle of confusion CoC...

In DoF calculations, the circle of confusion is mathematically defined to be

"diagonal size of medium" / CONSTANT

where the CONSTANT is constant reflecting an observer's biology (the human eye).

The only convention is which exact value is chosen for the constant.

The most common convention is the "Zeiss formula" which is CONSTANT=1730.



Obviously, with 1920x1080 HDTV, the constant ought to be 2200
i am aware of that. let me further clarify my choice to bring coc into the discussion (against your decision, which i did understand): if we are talking dof, for people to understand what the hell that is and how it works, they should definetly understand this one small detail too (which brings this "fuzzyness" factor into the dof discussion, and shows it is actually nothing set in stone). as long as you stay within some "reasonable limits", you can assume a certain coc for a certain medium size/type, but when joe thinks of dof refering to 10x15 cm prints, whereas jim is thinking of 1m wide prints, (from the same sensor), you will be in big trouble, and there will most probably be a fight (though both of them are right, the only problem is they are talking about a completely different thing, which they both happen to call dof). to avoid the bloodshed, and in the interrest of better understanding and less frustration, "i proclaim the circle of confusion a needed point to be discussed" :-P. enough said, i guess it should be clear

04-08-2009, 02:03 PM   #32
Veteran Member




Join Date: Sep 2007
Photos: Gallery
Posts: 975
QuoteOriginally posted by Wheatfield Quote
You can actually pretty much forget about the circle of confusion thing that people like to bat around.
Depth of field is controled by two things and two things only:
1: Image magnification on the sensor/film
2: Aperture being used.

Thats it.

The depth of field doesn't change depending on the size of the print, it's just that a bigger print makes it easier to see.
aahemmm.. no, actually that's simply wrong, and it effectively means that me, flaconeye, and some good people here (*) have been wasting time and bandwidth for nothing . how better (further) can it be explained..?

depth of field does not become more or less obvious depending on print size, it changes completely (namely, something which in a bigger print was 1m out of the dof zone is in a smaller print in focus). not only does the size of print change this, but, to make things even more confusing, viewing distance of the same print also does! look at one of your favourite 20x30 (roughly a4) prints, and look at the same picture on your computer, full screen, chose a print which has a wide field in the frame, and with both in and out of focus areas (the smoother the transition the easier to see the difference). for an even more extreme and obvious example, try the same experiment with an a4 (or even smaller, 5x7in or something) print and the same picture on your cameras back lcd. if you still don't agree that "the size does matter", i see no other escape for me than to just give up :/

---------
* sorry falconeye, the good people are always the "other" people, you're in the "us" bandwagon on this one, with me
04-08-2009, 03:29 PM   #33
Veteran Member
Sean Nelson's Avatar

Join Date: Sep 2008
Location: Vancouver BC Canada
Posts: 353
QuoteOriginally posted by Wheatfield Quote
The depth of field doesn't change depending on the size of the print, it's just that a bigger print makes it easier to see.
Nanok's right, this is simply wrong. Your statement reads as if depth of field means that everything within a certain range is in perfect focus and everything outside that range is not. That is not the case - in fact nothing is in perfect focus except objects that are in the lens's exact focus plane. Everything else is blurry - the further from the focus plane, the blurrier. "Circle of confusion" is the technical means of describing exactly how blurry.

Objects that are within the depth of field aren't perfectly sharp, it's just that the degree of blurriness is so small that you can't see it. But the more you enlarge the image, or the closer you get to it, the easier it is to see the blurriness, and this means that the depth of field is less.
04-08-2009, 03:44 PM   #34
Veteran Member
Sean Nelson's Avatar

Join Date: Sep 2008
Location: Vancouver BC Canada
Posts: 353
After the umpteenth attempt to explain with words, perhaps a couple of images will make it clearer. These two pictures are exactly the same image taken with the DA35mm on my K100D Super. The first is the entire image scaled down (bicubic scaling, no other changes) to a size that should appear as 4 x 6 inches on a 96dpi monitor. The second is a crop of the centre at 1:1 scale - if the entire image were displayed at this scale on a 96dpi monitor it would be 20 x 30 inches.


