Subject size: APS-C vs. FF

[tigerotor77w]

I'm not exactly sure where to begin searching, as this post could go in several different directions. I'm not looking for a discussion of crop vs. FF merits, but rather a technical discussion of basic composition. I apologize if I'm using some incorrect or imprecise terminology. I'm no expert in these things. :(

If I have two cameras with identical resolution -- say XSi and 5D -- then using the same focal length lens on the two bodies will result in a smaller subject with the 5D. That is, if I shoot a 70-200 at 70 mm on the 5D and on the XSi with identical composition and camera-to-subject distance, the XSi will seem to have a larger subject (this all follows from Skip D's, wilt's, Perry Ge, et al's excellent discussion on perspective and focal length). If the subject is a bird on a tree branch (assume the background is far, far away and infinite, so no other features to discern), then I would need a 70*1.6 mm lens on the 5D to equal the XSi's rendering of the subject. That much makes sense.

Question 1. If I were to crop the 5D image in order to get the bird to be the same size as that rendered by the XSi, by how much am I cropping? 1.6 linearly (i.e. the whole image area would be 2.56 times smaller)?

Question 2. Corollary of Q1: if I wanted a FF *sensor* that would render the subject to be the same size as that rendered by the XSi, I would need 12*1.6*1.6=30.72 MP, correct?

What I'm getting at for this scenario is for, say, BIF or air shows.

Question 3 -- also related to 1 and 2... In a given instant, if I have either an existed 5DII + 70-200 or an XSI + 70-200, I will still get more croppability out of the XSi than the 5DII, even though the 5DII has more pixels than the XSi? The crop factor makes up for the pixel advantage of the 5DII?

Question 4. So if I want more croppability, my options are to continue up the APS-C ladder and *not* the FF ladder?

This, of course, ignores all the DOF, sensor efficiency, and perspective commentary in other various threads here, so...

Q5. If I want a subject to be the same size in two pictures, one taken from the 5D and one from the XSi, by how much would I need to change the aperture of either in order to have the same DOF? (Both cameras at the same location.)

Q6. Suppose now that I deliberately change my perspective of the 5D to match the *subject size* of the XSi (i.e. I move closer to the subject). This will affect DOF, so now what happens to the aperture values? I assume the DOF equation that wilt has described in Daniel Browning's thread about sensors (June 2009) can be used to calculate an exact number, but is there a simple trick (associating distance from subject to aperture, loosely speaking) to get this down pat?

Q7. If I add a 1.6x TC (which I realize doesn't exist) to the 5D, it should match the 35 mm equivalent focal length of the XSi, correct? If that's the case -- and ignoring the fact that the TC will slow down the lens -- will DOF also now be equivalent between the two cameras?

Maybe there's a more concise way to write all this... but you can see how many different "what if" scenarios could be posed...

Merry Christmas in the meantime!

[tzalman]

#1 - Correct.
#2 - Correct.
#3 and 4 - The 5DII has the same pixel density as the 30D and XT. However, the answer to, "The crop factor makes up for the pixel advantage of the 5DII?" is not simple. Resolution is only one factor, the amount of enlargement is another. If the 5DII cropped to 8 Mp is like a 30D, the 7D cropped is like a high resolution P&S, and as the ratio print : digital image increases, quality decreases.
Q5. If I want a subject to be the same size in two pictures, one taken from the 5D and one from the XSi, by how much would I need to change the aperture of either in order to have the same DOF? (Both cameras at the same location.)

Contradiction.
#6 Go to Dofmaster.com and have a play with different models, focal lengths and apertures.

[tigerotor77w]

#3 and 4 - The 5DII has the same pixel density as the 30D and XT. However, the answer to, "The crop factor makes up for the pixel advantage of the 5DII?" is not simple. Resolution is only one factor, the amount of enlargement is another. If the 5DII cropped to 8 Mp is like a 30D, the 7D cropped is like a high resolution P&S, and as the ratio print : digital image increases, quality decreases.

Contradiction.
#6 Go to Dofmaster.com and have a play with different models, focal lengths and apertures.

For 3 and 4 -- let's say ignoring IQ degradation for my original question, but for the sake of discussion... what you're saying is if I crop down to a 600x400 section on an XSi and on the 5DII, the XSi will appear worse in a given print size?

For 5: I meant after cropping in post. The two subjects will be different-sized, but after cropping down.

