Need assistance with DOF

[kaelaria]

I took my new 30D and 17-85 to a local botanical gardens, to both try the new gear, and  shoot for the first time, closeup subjects.  It baffled me when I got home to finally take a good look at the results, that most of the shots seemed to be out of focus.  After testing under all kinds of different and similar circumstances today, I finally stumbled on the cause.  The whole frame wasn't out in most cases, I simply had an EXTREMELY shallow DOF.

Doing a little research and utilizing an online DOF calculator it's now apperant to me that what I was doing was essentially limiting myself to a few inches of focus in many of my shots.  I now understand the relationships between focal length, distance to subject and available DOF - something I was more of less oblivious to, since all my past experience had been with more distant subjects, usually landscape.

Below you can see a typical frame of what I was shooting along with a crop of the result you can see in detail.  First I'll explain what I was going for, then explain what I got, and then critically - ask if it's even POSSIBLE to do what I had wanted in the first place.

What I was going for was the whole visible flower in focus, with the background slightly and nicely out.  In this particular shot I was focused on the petals.

What I got was the petals in focus, but it was out even by the ends of the stamen.  

I was about 2 feet from the flower at 85mm (actual).  1/60 F8

How could I, if at all, have produced a relatively full frame flower shot like this size, with the whole flower in focus?  Using the calculator I see I could zoom out, but that obviously doesn't give the composition I wanted.  Is the solution to use a higher zoom and stand further back - or will that simply do the same thing?

[AJSJones]

Looks as though f/8 might have been OK but you focused on the petals.  If you wanted the stamens and petals to be the front and back, respectively, of the areas in focus, you should have focused between them.  Set up the composition (tripod if possible) , find a part of the central structure that represents the right distance to be the midpoint, then shoot with several different apertures.  You should find one with the desired areas in focus and the b/g sufficiently blurred.  Pay attention to the distance from lens to the object relative to the distance of the b/g behind the object (the latter should be larger than the former in your example).  Do the experiment a few times at different distances (by using different focal lengths) and you'll soon develop the requisite intuition  

[kaelaria]

Well, even plugging in F22, it only adds an inch.  What if I wanted for example, everything here in focus?  All the way back to the green.

If it's simply not possible, how in the world do you focus between the two acurately?  I sure can't see that kind of detail through the tiny viewfinder?

[howiesmith]

I took my new 30D and 17-85 to a local botanical gardens, to both try the new gear, and  shoot for the first time, closeup subjects.  It baffled me when I got home to finally take a good look at the results, that most of the shots seemed to be out of focus.  After testing under all kinds of different and similar circumstances today, I finally stumbled on the cause.  The whole frame wasn't out in most cases, I simply had an EXTREMELY shallow DOF.

Doing a little research and utilizing an online DOF calculator it's now apperant to me that what I was doing was essentially limiting myself to a few inches of focus in many of my shots.  I now understand the relationships between focal length, distance to subject and available DOF - something I was more of less oblivious to, since all my past experience had been with more distant subjects, usually landscape.

Below you can see a typical frame of what I was shooting along with a crop of the result you can see in detail.  First I'll explain what I was going for, then explain what I got, and then critically - ask if it's even POSSIBLE to do what I had wanted in the first place.

What I was going for was the whole visible flower in focus, with the background slightly and nicely out.  In this particular shot I was focused on the petals.

What I got was the petals in focus, but it was out even by the ends of the stamen. 

I was about 2 feet from the flower at 85mm (actual).  1/60 F8

How could I, if at all, have produced a relatively full frame flower shot like this size, with the whole flower in focus?  Using the calculator I see I could zoom out, but that obviously doesn't give the composition I wanted.  Is the solution to use a higher zoom and stand further back - or will that simply do the same thing?
[a href=\"index.php?act=findpost&pid=80770\"][{POST_SNAPBACK}][/a]

What is the print size and viewing conditions you are using to design your image?  What CoC are you using for the camera?  Then the print?

You say you used a "calculator."  What assumptions were made for the use of the calculator to be valid?  Did you meet all of them?

[kaelaria]

I plugged in the appropriate selections here: http://www.dofmaster.com/dofjs.html

This particular shot I would probably try 8x10 since it gets fuzzy if I chose to print.  Most of mine are 24x30, although I was not looking for a large print out of these, it was mainly a day of experimentation.  I did not shoot any of these with printing in mind, simply to view on screen.