In the first, the depth of field appears to be around 1cm, centered around the 10cm mark on the ruler:




In the second, the depth of field appears to be around 2mm:



If you say "the depth of field is actually the same, you just can't see it as well on the smaller print" then you're MISSING THE POINT. If you can't see it, it isn't there - for all practical purposes the smaller image has more in focus. If you're taking the picture with the intent of displaying it on the web at a small size, and if you want to limit depth of field so that the subject stands out against the background, you have no choice but to take into account the fact that your image shown at the smaller-than-normal size will have more depth of field than the standard DOF scales indicate.

To look at it the other way - if I had taken this picture with a K20D and displayed IT at 1:1 scale, the higher resolution should show that even less depth was sharply in focus. Conversely, if I show the image as a 120 x 80 pixel thumbnail, then even more will appear to be in focus. The whole DOF concept is relative - the more you magnify the picture, the less depth of field you get.

04-08-2009, 05:11 PM   #35
Pentaxian
falconeye's Avatar

Join Date: Jan 2008
Location: Munich, Alps, Germany
Photos: Gallery
Posts: 6,863
QuoteOriginally posted by nanok Quote
aahemmm.. no, actually that's simply wrong, [...] depth of field does not become more or less obvious depending on print size, it changes completely
QuoteOriginally posted by Sean Nelson Quote
Nanok's right, this is simply wrong.
I thought everything has been said. And now we start to contradict ourself

DoF does depend on how much an image is cropped. But it does not depend on how large an image is printed out.


Some posts up, my comment about HDTV and the Zeiss formula (i.e., the proximity of values 1730 and 2200) gave me a new idea and a new twist to our discussion.


THE ALL NEW DEFINITION OF DoF
=============================
Def.: Subjects are within the depth of field area of an image if they appear to be completely sharp on a FullHD HDTV screen displaying the total of this image.
The important point here is that if you approach the screen, you start to see pixels. Therefore, subjects don't start to appear soft. Of course, you shouldn't watch a FullHD HDTV screen so closely that you can see the pixels. Not if you don't like headaches

A larger screen (say 60" rather than 32") is viewed from a larger distance. So, DoF doesn't change.

The century old mathematical definition of DoF works exactly as my HDTV analogy. Hope it helps.
04-08-2009, 06:05 PM   #36
Pentaxian
Wheatfield's Avatar

Join Date: Apr 2008
Location: The wheatfields of Canada
Posts: 10,108
QuoteOriginally posted by nanok Quote
aahemmm.. no, actually that's simply wrong, and it effectively means that me, flaconeye, and some good people here (*) have been wasting time and bandwidth for nothing . how better (further) can it be explained..?

depth of field does not become more or less obvious depending on print size, it changes completely (namely, something which in a bigger print was 1m out of the dof zone is in a smaller print in focus). not only does the size of print change this, but, to make things even more confusing, viewing distance of the same print also does! look at one of your favourite 20x30 (roughly a4) prints, and look at the same picture on your computer, full screen, chose a print which has a wide field in the frame, and with both in and out of focus areas (the smoother the transition the easier to see the difference). for an even more extreme and obvious example, try the same experiment with an a4 (or even smaller, 5x7in or something) print and the same picture on your cameras back lcd. if you still don't agree that "the size does matter", i see no other escape for me than to just give up :/

---------
* sorry falconeye, the good people are always the "other" people, you're in the "us" bandwagon on this one, with me
You probably think that focusing screens change their transmissive properties when a brighter lens is put on the camera too.
This isn't a point I'm going to argue, I have better things to do than try to shoot down internet myths.
04-08-2009, 11:32 PM   #37
Veteran Member
Sean Nelson's Avatar

Join Date: Sep 2008
Location: Vancouver BC Canada
Posts: 353
QuoteOriginally posted by falconeye Quote
DoF does depend on how much an image is cropped. But it does not depend on how large an image is printed out.
I'd agree with this if you adjust your viewing distance so that the images printed at different sizes fill the same angle of view in your visual field. But if you print the same image at 8" x 10" and again at 80" x 100", then view the two prints from 2 feet away, you're going to see stuff not in focus in the larger print that appears sharp in the smaller one. And the reverse is also true, if you shrink the image to 0.8" x 1.0" and view it from 2 feet, you're going to have more apparent depth of field.

In terms of prints this kind of thing usually doesn't have that much impact because people generally hold or view a print at a comfortable distance so that it fills a similar proportion of their field of view. But, for example, in web page design when you choose how big to show an image people don't generally adjust their viewing distance to match. In cases like this you're essentially dictating the apparent size of the print, and so these effects on DOF are more important.