For 6 -- makes sense. :)

[SkipD]

If I have two cameras with identical resolution -- say XSi and 5D -- then using the same focal length lens on the two bodies will result in a smaller subject with the 5D. That is, if I shoot a 70-200 at 70 mm on the 5D and on the XSi with identical composition and camera-to-subject distance, the XSi will seem to have a larger subject (this all follows from Skip D's, wilt's, Perry Ge, et al's excellent discussion on perspective and focal length).There is a VERY simple solution to the problem you describe. DO NOT attempt to use the same focal length on different format cameras for the same framing. It will not work properly.

If you can properly frame something (from the best camera-to-subject distance for the perspective) with a 100mm lens on an XSi camera, then you would simply use a 160mm lens (or a zoom dialed in to 160mm) to get the same framing with a 5D.

The two cameras are different format cameras, and each requires a different focal length for the same framing from the same distance. There's no getting around this simple fact.

[toxic]

Q5. If I want a subject to be the same size in two pictures, one taken from the 5D and one from the XSi, by how much would I need to change the aperture of either in order to have the same DOF? (Both cameras at the same location.)

Q6. Suppose now that I deliberately change my perspective of the 5D to match the *subject size* of the XSi (i.e. I move closer to the subject). This will affect DOF, so now what happens to the aperture values? I assume the DOF equation that wilt has described in Daniel Browning's thread about sensors (June 2009) can be used to calculate an exact number, but is there a simple trick (associating distance from subject to aperture, loosely speaking) to get this down pat?

35mm has about 1 1/3 stops less DoF for the same framing and f-stop. It is the same for both moving the camera and using a longer lens. Basically, it loses 1.6 (or whatever crop factor) stops of DoF but gains some DoF back because it needs less enlargement (less enlargement = more DoF).

Q7. If I add a 1.6x TC (which I realize doesn't exist) to the 5D, it should match the 35 mm equivalent focal length of the XSi, correct? If that's the case -- and ignoring the fact that the TC will slow down the lens -- will DOF also now be equivalent between the two cameras?

A lens + 1.6x on a 5D would be the same AoV as the same lens on an XSi. The DoF would not be the same, for the same reasons why any larger format yields smaller DoF - you are either increasing the focal length or decreasing subject distance to achieve the same subject size. TCs increase the focal length, while different sensor sizes change the AoV and enlargement factor.

DoF for two different formats is the same if you use the same settings at the same distance and crop the larger format to match the smaller.

[tdodd]

Question 4. So if I want more croppability, my options are to continue up the APS-C ladder and *not* the FF ladder.
In my opinion this topic is a lot more involved than just counting how many pixels you have on your subject. I find it much better to think in terms of maximum acceptable enlargement factors rather than pixels.

For example, consider the days of 35mm film. You would have an image recorded on your film of 36mm x 24mm. That's about 1.5"x1". A typical "large" print would be 12"x8" for viewing from close up. That is an enlargement factor of 8X, which is really quite modest. That is a print size for which the DOF tables we commonly use will tend to work quite well. It is also a print size for which rules of thumb for minimum shutter speeds to prevent visible shake also work quite well. Of course, you could easily make larger prints, but not for scrutiny from 12-18" away.

Once you start making enlargement greater than 8X from captured image to final display you begin to bend/break those rules. Your DOF will not be as great as you thought, if you are still going to review the images from 12-18" away. You may more easily see the effects of camera shake. A lens that was sharp with an enlargment of 8X may tend to look soft at greater enlargement sizes. Then there is the matter of noise as well. People these days will routinely use ISOs that would have the old school film guys shaking their heads in horror.

So, still while ignoring pixels for the moment, let's agree that a reasonable enlargement factor for a high quality image is 8X. If you want to enlarge more than that you need a good (great) lens, with accurate focusing, supported correctly, with adequate shutter speed and you can't go mad with your ISO unless the camera really has the performance you need.

But if a 35mm camera with an enlargement factor of 8X produces a 12x8 print, what can we expect from an APS-C camera? Well, for a start, we should not assume we can easily enlarge the image from that camera by more than 8X either. Well an APS-C camera has a sensor of only ~22mm x 15mm, which is less than an inch on the long side. If we were to fill the frame of a 7D or XSi and then produce a print with a magnification of 8X we would have a print of approx 7"x5".

Now, I'd bet that if you viewed any of your images at a physical size of around 7x5 on your monitor then you would be pretty happy with the IQ. But people don't. Firstly, they tend to display to fill the screen at least, which might mean displaying at a physical size of something like 15"x10" or so. That's an enlargement factor of 17X, which puts greater demands on camera, lens and photographer when still viewing from just 12-18" away.