[howiesmith]

I plugged in the appropriate selections here: http://www.dofmaster.com/dofjs.html

This particular shot I would probably try 8x10 since it gets fuzzy if I chose to print.  Most of mine are 24x30, although I was not looking for a large print out of these, it was mainly a day of experimentation.  I did not shoot any of these with printing in mind, simply to view on screen.
[a href=\"index.php?act=findpost&pid=80779\"][{POST_SNAPBACK}][/a]

It is my opinion that if you intend to view the photos on screen, you need to use that as the print size and viewing condition.  What size image are you looking at?  From ehat distance?  In the case of a back lit image, it may be much brighter than a print and exibit less DoF.

What CoC did you use or was assumed by the calculator for you?  Don't forget, the calculator may assume you are making a 30cm print and viewing it from 50cm or an 8x10 or 12 anf viewing from arm's length.

If you change any of the calculators assumptions, the DoF you get may be differnet than expected.

Also, if you are close compared to the lens focal length, the calculator may neglect the lens focal length addition in the hyperfocal length calculation.  Not a big deal at all for distant shots, but maybe a game breaker for macro.

[kaelaria]

I'm confused - how does view distance from the print effect it's focus?

I chose Canon 30D which uses 0.019

[howiesmith]

I'm confused - how does view distance from the print effect it's focus?

I chose Canon 30D which uses 0.019
[a href=\"index.php?act=findpost&pid=80783\"][{POST_SNAPBACK}][/a]

The viewing distance does not affect focus, just DoF.  They are not the same.

CoC is a user selected number, not a camera variable.  You need to select your own CoC oy follow the calculators assumptions pretty close.

Forgot to ask before, but did you crpo the image from the camera to screen?  Any at all?

[kaelaria]

OK maybe I am misunderstanding your terms then - 'viewing distance' - do you mean how far I am to my screen or a print?

If you look on that page, the drop down box for coc has many presets including the 30D.  When you select the 30D it uses 0.019 for the calculation, I am not using it as a camera variable.

No, the Full image is not cropped at all.

[BernardLanguillier]

Well, even plugging in F22, it only adds an inch.  What if I wanted for example, everything here in focus?  All the way back to the green.

If it's simply not possible, how in the world do you focus between the two acurately?  I sure can't see that kind of detail through the tiny viewfinder?
[a href=\"index.php?act=findpost&pid=80776\"][{POST_SNAPBACK}][/a]

It's probably simply impossible.

The best method to focus between objects is to start by focusing on the far one (autofocus for instance), and then manually focus down to the close one.

Picking the middle point can either be done visually (but the viewfinder of the 30D is indeed small), or by looking at the distance scale on the lens barrel.

Cheers,
Bernard

[kaelaria]

It sounds like it's just time for more practice and experimentation!

[howiesmith]

OK maybe I am misunderstanding your terms then - 'viewing distance' - do you mean how far I am to my screen or a print?

If you look on that page, the drop down box for coc has many presets including the 30D.  When you select the 30D it uses 0.019 for the calculation, I am not using it as a camera variable.

No, the Full image is not cropped at all.
[a href=\"index.php?act=findpost&pid=80798\"][{POST_SNAPBACK}][/a]

Exaxtly.  Viewing distance is how far away you are form the final image.

OK.  The calculator gave you 0.019.  Is that the value you want or need?  I can see that is the vakue the calcualtor has selected, but it is based on the calculator's author's assumptions about what you are actually doing or trying to do.  That person may not know.

What is the distance fom the the front to the back of the parts aof the flower you want to be in the DoF.  If you can measure that, then the math will give you the rest (the near and far and difference are functions only of the hyperfocal distance and the focus distance,  You can measure focuse distance and then cakcuk=late hyperfocal distance.  You then know the hyperfocal distance, the focal length, and CoC you need.  Simply calculate the needed f/stop.  Is it available on your lens?  (Is it possible?)

By the way, I really don't like your picture you have chosen for your posts.  You look very angry and combative.  Not exactly the face I want to try to help.