It's true that at when you get close enough or print large enough you will hit the resolution of the media, at which point everything starts to look soft. But that's an just an artifact the media (including HDTV) that doesn't invalidate the general fact that the more you magnify an image, the less the apparent depth of field.
04-09-2009, 02:51 AM   #38
Pentaxian
falconeye's Avatar

Join Date: Jan 2008
Location: Munich, Alps, Germany
Photos: Gallery
Posts: 6,863
QuoteOriginally posted by Sean Nelson Quote
I'd agree with this if you adjust your viewing distance so that the images printed at different sizes fill the same angle of view in your visual field.
I am with Wheatfield "This isn't a point I'm going to argue, I have better things to do than try to shoot down internet myths.".

However, if one made a statement one should be prepared to invest some time to explain it.

Sean, you seem to be rather knowledgable. So there is a chance we identify a common source of confusion here.


What you talk about is "Will I see softness from out of focus subjects?". Of course, if you approach / enlarge an image and if there were more pixels than the Zeiss formula requires (~3 MPixels), than you will.

But this is NOT what is called Depth of Field (DoF). You need to coin another term when speaking about this property! DoF is a well-defined term for over a century now and you cannot change its meaning at will.


Me, what I talked about is "What subjects are in the DoF area?"
And there is only one answer. A professor in optics wouldn't argue with you. You would just receive 0 points for a wrong answer

To help people understand how DoF is defined I used my HDTV example.


And you are right: If you blow up an image, you must increase your viewing distance accordingly. This really is part of how DoF is defined.



The area you talk about (Will I be able to see softness from out of focus subjects?") requires you to chose the Circle of Confusion to be 1 pixel and apply the DoF formulae. What you end up with is a new property which depends in a different way on factors than DoF does. You may call it "Digital depth of field" or DDoF.

DDoF is narrower than DoF for higher than HDTV resolutions. But you cannot see the difference from a viewing distance which let you keep the entire image in view. As you cannot see the pixels from an HDTV screen when watching a Bluray movie.


Please, in this forum, let's not confuse DoF with DDoF (or whatever you may want to call it).

04-09-2009, 03:14 AM   #39
Pentaxian
falconeye's Avatar

Join Date: Jan 2008
Location: Munich, Alps, Germany
Photos: Gallery
Posts: 6,863
QuoteOriginally posted by Wheatfield Quote
You can actually pretty much forget about the circle of confusion thing that people like to bat around.
Depth of field is controled by two things and two things only:
1: Image magnification on the sensor/film
2: Aperture being used.
@Wheatfield, this may be a corrolar to my two points rule (mm-aperture & FoV).


Can you please clarify what you mean by:

1. Image magnification on the sensor/film?
(Do you mean, e.g., 24mm / 50km for a mountain chain on APS-C etc.?)

2. Aperture?
(Do you mean relative or absolute aperture?)
04-09-2009, 04:37 AM   #40
Veteran Member




Join Date: Sep 2007
Photos: Gallery
Posts: 975
i agree, the dof is defined assuming a certain size of print should be viewed from a certain distance, which is why falconeye's "simple" statement is okay (as i said), what i was trying to point out (again), is that people trying to understand this should be able to understand the concept of circle of confusion and it's connection to dof, stating we can just ignore everything else only leads to trouble, as far as understanding is concerned, it leads to less headache after you understand what is involved and you are able to use the simplified statement effectively, being aware of what's behind it, but only after, not before. incomplete understanding and trying to reverse engineer the formula after without understanding what's behind it is a common source fo trouble (example, field of view simplified statements, it has happened in this forum: a 200mm "full frame" lens on aps-c is equivalent to a 300mm on film, a 200mm lens on medium format is equivalent to 100 on film, thus a 200mm medium format lens should be equivalent to a 400mm on aps-c )

what i am trying to say is that it would be best to help people understand what is behind, we can move to simplifying things for easier calculation after (such as, unifying print size + viewing distance, and getting rid of CoC altogether, by hiding it behind FOV). jumping from explaining in detail to very condensed rules all the time can only lead to confusion.

i do admit that using the dof term to talk about a "dof" which changes with viewing distance might be a bit uninspired, and in contradiction, as you said, with a century old definition which is accepted and works, so thanks for calling me on that one.

wheatfield: i am unsure what your last statement is supposed to mean, it is probably meant as an insult. i am trying to help out, not to prove that i am all-knowing, so i am okay with being corrected myself, i am sorry though if i have hurt your ego by disagreeing with you.
04-09-2009, 07:12 AM   #41
Pentaxian
Wheatfield's Avatar

Join Date: Apr 2008
Location: The wheatfields of Canada
Posts: 10,108
QuoteOriginally posted by falconeye Quote
@Wheatfield, this may be a corrolar to my two points rule (mm-aperture & FoV).