But even that is not enough for some people. They want to go larger, by viewing at 50% or 100%, which will probably end up as an enlargement factor of up to 40X or so, yielding a virtual image of 36" or more on the long side. Yet still they want to review the image from just 12-18" away. This is unlikely to result in complete happiness. That level of magnification means there is nowhere to hide. The tiniest failure in focus/shake/blur/inadequate DOF, noise is going to be revealed. Unless you are shooting with a prime, in great light, with low ISO, and perfect focus, from a stable tripod, prepare for some disappointment.

So, back to the original question - The intrinsic IQ of an image does not come from the number of pixels alone. There are many more factors involved that all need to be sorted. Otherwise your power to crop will be limited. If you are shooting BIF with a 400mm lens on a 5D2 or a 400mm lens on a 7D, the subject itself will be exactly the same physical size on the sensor. A 400mm lens really does magnify like a 400mm lens. It does not matter which body you fit it to. So let's assume our subject is small in the frame and, with a sympathetic crop to improve composition you end up with a useable area of your sensor that amounts to only 12mm x 8mm. That's about 1/2" on the long side. If we stick to our 8X guideline for enlargement for viewing from 12-18" away then we should expect to make a nice 4"x2.7" print. Of course, viewed on a monitor you would still only want to display the image at a physical size of 4"x2.7". If you tried to view the 5D2 and 7D files, cropped to 12mm x 8mm, at 100% then you would be enlarging both files way beyond what one should reasonably expect, and the 7D file would be significantly more tortured than the 5D2 file.

Now the truth is that with modern equipment people are able to get good results at enlargements above 8X, but let's not go nuts and expect the impossible. I'm not sure the number of pixels matters that much - it's how much you enlarge that matters and how good your original capture was (blur/shake/misfocus/noise etc.). So long as you have sufficient pixels then that is enough. You do not need more than enough.

For example, my own experience when shooting BIF with my 100-400 lens and my 1D3 and 50D bodies is that I always had just as much cropping power with the 1D3 as with the 50D. It was fairly easy to use the 1D3 files at 100%, with those big, fat, juicy pixels, relatively few in number. The pixel sharpness was good. The pixel noise was good. However, due to shake/blur/noise/focus error etc. it was not possible (usually) to use the 50D files at 100%, not when shooting BIF. By the time the 50D files were sized to a level where the IQ was acceptable I was right back to where I already was with the 1D3. In other words, the extra pixel density of the 50D bought me no more cropability - for BIF. I could get the same level of enlagement from both cameras. I could not get more from the 50D.

However, when I would shoot perched birds in good light, with low ISO, and using a solid tripod then I was able to make those extra pixels count. The 50D did comfortably outgun the 1D3 when the conditions were right. But as soon as you start shooting action, or the light dims, the gap closes. The problem is not the camera, the lens, the light, or even me. It's just the practical constraints of real world physics.

Sorry for the waffle. I hope that makes sense.

[tigerotor77w]

Oh boy; this is going to be a long one. Thanks for everyone's replies -- it's already given me some good insight.

There is a VERY simple solution to the problem you describe. DO NOT attempt to use the same focal length on different format cameras for the same framing. It will not work properly.

If you can properly frame something (from the best camera-to-subject distance for the perspective) with a 100mm lens on an XSi camera, then you would simply use a 160mm lens (or a zoom dialed in to 160mm) to get the same framing with a 5D.

The two cameras are different format cameras, and each requires a different focal length for the same framing from the same distance. There's no getting around this simple fact.

Agree with your post -- the reason I asked, though, if I have the 70-200 and want more croppability with a FF body, this means that my own choice is to buy a longer lens -- simply having more pixels to crop from won't get me the framing I'd want (besides the point that no Canon body has 12*1.6*1.6 MP right now).

35mm has about 1 1/3 stops less DoF for the same framing and f-stop. It is the same for both moving the camera and using a longer lens. Basically, it loses 1.6 (or whatever crop factor) stops of DoF but gains some DoF back because it needs less enlargement (less enlargement = more DoF).

A lens + 1.6x on a 5D would be the same AoV as the same lens on an XSi. The DoF would not be the same, for the same reasons why any larger format yields smaller DoF - you are either increasing the focal length or decreasing subject distance to achieve the same subject size. TCs increase the focal length, while different sensor sizes change the AoV and enlargement factor.

