DoF question

rubo · March 8, 2012

Hi all,

my math is a little rusty on the subject, so i need your help.

Right now i have a 60D.

When i'm shooting landscapes i mostly keep the lenses @ f/8-11, more than that diffraction blur tends to become obvious.

So here is my question:

After i get the 5d Mkiii, what would be the right f/stop to get the same DoF i'm getting right now lets say @ f/8 on a 60D, assuming the framing is the same , so instead of 30mm on a 60D (which is 48mm equivalent on FF) it would be 50mm?

Thanks for any help.

robin_sibson1 · March 8, 2012

To produce images that are indistinguishable after being scaled to the same size (say, a 4" by 6" print, or for that matter a 40" by 60" print), the recipe is as follows as between a FF body and a 1.6-factor crop body. To move from 1.6-factor to FF you need (a) a lens with a focal length longer by a factor of 1.6, and (b) to use an aperture narrower (that is, numericlly larger) by a factor of 1.6, which in practice means one-and-one-third stops. For example, 50mm at f/4 on 1.6-factor corresponds to 80mm at f/6.3 on FF. This assumes you are shooting at distances much larger than the focal lengths involved. Also, to say that the images are indistinguishable simply means that the angle of view and depth of field are the same. The quality of the out-of-focus blur (bokeh) may be different because you are using different lenses (or the same lens at different focal lengths), and the resolution and other aspects of sensor behaviour may vary (but of course that can happen with different cameras of the same format as well).

JosvanEekelen · March 8, 2012

Regarding diffraction, see this link <a href="http://www.the-digital-picture.com/Reviews/Canon-EOS-Rebel-T3i-600D-Digital-SLR-Camera-Review.aspx">http://www.the-digital-picture.com/Reviews/Canon-EOS-Rebel-T3i-600D-Digital-SLR-Camera-Review.aspx</a> for a table of several EOS bodies and their approximate diffraction limit. 5D3 is not listed yet but it will be in the same range as 5D2 and 1DX, about f/11.

tudor_apmadoc · March 8, 2012

Some of this is going to depend on the specific lens. dpreview has one of the best test report mechanisms i've seen. This is an example for a 70-200 2.8 II lens. They don't have all the Canon Lenses covered, but should give you some ideas

http://www.dpreview.com/lensreviews/canon_70-200_2p8_is_usm_ii_c16/5

You can look at APC/ Full Frame, and really see how f stop changes things

g dan mitchell · March 8, 2012

About "diffraction limit" - be cautious about what you accept regarding this value. You'll often hear people tossing around the term "diffraction limited aperture." Some mistakenly take this to me that aperture beyond which you'll have a diffraction blur problem as you stop down. This is a false interpretation of the term, for reasons I won't go into right now. (Unless someone really wants to know...)

I bring this up because I read above a mention of f/11 as a "diffraction limit" for the 5D3, 5D2, and 1Dx. I shoot the 5D2 and I am quite critical about image resolution, and you have little at all to fear from shooting at f/16. You most certainly would not be able to detect any difference in sharpness in, say, 13" x 19" prints made at f/11 and f/16, and in some cases the overall effect of the smaller aperture might be to create an image that seemed slightly sharper because objects near the edges of the DOF zone might be slightly sharper.

Be especially suspicious (as in "disbelieve completely") anyone who suggests, for example, that a 12MP FF body has "less diffraction" at some aperture than a 22MP FF body, or that you'll have a diffraction problem sooner as you stop down on a higher MP camera. This is completely wrong.

Dan

scott_ferris · March 8, 2012

Rubo,

The<a href="http://www.josephjamesphotography.com/equivalence/"> theory of equivalence</a> says that for same generation sensors,

<ul>

<li>1.5 crop camera, 50mm, f2 at 100iso equals</li>

<li>FF camera, 85mm, f2.8 at 200 iso.</li>

</ul>

This works pretty well for same generation sensors, but your 5D MkIII will probably be more than one stop better for noise than your 60D, and the rounding up figures should give the ff camera another 2/3 stop latitude because Canon's are 1.6 crops and the above rule of thumb is for a 1.5 crop factor Nikon.

In practice I suspect you will get much better images, but very similar framing, dof etc, from the 5D MkIII by closing down one stop, upping the iso one stop if you need to, and using the 1.6 focal length calculation.