Can you please clarify what you mean by:

1. Image magnification on the sensor/film?
(Do you mean, e.g., 24mm / 50km for a mountain chain on APS-C etc.?)
Not quite certain what you are asking, but I mean to say the ratio of object size to image size on the capture medium.
Longer lenses and shorter object distances increase magnification, shorter lenses and longer distances decrease magnification.

QuoteQuote:
2. Aperture?
(Do you mean relative or absolute aperture?)
Again, not sure about your terminology, I suspect I mean relative though.
Having said that, I mean for any given f stop number.

COC selection does not change, improve, or degrade the DOF in an image,
it just changes the way you measure it.

Last edited by Wheatfield; 04-09-2009 at 08:23 AM.
04-09-2009, 09:10 AM   #42
Veteran Member
Sean Nelson's Avatar

Join Date: Sep 2008
Location: Vancouver BC Canada
Posts: 353
QuoteOriginally posted by falconeye Quote
And you are right: If you blow up an image, you must increase your viewing distance accordingly. This really is part of how DoF is defined.
My point is this: DOF calculations ASSUME that the viewer will view the image at a certain apparent size - that's usually a reasonable assumption but there are exceptions. The case I cited of a web page layout is one, photos in a magazine would be another. To give an HDTV example, a inset picture within the frame would be another.

To go to the other extreme, consider an Omnimax theatre. The 70mm x 48mm frame is projected onto a spherical screen that typically covers 180 degrees horizontally in the audience's field of view. Omnimax cinematographers have to be very careful with depth of field because it's much shallower than the "standard calculations" would imply.

In these situations, the standard assumptions don't apply because the viewer doesn't adjust his viewing distance. This means that for images in which DOF is important you have to be aware of the difference.

This is why I have some trouble with a blanket statement "DOF does not depend on print size". While "print size" is largely irrelevant (it's really APPARENT size that's important), IMHO it's misleading to simply claim it doesn't matter and leave it at that.
04-09-2009, 09:38 AM   #43
Pentaxian
falconeye's Avatar

Join Date: Jan 2008
Location: Munich, Alps, Germany
Photos: Gallery
Posts: 6,863
QuoteOriginally posted by Wheatfield Quote
I mean to say the ratio of object size to image size on the capture medium.

I mean relative though. Having said that, I mean for any given f stop number.

COC selection does not change, improve, or degrade the DOF in an image,
it just changes the way you measure it.
Thanks for the clarification.

All the math can be found on Wikipedia:
Depth of field - Wikipedia, the free encyclopedia

My "two factors only" rule was derived and explained here:
https://www.pentaxforums.com/forums/214871-post47.html
Only the (relative simple) formula for hyperfocal distance H is required here.


Your rule is correct as well. But we must take the following into account when applying it:
1. Subject distance s must be small and comparable to the focal length f rather than the hyperfocal distance. Otherwise, the math doesn't hold true anymore.
2. The Circle of Confusion must be considered a constant, i.e., the size of the recording medium must be constant.

Then indeed:
DoF ~= 2 F c ((m + 1) / m^2)
where F is the f-stop number, c is the Circle of Confusion and m is the magnification.

This does help a lot in practice.


Of course, in discussions about the influence of size of the recording medium, the formula must be re-expressed and becomes a lot more intricate (because then both c and m become dependent on another variable).

Also, one would have to determine if FoV or magnification should be kept constant when changing the size of the recording medium. One would probably request a higher magnification and a larger Circle of Confusion for a larger recording medium. In order to produce an equivalent image.

Last edited by falconeye; 04-09-2009 at 09:43 AM.
04-11-2009, 03:24 PM   #44
Veteran Member
Eaglerapids's Avatar

Join Date: Apr 2007
Location: Idaho,USA
Photos: Gallery
Posts: 1,588
Thanks guys, now I've got a massive headache:-).
04-11-2009, 07:00 PM   #45
Veteran Member




Join Date: Sep 2007
Photos: Gallery
Posts: 975
QuoteOriginally posted by falconeye Quote
(...)