DoF for two different formats is the same if you use the same settings at the same distance and crop the larger format to match the smaller.

I'll have to think about your post... I think it makes sense, but give me some time...2

In my opinion this topic is a lot more involved than just counting how many pixels you have on your subject. I find it much better to think in terms of maximum acceptable enlargement factors rather than pixels.

For example, consider the days of 35mm film. You would have an image recorded on your film of 36mm x 24mm. That's about 1.5"x1". A typical "large" print would be 12"x8" for viewing from close up. That is an enlargement factor of 8X, which is really quite modest. That is a print size for which the DOF tables we commonly use will tend to work quite well. It is also a print size for which rules of thumb for minimum shutter speeds to prevent visible shake also work quite well. Of course, you could easily make larger prints, but not for scrutiny from 12-18" away.

Once you start making enlargement greater than 8X from captured image to final display you begin to bend/break those rules. Your DOF will not be as great as you thought, if you are still going to review the images from 12-18" away. You may more easily see the effects of camera shake. A lens that was sharp with an enlargment of 8X may tend to look soft at greater enlargement sizes. Then there is the matter of noise as well. People these days will routinely use ISOs that would have the old school film guys shaking their heads in horror.

So, still while ignoring pixels for the moment, let's agree that a reasonable enlargement factor for a high quality image is 8X. If you want to enlarge more than that you need a good (great) lens, with accurate focusing, supported correctly, with adequate shutter speed and you can't go mad with your ISO unless the camera really has the performance you need.

But if a 35mm camera with an enlargement factor of 8X produces a 12x8 print, what can we expect from an APS-C camera? Well, for a start, we should not assume we can easily enlarge the image from that camera by more than 8X either. Well an APS-C camera has a sensor of only ~22mm x 15mm, which is less than an inch on the long side. If we were to fill the frame of a 7D or XSi and then produce a print with a magnification of 8X we would have a print of approx 7"x5".

Now, I'd bet that if you viewed any of your images at a physical size of around 7x5 on your monitor then you would be pretty happy with the IQ. But people don't. Firstly, they tend to display to fill the screen at least, which might mean displaying at a physical size of something like 15"x10" or so. That's an enlargement factor of 17X, which puts greater demands on camera, lens and photographer when still viewing from just 12-18" away.

But even that is not enough for some people. They want to go larger, by viewing at 50% or 100%, which will probably end up as an enlargement factor of up to 40X or so, yielding a virtual image of 36" or more on the long side. Yet still they want to review the image from just 12-18" away. This is unlikely to result in complete happiness. That level of magnification means there is nowhere to hide. The tiniest failure in focus/shake/blur/inadequate DOF, noise is going to be revealed. Unless you are shooting with a prime, in great light, with low ISO, and perfect focus, from a stable tripod, prepare for some disappointment.

So, back to the original question - The intrinsic IQ of an image does not come from the number of pixels alone. There are many more factors involved that all need to be sorted. Otherwise your power to crop will be limited. If you are shooting BIF with a 400mm lens on a 5D2 or a 400mm lens on a 7D, the subject itself will be exactly the same physical size on the sensor. A 400mm lens really does magnify like a 400mm lens. It does not matter which body you fit it to. So let's assume our subject is small in the frame and, with a sympathetic crop to improve composition you end up with a useable area of your sensor that amounts to only 12mm x 8mm. That's about 1/2" on the long side. If we stick to our 8X guideline for enlargement for viewing from 12-18" away then we should expect to make a nice 4"x2.7" print. Of course, viewed on a monitor you would still only want to display the image at a physical size of 4"x2.7". If you tried to view the 5D2 and 7D files, cropped to 12mm x 8mm, at 100% then you would be enlarging both files way beyond what one should reasonably expect, and the 7D file would be significantly more tortured than the 5D2 file.

Now the truth is that with modern equipment people are able to get good results at enlargements above 8X, but let's not go nuts and expect the impossible. I'm not sure the number of pixels matters that much - it's how much you enlarge that matters and how good your original capture was (blur/shake/misfocus/noise etc.). So long as you have sufficient pixels then that is enough. You do not need more than enough.

For example, my own experience when shooting BIF with my 100-400 lens and my 1D3 and 50D bodies is that I always had just as much cropping power with the 1D3 as with the 50D. It was fairly easy to use the 1D3 files at 100%, with those big, fat, juicy pixels, relatively few in number. The pixel sharpness was good. The pixel noise was good. However, due to shake/blur/noise/focus error etc. it was not possible (usually) to use the 50D files at 100%, not when shooting BIF. By the time the 50D files were sized to a level where the IQ was acceptable I was right back to where I already was with the 1D3. In other words, the extra pixel density of the 50D bought me no more cropability - for BIF. I could get the same level of enlagement from both cameras. I could not get more from the 50D.