JosvanEekelen · March 8, 2012

@Dan: the word "limit" is indeed a bit misleading since smaller apertures don't ruin an otherwise good picture. It's only the f/stop where diffraction starts to kick in at the pixel level, see the abovementioned link. Personally I find the dust specs more reason to avoid smaller openings but in the end the aperture is just one of many variables that lead to a picture.

arie_vandervelden1 · March 8, 2012

If you like f/8 to f/11 on crop, try f/11 to f/16 on full-frame gear.

wade_keenon1 · March 8, 2012

Not to hijack the thread, but can someone elaborate on the "diffraction limit" issue, as I'm confused about how much images degrade (if at all) at apertures past the limit.

I shoot a 40D, which apparently has a DLA of f/9.1, although I routinely use f/11 and f/16 with (to me) acceptable results. Still, I've been reluctant to consider an upgrade to a 7d, which has a DLA of f/6.9, thinking that it might produce "blurrier" images at smaller apertures. Does it?

Sorry, if this question if veering too much from the OP's issue.

john_wheeler6 · March 8, 2012

Hi Wade

I don't think you question is off track since diffraction was part of the question. Here is a link to an article that does a decent job and includes an online calculator

http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm

I think part of the confusion stems from two separate aperture limits are talked about with diffraction and get mixed up in many conversations. More on that below.

First - the size of the "Airy Disk" or diffraction blur on the film/sensor is only a function of the Aperture of the lens. It does not matter the size of the pixels, focal length, sensor size etc. You can try this with the online calculator.

There are two directions on when to consider what is the aperture limit for diffraction

1) You want to take an image where the diffraction blur in the image data is below the sensors pixel size so that diffraction blur will never come in to play no matter what size enlargement print you make in the future. This is the more restrictive of the two ways to calculate the Aperture limit and you do so by setting the checkbox in the calculator to have the circle of confusion set to the sensor pixel size.

2) You want the diffraction blur to be smaller than the acceptable blur as set by the Circle of Confusion (CoC) used in Depth of Field calculations. The standard value used in Full Frame sensors is .03mm and for 1.6X crop sensors .019mm is used. Both of those values are based on viewing an 8x10 image at a distance of 10 inches (25cm) for the average viewer. This is what the rest of the online calculator covers, figuring out the CoC for a given size print, at a given viewing distance, starting from a given sensor size (determines enlargement required). Under those circumstances, the Diffraction blur is typically not as large and the Aperture can be stopped down further without producing visible diffraction blur. Note: The .03mm CoC used as standard for a full frame sensor is the same as 30 microns. This is much larger than the typical sensor pixel size. This means that for smaller print sizes, you can tolerate more diffraction blur and you will never see it.

Hope this is helpful

robertbody · March 8, 2012

Same sized cropped sensor, higher pixel density going from 40D to 7D would affect the DLA. Bigger sensor is a better upgrade when addressing diffraction and wanting to use smaller apertures.

g dan mitchell · March 8, 2012

For those concerned about something called "diffraction limit" or a system described as being "diffraction limited" at some aperture, you might step away from that for a moment and think about the following:

Take two full frame cameras, one with a 12MP sensor and one with a 21MP sensor. Put the same lens on each, set it to the same aperture, and photograph the exact same scene. Now make two prints at the same size from both captures.

Both images will have precisely the same "amount" of diffraction. This is true if you shoot both lenses at any aperture.

Dan

(The diffraction blur, to the extent that it exists, will be rendered more accurately by the 21MP camera, but that isn't important here and it is irrelevant to question of "how much" diffraction there is.)

