Focus stacking, in-depth question

[Seapy]

The reason I returned to my interest in photography (apart from the 'free film' provided by digital ) was that I wanted some illustrative input to add to my late fathers unpublished 1973 manuscript on commemorative plant names.

Since 2006 (when I got my D1) I have been visiting botanic and other gardens gathering photographs of commemoratively named plants with the eventual aim of incorporating them when I publish his work, which amounts to about 240,000 words, cataloging and describing the work and lives of the people who have been commemorated by naming a plant, species or genus after them, usually in a Latinised form.

Photographing plants and flowers well isn't particularly easy as I have discovered, particularly outdoors.  I try to keep the background in soft focus to emphasise and separate the subject.  I try to make the picture pleasing, sharp and detailed, a mixture of art and accuracy.  To this end I have tried to use good, fast manual lenses like the Nikkor 105mm f/2.5, micro Nikkor 55mm f/2.8 and occasionally, where access is limited the Nikkor 300mm and 180mm -  f/2.8.  Apart from the inevitable breeze which springs from nowhere the moment I have finished composing, focusing and place my finger on the shutter release, is the consideration of depth of focus.  For a single flower or if I am lucky enough to find a bunch of flowers in the focal plane, that's usually OK, but for some plants and trees it becomes an issue.

This Magnolia Sprengeri is pretty typical.  Taken with D200, micro Nikkor 55mm f/2.8 @ f2.8 ISO200. Two exposures, one focused on the back flower, the other on the front flowers.





I overlaid the two images in Photoshop and manually blended the two to form one image.

This is what I came up with...



But I felt this was better...



I have used this ad-hok method occasionally where the situation demands but it's pretty hit or miss and very time consuming.

I have been following Michael's activities for many years with interest, for me I have felt that although his work is wonderful, it goes beyond my needs.  However, for some specific occasions where I can access nice examples of plants and flowers in a controlled environment, indoors, I am beginning to wonder if a motorised stacking device might be worth investigating, to create some stunning illustrations.

What I am unsure about is the difference between moving the focused camera to achieve the stack or to alter the focus.  Maybe it's mechanically easier to move the camera/lens.  But, what about making a simple mechanical step focusing attachment which turns the focus ring with a stepper motor???

With a longer focal length the scene won't change very much but with a shorter focal length, moving the lens/camera forward will alter the scene by parallax.  Does the stacking software compensate for the parallax changes, does this cause problems, artefacts?  Even re-focusing can alter the parallax a bit esp. with CRC lenses, where the focal length changes.  I don't want to invest time and effort into making a stacking slide to find re-focusing stacking is better, I have plenty paperweights and doorstops!

Thoughts or comments please.

[basker]

What I am unsure about is the difference between moving the focused camera to achieve the stack or to alter the focus. 
Thoughts or comments please.

Robert,

I really like what you are doing, and the question is one I would like to have answered as well. No claims of expertise, but here is what I think. This is not intended to address tilt, shift, bellows, internal focus etc. It is just notes I made to see if I could justify buying a rail.

Assuming the lens is focussed by moving the optics as one block (like your 105/2.5 does)
* If the camera is stationary and  the lens focus is changed, then both the position of the focal plane and the magnification is changed.
* If the lens focus is locked and the camera is moved along the axis of that lens, then the focal plane simply moves an equal distance.
* Parallax would only occur if the lens axis is not parallel to the direction of movement.
* Perspective is just about how close the entrance pupil is to the subject. (per Bjørn) That would apply equally to either case.

It seems that using a rail requires careful alignment of the lens, but it does avoid changing magnification. Apart from what the stacking software expects, I think rail vs. no rail mainly depends on subject size and distance.

The handcrafted results on the  magnolia are really nice. Not having to do it every time sounds pretty good too. Feel free to correct any technical mistakes, I am just trying to get the first pickle out of the jar to make the rest easier.