However, when I would shoot perched birds in good light, with low ISO, and using a solid tripod then I was able to make those extra pixels count. The 50D did comfortably outgun the 1D3 when the conditions were right. But as soon as you start shooting action, or the light dims, the gap closes. The problem is not the camera, the lens, the light, or even me. It's just the practical constraints of real world physics.

Sorry for the waffle. I hope that makes sense.

Makes a lot of sense, but I do have a few questions.

1) You said that if I cropped down to an image of 12mm x 8mm on both a 7D and a 5DII that the 7D pixels would be far more tortured -- is this because 12x8 is a bigger percentage of the 7D's sensor than it is of the 5DII's sensor?

2) Ultimately, on a larger-sensor body... the worse the lighting, the bigger the differentiation between it and a smaller-sensor body? If the 50D and 1DIII give you the same level of enlargement, why not just buy the cheaper body (ignoring features and considering only IQ). I realize this is kind of a dumb question. I bought my XSi instead of the 40D (with a much small price difference than 50D -> 1DIII) for this reason, and while I don't regret it, my next decision may be between the 7D and 5DIII or 7D and 3D (hypothetically). If the 7D is capable of the same printable magnification of the 5DIII or 3D, why spend the extra money on a FF body?

And perhaps this is why Daniel Browning's thread came into being.

[tzalman]

If the 50D and 1DIII give you the same level of enlargement, ... If the 7D is capable of the same printable magnification of the 5DIII....
Reread Tim's post. The whole point is that an image from a smaller sensor needs to be enlarged more in order to get the same sized print. An FF needs 8X for an 8x12, an APS-C needs 13X.

[tigerotor77w]

Reread Tim's post. The whole point is that an image from a smaller sensor needs to be enlarged more in order to get the same sized print. An FF needs 8X for an 8x12, an APS-C needs 13X.

I realize the last point... so provided the APS-C body has enough pixels to crop from, then why not choose the smaller, cheaper body, ignoring ergonomics/performance factors?

Tim's precise quote:

In other words, the extra pixel density of the 50D bought me no more cropability - for BIF. I could get the same level of enlagement from both cameras.

This means I could potentially print from a 50D what I could from a 1DIII, no? He also says that in good light, the 50D has more croppability and that it's in low-light or challenging light where the 1D (and by extrapolation the 5DII) shines.

[Edit: I don't want to take this into the realm of merits of FF vs. APS-C... I'll leave these questions as rhetorical. I didn't mean to twist it into the direction I twisted it; it was somewhere in my subconscious. My apologies!]

[tdodd]

If you are cropping then you immediately remove the major advantage to IQ offered by the cameras with larger sensors, so yes, you pull the rug from under the feet of the APS-H and full frame bodies, as far as IQ is concerned.

However, in the times when you are not focal length limited and you can fill the frame then the larger sensors will offer a real improvement in IQ because they can record a larger image, needing less enlargement later on, and also gathering more light. As you said, a full frame camera gathers 2.56X more light than APS-C, which should give it about 1 stop better performance at the image level.

But there is more to getting good images than the sensor alone. If you want to shoot BIF or other fast action then you will want great (customisable) AF, possibly high frame rates and maybe a deep buffer. Ergonomics will be important so that you can quickly adjust the camera without removing it from your eye. A bigger, brighter viewfinder will make the whole task easier and more enjoyable. Better build will mean the camera can handle the knocks of the natural environment and also withstand shooting in the rain. Having a 3 stop meter and a tight spot for metering will also make things much easier, if you shoot using the zone system or some variant of it. Then there are things like shutter lag and viewfinder blackout times to consider and longer battery life, plus a proper battery power meter. I'm sure the list goes on.

So while the IQ will be roughly similar when the shooting is easy and you are cropping, cameras like the 1D3 and 7D will allow you to keep on shooting when the going gets tough. That can be the difference between getting a great shot, an OK shot or no shot at all. All the IQ in the world won't help you if the shot is out of focus or mistimed, or you get shutter failure after 50,000 images.

