Calculating DOF for macro work

[Stagecoach]

ControlMyNikon doesn't do the stacking, only the photo's.

George

That's the reason I use it.

[PhotomanJohn]

From what I recall the FL for my 105mm at MFD of 1:1 magnification ratio is 70mm.

I see you have both a 105mm AF-S and the "D" version as well. Which one is the 70mm for? I would assume that the "D" versions would not have as large a focal length reduction as the AF-S since it physically extends as you focus closer.

Trying to sort out all that has been discussed in this thread is giving me a headache.

John

[Stagecoach]

I see you have both a 105mm AF-S and the "D" version as well. Which one is the 70mm for? I would assume that the "D" versions would not have as large a focal length reduction as the AF-S since it physically extends as you focus closer.

Trying to sort out all that has been discussed in this thread is giving me a headache.

John

John, the 70mm (or something like 68.8 if I remember correctly) is for the 105 AF-S VR. I do not have a clue about the D version.

Macro DoF when discussed in these forums always goes to an extreme and whilst basic DoF calculations of varying degrees are there you can't beat trialing and noting what works for you. Before undertaking anything that may require 30 shots or more I will take say 5 at certain settings, stack and view the result to confirm.

I have just been playing with Helicon Remote, it's default Auto step calculation settings seem pretty good (for my lens).

[george013]

There is also another thing using the dof-calculator with macro. Be aware that distances are measured from the optical center of the lens, which is mostly unknown.

If you want to know the focal length of your lens at 1:1, make a photo at 1:1 and measure the distance from your object to the phi-sign on your camera. That's the sign indicating the sensorplane. At 1:1 that distance should be 4 times the focal length.
And once you know the focal length, you know the optical center of the lens, 2 times the focal length from the phi-sign. And you can play with the dofcalculator. But only at 1:1.

George

[ajohnw]

This might help Dan

DOF calculator

My gut feeling is that you will find that some experimentation is required depending on your camera/lens combination. if you use nikon your in luck in that a program called control my nikon will automate this for you!

The trouble with a LOT of the posts in that thread including mine is that they are APPROXIMATIONS but people seem to treat them as absolute and they aren't. They work reasonably well at normal focal distances and are based on a single simple lens replacing the real one that is actually used. If some wants a loose idea of how this is done look at principle points and thin lens here.

http://www.britannica.com/EBchecked/...eory-of-lenses

In essence the complex lens is replaced with a single thin lens as it makes the calculations possible. The actual dimensions of the internals of a complex lens and the various refractive indexes would need to be known to get an accurate answer along with some pretty complex software. The approximations work in situations where distances are much larger than the focal length of the lens and get less accurate as the distances are reduced. It's a bit pointless wondering about where they are measured from as the fictitious lens used to perform the calculation bears no real relationship to the real thing.

Some pundits seem to advocated using magnification to determine depth of field for macro work. This is just based on a ratio of distances from the same single lens but applies to any focal length. It's - image distance / object distance or focal length / ( focal length - Object distance ). i reckon in practice I get more based on full resolution crops but haven't actually measure it. Full res results take crop factors out of it. What is recorded is determined by the pixel density.

In terms of stacking software I think people will find that the important aspect is that some is out of focus and the more rapidly this happens the better. Then there needs to be some overlap. On that basis the magnification method should be ok and I vaguely remember that the Zerene site makes suggestions. The general consensus of late seems to be that too many images to stack doesn't matter and is a better option than too few. Privately I have been told that certain subjects benefit from a visual approach but that's when taking shots via a microscope so seeing what is in and out of focus is pretty easy. It is probably possible to do the same thing on a normal camera macro set up by using magnified live view at a level where what is actually getting to the pixels can be seen. That's what the software is working on. Extreme comment and not intended to be literal - If conventional DOF is used based on 10x8 prints etc then best reduce them before stacking otherwise there wont be much that really is in focus - but there still needs to be some areas that are out of focus.

John
-

[rudi]

Dan, for stacking I use the DOF estimates table on the Zerene website http://zerenesystems.com/cms/stacker.../macromicrodof
which works for me. But I guess you are aware of those tables.

I am sure that Rik Littlefield will help you out with any technical questions.

[DanK]

Dan, for stacking I use the DOF estimates table on the Zerene website http://zerenesystems.com/cms/stacker.../macromicrodof
which works for me. But I guess you are aware of those tables.

I am sure that Rik Littlefield will help you out with any technical questions.