michael_young3 · March 8, 2012

To answer the OP question first and simply: DoF has an inverse linear relationship with focal length. Period. The definition is precise and exact. Doubling the focal length halves the DoF. Changing focal length by the crop factor changes the DoF by the same factor. To expand on this a little, hopefully without confusing or confounding the issue, DoF relates to the *size* of the aperture opening, measured in length units: inches or mm as is your custom. This holds true at every focal length. Confusion here arises because aperture is marked and almost universally expressed as a ratio of the opening size to the focal length, the f-stop number f/N. Your zoom lens changes the aperture *size* as you change the focal length, to keep the same selected aperture *ratio*. Conversely, changing the the aperture ratio with focal length to keep the same aperture size results in the same DoF as the other focal length. Moving on to what we mean by DoF and DLA, they are fundamentally the same concept, with the (insignificant in context) difference that one speaks in terms of CoC, while the other of Airy disks. For both, the chosen limit is how big that distance is before a detail is "fuzzy". They come back to what we consider a "pixel". GDan's opinion, equally valid as any other, is that we are concerned only with what the viewer's eye can perceive in a print. The "pixel" for GDan is the eye's ability to discern detail, at whatever viewing distance, in that lighting condition. The other definition of "pixel" is simply the common definition of what we see on the sensor. This is the realm of the pixel peeper. (I am one myself, and often evaluate detail in an image at 1:1 in my editing software. There is no shame or disrespect implied or intended in the term.) These two views are not at all independent. We tie both worlds together every time we divide the cropped image pixel dimensions by (say) 300 dpi to determine a print size. Here is a more moderate view. Considering diffraction effects only, at f/16, my 7D's 18 MP sensor captures at most 8 MP of detail. The consequence is this limits the largest print I can make from that image to that of an 8 MP camera, for example the 20D of 8 years ago. The other consequence is that the math says I need not worry about diffraction at apertures larger than f/6.9 on the 7D. All is well if that suits the image. If not, I am fully conscious of the tradeoff in selecting a smaller aperture. The tradeoff can be expressed as an equivalent sensor size in MP. The problem with DLA is that it not commonly well understood. It is expressed as a single number, and says nothing of the effect at smaller apertures. For the 7D, f/7.1 is my default aperture. At f/11, the fuzzies make it equivalent to about 12 MP. f/16 is about 8 MP. f/22 is 6 MP. This gives a good feel of the trade off, and the print size that can be produced.

rubo · March 8, 2012

Thanks, everyone.

I know full well how the "diffraction limit" of f/6.9 for 60D does not mean "don't shoot" at smaller f-stops.

All that is happening the sharpness gradualy decreases after that and i for myself find the loss in sharpness after f/11 to be tad too much of a trade of.

I'm not saying the images are useless after that, just not good enough for me. i have shot past f/11 if i want to have a certain look or effect or feel or what ever you want to call it, but it's not something i will do if i can avoid it :-)

Scott,

that was very interesting, i had not come accross the article before. Now my head is even more messed up :-)

Thanks again, everyone :-)

g dan mitchell · March 8, 2012

Someone attempted to interpret what I wrote thusly:

"GDan's opinion, equally valid as any other, is that we are concerned only with what the viewer's eye can perceive in a print. The "pixel" for GDan is the eye's ability to discern detail, at whatever viewing distance, in that lighting condition."

</blockquote>

Well, yes. But also, no. Mostly "no" in the context of what I wrote here.

In entirely objective and measurable terms, what I wrote is correct. This isn't about it "being different but so little you cannot see it so it doesn't matter." It isn't my "opinion" about resolution or diffraction blur or anything else. It is literally a verifiable fact that the amount of diffraction in the two examples I describes is precisely the same.

Diffraction is completely unrelated to pixel density. Diffraction is an optical phenomenon that affects the light that falls on the sensor or the film. The same "diffraction blur" falls on the 12MP sensor as on the 21MP sensor. Imagine a hypothetical situation in which the combination of lens, aperture, focal length produced diffraction blur that was ".1mm across." (To make up a simplified way of describing this.)

It is .1mm across whether it falls on a 1GB, 12GB, 22GB, or 100GB sensor. Its size has not changed at all. It becomes no larger or smaller when it strikes the different sensors. If we have an image area that is, say, 36mm across, the diffraction blur is still 1/360 of the image width in all cases.

This isn't really even open to debate.

Dan

john_wheeler6 · March 8, 2012

Hi Michael and thanks for your post. I followed most of it (I think) yet in regards to your statement below:

DoF has an inverse linear relationship with focal length. Period. The definition is precise and exact. Doubling the focal length halves the DoF. Changing focal length by the crop factor changes the DoF by the same factor.

</blockquote>

All the DOF equations that I have seen and results from online calculators don't match your statements. I may have misinterpretted what you stated. I am willing to learn. Here are the references that I have used for

-DOF calculator: http://dofmaster.com/dofjs.html

- Detailed equations: http://dofmaster.com/equations.html

- Approximate equations: http://dofmaster.com/equations2.html

BTW - the detailed equations in the above reference exactly match the equations from the 1950s seminal book on the topic "Photographic Optics" by Allen R. Greenleaf

Thanks in advance if you have the time for any additional information.