Also, one would have to determine if FoV or magnification should be kept constant when changing the size of the recording medium. One would probably request a higher magnification and a larger Circle of Confusion for a larger recording medium. In order to produce an equivalent image.
i think maybe you meant "lower magnification and larger circle of confusion", typo maybe? (that is, unless by magnification you mean ratio between recorded image and "real life size")

in this respect, i think "your" formula (using fov) might be more straight forward for going back and forth across formats, though weatfields way of stating it has it's merits (for instance, when the requirement is to reproduce some real life object at a certain magnification ratio, it makes more sense to talk about magnification).

for the rest of us though, the "thickheaded crowd" , i think it should be enough to understand the basic concepts behind determining DOF for a given situation, rather than (or at least before) taking in the concentrated pill, which is the formula, most photographers will not need to actually compute dof, but it will help to know what influences it, to what extent, and thus how to (empirically) control it on the spot. that's why i insisted on not excluding CoC.

it would go something like this:

- know your intended final "product" (small print? web display? "gallery" print? highway billboard? slide projection?-- and this is not in any particular order ). bigger will most likely mean less dof (yeah, i know, heresy..) -- a 10x15cm print of the same picture will (appear to) have more dof than a 20x30cm one, hanging on the same wall; cropping will effectively decrease dof (as demonstrated above)

- smaller (relative) aperture will mean more dof (aka higher "f number" )

- closer subject (focus) will mean less dof (in the above formula it is also obvious: dof is dimensionally inverse proportional to the magnification)

- longer focal length will mean less dof, but this is something which must be taken with a big grain of salt: some people use different focal lengths depending on the perspective they want to get (that is, i choose a wideangle if i want to get close, i choose a longer lens if i want to move further -- and flatten the perspective while i do so -- this means the magnification of my main subject stays the same more or less, but the ratios between the various subjects in the frame change, it is a bit confusing to me to talk about one magnification when talking about DOF for this reason), other people will just use them to frame more or less, which means that perceived dof change will be different, as the distance does not change. i think it is safe to keep the statement though, qualitatively at least

- when crossing formats, the first problem is to figure out what we are comparing, and what is the comon ground. the most often common denominator seems to be field of view (which is supposed to be expressed in degrees of circle, not mm, btw ). if we nail fov down, larger format means less dof. it would be interesting to nail dof down, in that case (assuming lenses of any fov are available) it would mean larger format just means more fov (so basically, a larger image of a larger portion of reality, which is funny, because it is actually so simple and straightforward a way to put it that it almost takes away the "magic" of larger formats)


(this goes a bit beyond "what the formula means for photographers" and more towards "what to keep in mind while shooting)
- one obvious point is that DOF is not specific to a (3d, real-life) scene, but to the way we choose to look at it (distance, angle of view, mainly). that is to say, choosing a lens and aperture at a certain scene does not define the dof alone, the choice of "where to shoot from" is essential (note that i am abusing the term dof really: this is dof as in "objects in the real life scene which are perceived as in focus in the projected image", which is probably quite a leap). this does not change dof formulas, but it does change the final outcome so i dared to mention it

- another interesting consequence of the second point (distance) is that, while keeping the (assumed single) subject within the dof range (so "in focus"), one can get more dof by focusing further (less separation from the background) or less by focusing closer (in extremis, one might choose to leave the subject near the upper limit of the DOF range, to gain better separation, though this rarely seems to happen)

(looking at it now after i wrote it, i wonder if it can help anyone: read: if anybody will actually read all this ; what do you think?)

Last edited by nanok; 04-11-2009 at 07:12 PM.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
camera, crop, dof, forum, k20, k20d, knowledge, lenses, medium format, sensor, shot
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Sensor size vs. DOF future_retro Photographic Technique 24 09-16-2010 04:30 PM
Do you shoot and crop in camera or shoot loose and crop in photoshop? hockmasm Pentax DSLR Discussion 23 05-02-2010 06:10 PM
DOF and Crop Sensors edl Pentax SLR Lens Discussion 16 11-28-2009 01:11 PM
DOF & Crop factor octavmandru Pentax DSLR Discussion 14 09-01-2009 08:44 AM
DOF and sensor size simons-photography Photographic Technique 60 04-23-2008 10:55 AM



All times are GMT -7. The time now is 07:38 AM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top