Hi Rudi,

thanks very much. I didn't know about that table.

Here's how the estimates compare, again using 100mm (nominal), 1:1 (approximately), and f/8 (nominal):

1. My correction of the DOFmaster estimate, using the factor of 2 suggested by the tutorial, and correcting for both f/stop and focal length: 0.9mm

2. The default steps of Helicon remote (I still haven't tested whether they need adjusting): 1.9 mm

3. The Zerene table: 0.56 mm

Some variance is to be expected, but this much seems a lot to me. Next step is for me to test the Helicon steps. That is a bit of a production. You have to create stacks of shots of something showing detail from front to back, using the default and various adjustment factors. I figure it could take a few hours, once I find a suitable target.

This is only relevant for my work for two reasons. First, I never find that my manual steps are too far apart. However, that suggests that they are closer together than they need to be. Fewer shots would be easier to manage. Second, now that I have tried Helicon remote, I think I will use it. If I can get the steps right, it will remove the guesswork, and it will avoid what I now have, which is lots of extra shots at the back end because I am usually not certain where to stop.

At least this does seem to answer my initial question: apparently, you do have to correct for both effective f/stop and effective focal length.
Dan

[xpatUSA]

Hi Rudi,

thanks very much. I didn't know about that table.

Here's how the estimates compare, again using 100mm (nominal), 1:1 (approximately), and f/8 (nominal):

1. My correction of the DOFmaster estimate, using the factor of 2 suggested by the tutorial, and correcting for both f/stop and focal length: 0.9mm

2. The default steps of Helicon remote (I still haven't tested whether they need adjusting): 1.9 mm

3. The Zerene table: 0.56 mm

Some variance is to be expected, but this much seems a lot to me.
Dan

Dan, as you know I've studied DOF quite a bit. I find, in all the methods and calculators a very large variance in the so-called Circle of Confusion or, in Richard Lyon's angular terms, Cone of Confusion. The said Circle of Confusion is based on several methods, not just one, and a paper by Kerr says it all in that respect (heavy reading though).

I've been trying out several methods lately and, for what it's worth, I get 0.5mm and 0.6mm using two different methods for your example above, so I lean toward your result # 3.

Currently, I am leaning toward pixel pitch as a basis for the CoC on my camera (Sigma SD14) which I multiply arbitrarily by Rayleigh's criterion x 2 (2.44). It comes to a CoC in the image plane of 0.019mm. Some folks just double the pitch. I have also tried a factor of 2 x root(2) = 2.828. I like the pixel approach because it frees me from assumptions like "1/1713 of an8x10 print diagonal viewed at 250mm" or "the sensor diagonal / 1500", etc., ad nauseam. I hasten to emphasize that I am only interested in the use my own camera and viewing only my 1280x1024px monitor with it's dot pitch of 0.294mm, so I'm not forcing my methods on anybody and am not even claiming they are better than anyone else's.

If anyone wants to play with multiples of pixel pitch as a CoC, my Open Office spreadsheet is here. Excel available on request.



.

[DanK]

Well, I decided to make this an empirical question.

Ted, thanks for your post. Unsurprisingly, you end up in the same ballpark as Rik, the author of Zerene. However,

I photographed a stack using Helicon remote at its default setting. I then doubled the correction factor, as described here: http://www.heliconsoft.com/focus/hel...CUS_BRACKETING. I then ran both stacks through Zerene PMax, which is the best option for preserving fine detail. I did this with my standard Canon 100mm L macro, with no extension. I photographed a piece of denim, because that has detail front-to-back.

The default setting gave me no visible gaps. A correction factor of 2 (reducing the number of slices by a bit over half) did, as described in the link above.

So, this first test suggests that whatever internal calculation Helicon remote uses is OK. I intend to use the software a good bit (in combination with Zerene), because it really is much easier than making a large number of manual adjustments. That will give me ample opportunity to test this setting under different conditions.

I'll also do a few more measurements; it's possible that I simply messed up the measurements in my first trial with a clip, or had the aperture set differently than I thought.

[xpatUSA]

Well, I decided to make this an empirical question.

Ted, thanks for your post. Unsurprisingly, you end up in the same ballpark as Rik, the author of Zerene.

<>

I'll also do a few more measurements; it's possible that I simply messed up the measurements in my first trial with a clip, or had the aperture set differently than I thought.

You probably didn't mess up and I certainly wasn't suggesting that any of your results was wrong. I just couldn't resist having a look using my own stuff plus taking the opportunity to sound off a bit