[dobson]

Depth of Field can be something of a catch-22 in close-up/macro photography. The more magnification you have the narrower the DOF, and stopping down the lens imparts it's own problems. When I was taking macros of mineral specimens, (300mm lens with a reverse mounted 50mm), the DOF was under a millimeter deep. And this was at F/40! At those high F-stops diffraction becomes a big issue and images never look as sharp as you would like. The only way I have found to remedy this is to bracket focus and combine the images digitally via specialized software, but that's a real pain.


This image shows how narrow DOF can get at high magnifications. Field of view is about 4mm across. 50mm reverse mounted on a 300mm lens, F25, .2 sec

[attachment=1051:attachment]

[deelight]

If you wanted the stamens and petals to be the front and back, respectively, of the areas in focus, you should have focused between them.  Set up the composition (tripod if possible) , find a part of the central structure that represents the right distance to be the midpoint, then shoot with several different apertures. 

Generally a good idea, only the splitting is not that correct:

Do not focus on the midpoint between the nearest and farest point of sharpness. The DOF starts 1/3 in front of the focused point and ends 2/3 behind it.

Just for the record...  

Best,

Clem

[kaelaria]

OK I researched coc deeper and now understand what it is, how to calculate it and what factors effect it.  Yes that calulator above is assuming 5/mm, 30cm @ 50cm.

I guess what this boils down to is, I did not have a print size in mind when I was shooting, and obviously was oblivious to coc.  

What I still can not find is a qualitative answer to what is an acceptible sharpness level, if that even exists.  I see the numbers and can alter them for circumstances - but I can't relate that to what the darn thing actually looks like!  Is there anything out there that actually gives examples - like a series of images to print or anything - so I can see what the various changes actually do?  Or do I simply have to make my own prints and experiment?

Is there, for eample, a chart that gives you average coc values for a given sensor size, with various full frame enlargement sizes?  Or do you simply figure your own at say, 8x10, and extrapolate from there?

[howiesmith]

What I still can not find is a qualitative answer to what is an acceptible sharpness level, if that even exists.  I see the numbers and can alter them for circumstances - but I can't relate that to what the darn thing actually looks like!  Is there anything out there that actually gives examples - like a series of images to print or anything - so I can see what the various changes actually do?  Or do I simply have to make my own prints and experiment?

[a href=\"index.php?act=findpost&pid=80846\"][{POST_SNAPBACK}][/a]

What is "acceptable sharpness" is a very personal thing.  Only you can and should decide that for yourself.  There are some very picky folks, and some not so picky.  Even the very same print will have different DoF for different folks.  Some too much, some not enough, same just right.  Sort of a Three Bears thing.  You will need to decide (or have someone decide for you if you trust them more) who your audience is and what they accept as "sharp" or "sharp enough."  For me, I am the audience, so "sharp enough" is very easy to determine.

There are likely some people who would say your choice of camera and lens will never make a "sharp" print.  You will certainly never please them.  There are others that would say you have a very fine camera - "Must take clear pictures."  End the end, only you can decide what is good enough.

That's one reason I don't exhibit my work anymore.  I just got tired of some folks saying the print wasn't "sharp" as they looked at my print with a loupe and their nose pressed against the paper.  I would say that the print wasn't designed for that scrutiny, but that didn't stop them.

[kaelaria]

I understand that now.

So I guess I'll just have to examine my own prints to determine where my 'just right' point is.  Without seeing the results yet, I would guess I swing towards the picky side, but that may change if indeed it's fine under my typical viewing conditions, we'll see

Thank you very much for the info everyone!

[howiesmith]

I understand that now.

So I guess I'll just have to examine my own prints to determine where my 'just right' point is.  Without seeing the results yet, I would guess I swing towards the picky side, but that may change if indeed it's fine under my typical viewing conditions, we'll see

Thank you very much for the info everyone!
[a href=\"index.php?act=findpost&pid=80855\"][{POST_SNAPBACK}][/a]

It gets even worse.  What is "sharp enough" also depends on the subject.  A very mature subject (been photographed a lot) needs to be sharper than an immature subject.  A bald eagle (millions of images) must be very much sharper than narwals mating (never photographed that I know of).  Look at the change in what is an acceptable photo of a polar bear over the last fifty or so years.

And a grandmother will accepted nearly any cute (but fuzzy) picture of the precious baby.