</blockquote>

rubo · March 8, 2012

G Dan,

But isn't the "diffraction blur" more pronounced the higher the pixel density of the sensor (i.e. the higher the resolution)?

So, for a 12MP 5D you would start to notice the effects of the "diffraction blur" @ f/13, while for a 21/22MP 5D Mkii/5D Mkiii it would be around f/10.

Or am i missing something?

I know for a fact, a picture taken with 60D is visualy softer (viewing at 100% magniffication on a computer screen) @ f/16 compared to the same picture @ f/8, does not matter which lens i use.

rubo · March 8, 2012

John,

About the first link, while the main calculations seem to be correct, i'm having a hard time understanding the "hyperfocal distance" and the recomended focusing distance.

How are you suppose to use that in real life?

scott_ferris · March 8, 2012

Rubo,

No the diffraction is no worse when using a higher density sensor, it is just resolved more accurately. A diffracted blur of a given size might cover four pixels (low density sensor) or sixteen pixels (high density pixel), the airy disc is the same size, so, the diffraction is no worse. The reason the higher density sensor is said to be worse is because it can resolve smaller airy discs, so wider apertures, but at any given aperture the diffracted blur is the same size.

Put another way, the higher density pixel pitch is able to resolve smaller airy discs, so is considered to "see" diffracted blur sooner, but that is good, not bad.

Exactly as G Dan says, for any given enlargement from any density sensor the diffraction is exactly the same.

g dan mitchell · March 8, 2012

Rubo, it is easy to become confused about this as there is a lot of misinformation floating around on this topic. The amount of diffraction blur at, say, f/13 is exactly the same regardless of photo site density (i.e. - "number of megapixels) on cameras with the same sensor size.

Here's the bottom line. When it comes to resolution, there is no bad news from using a sensor with more photo sites. In regards to diffraction blur, in the worst case the amount of diffraction blur is precisely the same with the higher photo site sensor. In the best case - and not related to diffraction - the higher density sensor may be able to produce a very slightly better image at larger apertures if you use a very good lens.

Dan

john_wheeler6 · March 8, 2012

Hi Rubo

First I would like to say great advice from Scott and Dan.

As far as your question:

John, About the first link, while the main calculations seem to be correct, i'm having a hard time understanding the "hyperfocal distance" and the recomended focusing distance. How are you suppose to use that in real life?

</blockquote>

I may not totally understand your question yet here is my answer based on what I think your are asking.

For a given sensor size, the Hyperfocal distance is determined by both the lens focal length and the Aperture setting. If you set you focus point to the Hyperfocal distance with said lens and said Aperture, your depth of field will be from 1/2 the hyperfocal distance to infinity.

If you have an image in which you want to have very far distances in focus (assume infinity) and as much as possible near you in focus, then you should set the focus distance to the Hyperfocal distance.

You can test this quite easily with the online calculator. Select your camera, lens, and Aperture. That results in a displayed Hyperfocal distance. Plug the Hyperfocal distance into the subject distance input and it will result in the near DOF point being 1/2 the Hyperfocal distance and the far DOF point to be exactly infinity. That's its use. If I did not understand your question just ask again. Hope this helps.

michael_young3 · March 8, 2012

John, I can't answer in specifics about your links and calculators. I would instead point you in a slightly different direction, to Harold Merklinger's writing on focusing and depth of field. A good starting point is here: http://www.trenholm.org/hmmerk/HMArtls.html.

More generally, the OP question of duplicating DoF with a different focal length can be answered without online calculators. Just shoot it with your equipment as demonstration of the proof. I asserted that the relationship is simply linear with focal length. This comes from the definition of aperture f-numbers, expressed as f/N, or focal length divided by aperture size. So long as we keep the aperture *size* the same, the DoF will be the same.

A numeric example will make this more clear. f/2 at 50mm is nominally (50 / 2 = 25mm) aperture size. With a 100mm lens, f/4 gives the same aperture size (100mm / 4 = 25mm). Mount the camera on a tripod; turn off the autofocus. Shoot one shot at 50mm at f/2, and a second shot 100mm at f/4. Examine the two shots and draw your own conclusions.