By the way, the 50D advantage referenced in the earlier post does, in my opinion, not just require good light. It requires in particular a very low ISO - 200 max - and an extremely stable, shake free platform for the camera. Also the lens must have great IQ and the selected aperture must not push too far into diffraction. So the 50D advantage only applies to a few types of photography under optimised conditions - IMHO. A camera like the 1D3 should deliver excellent results in a broad ranging set of shooting scenarios and will not waste storage space and processing time on an excess of noisey/blurry/soft pixels that do not add value to the iQ.

[Radtech1]

Q5. If I want a subject to be the same size in two pictures, one taken from the 5D and one from the XSi, by how much would I need to change the aperture of either in order to have the same DOF? (Both cameras at the same location.)


You do not have to change the aperture at all.

Not one tiny bit.

Assuming the same focal length, the image projected by the lens does not care what size sensor it falls on. It had the same DoF regardless.

It is an often chanted mantra that APS-C sensors have a different DOF than FF, and they often cite DOFMaster, because it's a website, and well, that makes them right. But rather than blind belief, I encourage you to think it through.

Then lens on the camera does works just the same way the lens of a movie or slide projector. Imagine then that you are projecting an image. You have a large screen, big enough to accommodate the entire image projected. That is analogous to the FF sensor.

Then, you place a smaller screen. Some of the image falls off of the smaller screen - that is your "crop factor". BUT, does the act of placing a smaller screen change the quality of the image, including the DOF of that image, coming from the lens. Hell to the power of NO! Likewise, the same lens, at the same focal length and f/stop will project the same image (DoF and all) regardless of the size screen - oops, sensor - that is there to catch the image.

Rad

BTW, I checked this for myself by shooting with a 5D (FF) and a Rebel (1.6) and compared the focus and focus falloff. Not surprisingly, the results were grossly the same camera to camera.* And, also not surprisingly, many people howled that I was wrong. If interested, check here: http://photography-on-the.net/forum/showthread.php?t=472572

*Absolute resolution of the two different cameras gave the 5D a slightly softer look all the way through, but that is a different issue than the "sensor size/DoF question".

[Wilt]

You do not have to change the aperture at all.

Not one tiny bit.

Assuming the same focal length, the image projected by the lens does not care what size sensor it falls on. It had the same DoF regardless.
Your oversight is that while the image at the focal plane from the same lens is identical, when you enlarge the APS-C 15x22.3 image to final size (like 8x10) and you enlarge the FF 24x36mm frame to same final size, the APS-C image is magnified by 1.6x greater amount and THAT enlarges the COC so that the eye perceive the out of focus discs better...decreasing the DOF!

[tigerotor77w]

If you are cropping then you immediately remove the major advantage to IQ offered by the cameras with larger sensors, so yes, you pull the rug from under the feet of the APS-H and full frame bodies, as far as IQ is concerned.

By the way, the 50D advantage referenced in the earlier post does, in my opinion, not just require good light.

Thanks. This makes sense.

[I][B]You do not have to change the aperture at all.

Gonna have a look at your other thread.

Your oversight is that while the image at the focal plane from the same lens is identical, when you enlarge the APS-C 15x22.3 image to final size (like 8x10) and you enlarge the FF 24x36mm frame to same final size, the APS-C image is magnified by 1.6x greater amount and THAT enlarges the COC so that the eye perceive the out of focus discs better...decreasing the DOF!

So... regardless of whether I'm looking at an 8x12 or my screen, the CoC changes and hence the DOF changes as well?

[Wilt]

Thanks. So... regardless of whether I'm looking at an 8x12 or my screen, the CoC changes and hence the DOF changes as well?


DOF calculations assume that one is producing an 8x10 print viewed from a certain distance. Whenever you change the viewing distance or the size of the image from that assumption, the DOF calculations are invalid. A few DOF calculators permit you to choose the image size to be viewed, the distance from which you are viewing it, and the level of visual acuity (Human vision is actually MORE precise than what lens manufacturers assume, so what they state as 'acceptably small CofC' likely is NOT truly well in focus to you when viewed at the size and distance in the assumptions.)

[tigerotor77w]

Ok. It's time to do some more reading. :)

[Wilt]

The key principle that is often lost about DOF is the fact that is all based upon the size of the out-of-focus disks perceivable by the human eye. And the size of the out-of-focus disks perceivable changes when the size of the print changes because the disks are more greatly magnified, when the viewing distance to the print changes, and when one person has higher visual acuity than another. If that key principle is remembered, then the 'mystery' of DOF dissolves, and one easily understands why different format cameras have different CofC sizes, and why one has to base DOF calculations on what final print size and distance of the viewer will be.