Sam

[Bent Hjarbo]

First the magnolia image is very nice made.
I am definitely not the expert on this, but I have had the same speculations regarding the size of the image when moving focus the get from the top to the bottom of a flower.
I have use Michaels notes/posts/books as a source for finding my way.
Today I just for the purposes of looking into the problem two set of pictures of a plat my wife jut got.
First by moving the PB-4 on the slide rails, keeping the the the lens at a fixed place and moving it all.
Next by moving the lens to change the focus.
I used my 105mm f4 bellows and a D800.
The filenames refers the the method used. It can be seen that the size of the flower change regardless of the method.
I have used LR and Photoshop to do the stacking, I have not done any manual improvements on the output files, which can be seen, other software may do a better job?

[basker]

...I have had the same speculations regarding the size of the image when moving focus the get from the top to the bottom of a flower.
... It can be seen that the size of the flower change regardless of the method.

Bent,

I agree with your observation. Both methods show a change in apparent size. In the first example, I think the size change is caused entirely by perspective change. In the second case, moving only the lens, I think both magnification and perspective changes. I would be interested to know you have a preference regarding which is the better stack?

Michael's writing and his work are impressive, and he is undoubtedly committed to improving and documenting his methods. I want to be as open minded as he is.

Best regards,

Sam

[Bent Hjarbo]

My stacking experience is not very long, so I have no preferred method yet.
PS do the resizing so the stack can be done either way.
PS might not be the best software fir the task.

[Seapy]

Sam and Bent, sorry for not engaging in this discussion, I have been struck by a returning infection, thinking straight isn't one of the things I can do right now!

I think we can conclude that no matted how focus is stepped, the image changes in size and parallax movement occurs.  The only way to prevent parallax movement errors is to lock down the entrance pupil relative to the subject and move the sensor (camera) to step the focus, which in reality is't really practicable.  Also the act of focusing would likely still change the relative size of the recorded image.

In which direction do you stack Brent?  From the front to the back or from the back to the front?  My immediate expectation would be to start at the back and work out towards the front.

That way the exposures would be stacked like looking down on a Christmas tree, so to speak, with the 'largest' at the back, which it would seem to me should reveal the best vision. But like I say I am not thinking clearly right now.

For stacking landscapes, clearly moving the camera isn't an option so stepped focus has to the the only way, perhaps it's horses for courses.

Sorry for rambling. 

[Bent Hjarbo]

My stacking is from front to back.
One could reverse the order of the images if that would give a nettet result

[basker]

Sorry for rambling. 

Robert, I enjoy a good ramble. 
I wish you a quick and complete recovery.

[Krishna Kodukula]

Robert,

That is a phenomenal project. Congratulations! I would love to acquire a copy when you publish the work.

Similar to you, I have been following Michael's work for quite some time and enjoyed his detailed writings. However, my stacking experience is not very long (learning the technique and slowly acquiring necessary tools) but I seem to like stacking flowers from front to back. It's just that I tend to focus on the nearest point first and then go farther and farther away.

Krishna

[Michael Erlewine]

The stacking approach is tailored for the software we use. It does not seem to matter whether you stack from front to back or from back to front. What matters is what method you do use. According to Rik LIttlefield, designer of Zerene Stacker the software I find most useful for my work says the methods to stack, from best to worse, are:

(1) Best is to use a bellows system, fixing the front standard (the lens) and moving the camera with the rear standard.

(2) The second best method is to use the focus barrel on your lens.

(3) And the third best (which also is known as "the worst") method is to mount your camera and lens on a focus rail and move the whole group forward on the rail.

In my opinion, the crucial thing to learn is how to compose the shot properly. What are you taking? Is it just a photo? Is the composition as important as the subject? Is the composition MORE important than the subject? What is the photo saying and to whom?

Is stacking from front to back to make a perfectly-in-focus photo what you are after? Or, do you want parts of your photo in perfect focus and the rest a dreamy bokeh, which is my favorite. And what does that say? To me, that says this world we live in is like a dream and I am preoccupied by (the in-focus parts) this or that part of the subject. Is focus stacking a craft or an art? Or both? I believe it is both.