It didn't require heavy math and calculating CoC size or hyperfocal distance.

The remainder of the discussion is about the fuzziness of detail, whether it is DoF and CoC related, or diffraction and Airy disk related. The result of either is to "fuzz" the detail across more than one pixel. If the "fuzz" is small enough to not scatter beyond one pixel on the sensor, or too small to be visible in a print, we don't care. Our tolerance for "fuzz" is actually much more generous than that. But there is a limit that most of us will agree is "not in focus" or diffraction softened.

In any case, I'll stand by my earlier statements, and brush off GDan's bland accusations of ignorance in everyone but himself. Diffraction effectively limits the detail that can be captured. If the diffraction limits the detail that can be captured on a 12 MP camera to equivalently 6 MP, the usable print size is effectively the same as that captured by a 6 MP camera. This is the diffraction limiting that he denies exists.

The bottom line for me is to not fret about hyperfocal distance. I don't know at the point of clicking the shutter how large I will print. I don't know that the framing I chose when shooting is how I will choose to crop. Not knowing these things, precise calculations of CoC referenced to print detail, such as you'll find in online calculators, are no help at all. One extreme of "giving up" this guessing game is my strong preference for TS-E lenses. If it suits the subject and I have a need for crisp, sharp detail, tilt is the answer, not precisely calculated fiction and guesses about what is "close enough".

Beyond this, you might find Merklinger's writing helpful and meaningful. Among his brilliant insights is a revelation that for objects closer than the focus point, details larger than the aperture opening will be rendered in recognizable sharpness. He also usefully relates image detail size to distance beyond the focus point.

rubo · March 9, 2012

Scott and G Dan,

we are esentialy saying the same thing, just because of my english (it not being my native language) we are not CLEARLY understanding each other :-)

John,

i understand the theory, but how would i do it in practice?

on my 24-105mm the longest distance marked is 20ft then a little space |___ and the infinity sign and the lens has room past the sign also.

So how would i set the focus for lets say 470ft?

Michael,

not sure if i agree with the amout of detail lost after "diffracion blur" becomes apparent, but i do agree it starts to blur the shot significantly & it looks "fuzzy".

Thank again, every day i learn something new :-)

michael_young3 · March 9, 2012

not sure if i agree with the amout of detail lost after "diffracion blur" becomes apparent, but i do agree it starts to blur the shot significantly & it looks "fuzzy".

</blockquote>

Rubo, that's a good start for a conversation. :) The diffraction softened image contains less detail -- that is to say, has less information -- than if diffraction were not a factor. The questions remaining are how much less, and how best to express it meaningfully.

john_wheeler6 · March 9, 2012

i understand the theory, but how would i do it in practice? on my 24-105mm the longest distance marked is 20ft then a little space |___ and the infinity sign and the lens has room past the sign also. So how would i set the focus for lets say 470ft?

</blockquote>

Hi Rubo

Good question especially now that many lenses these days provide less and less detail on the focus scale and less and less information on a DOF scale. Others should jump in with their approach. The approach I learned is based on lenses from decades ago (mostly manual) that included a lot more distance detail and a DOF scale as per this image linked from Wikipedia:

http://upload.wikimedia.org/wikipedia/commons/e/e4/DOF_scale_detail.png

As opposed to setting the focus point at the Hyperfocal distance, what I did is set my Aperture and then use the DOF scale. The DOF scale has two marks on it for each Aperture (one closer to infinity and one closer to zero). I lined up the DOF mark that was closer to infinity on the infinity mark and then the DOF mark closer to zero showed me the near DOF limit. When you take this approach, the focus point is automatically set at the Hyperfocal distance. Very simple way to maximize your DOF with the far end at infinity. Again, with limited information on both distance scale and DOF scales on more recent lenses this approach may not be as useful.

As a side note, the extra space past infinity on good lenses has several purposes

- That gives some maneuvering room for the auto-focus mechanism so it does not bang up against infinity

- The focus point for infra-red image taking is past the normal light infinity point

- For generic lenses it may be just manufacturing slop.

Sign In

DoF question

Recommended Posts

Link to comment

Share on other sites

Top Posters In This Topic

Popular Days

Top Posters In This Topic

Popular Days

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Create an account or sign in to comment

Create an account

Sign in