Focus Stacking

[stephencamptonjones]

It has been a long time between posts
This is my second attempt at using a program called "Zerene Stacker".
I stacked 93 jpg images for this image taken with a Nikon D80 50mm f1.4 @f2 (I thought I had set the aperature to f4,mmm)1/40 sec ISO 100, manually focused.
I have a lot more to learn in the area of retouching and I had to start with something difficult.



First attempt
119 images



Thanks for looking, Stephen

[Stagecoach]

Stephen, all I can say is very impressive !

It certainly adds interest to the 'manual lens focus' versus 'focus rail' procedure discussions we see so often. 93 and 119 are also the greatest number of slices used for a stack I recall having ever seen.

How did you manage to get equal increment rotations of the focus ring, was it by using the grip serrations ?

[Andrew76]

Very nice shots Stephen, I too would like to know how you managed the manual focus. Thanks for sharing!!

[stephencamptonjones]

Grahame
I just turned the focus ring a piffel (a carpenters term for small bit). I don't think the slices have to be exact but an aperature of f4-f8 is a good choice.
I also have a Nikon PB6 bellows unit that I could use as a rail

[DanK]

Stephen,

Beautifully done.

Like you, I use small changes in focus rather than a rail. Works fine.

If you don't mind a few suggestions:

I think you are making things harder for yourself than they need to be. There is nothing gained by opening wider than f/8 or so (very few lenses will be sharper at f/4 than at f/8), and at that aperture, I have never found it necessary to do more than 15 images for a single flower (your second image is deeper), and in many cases, 3 to 5 does it fine.

So you might try fewer, perhaps with slightly bigger changes in focus. If it works, it will save you time and labor.

A second suggestion is that you save or export the original images as TIFFs rather than jpegs. JPEGs are created by lossy compression, so they lose detail. Of course, with nearly a hundred images, TIFFs would eat up a huge amount of space, but with 3-15 images, even 16-bit TIFFs are manageable.

There are quite a few threads on this in the postprocessing forum, so you might want to check that out.

Dan

[Inkanyezi]

There is nothing gained by opening wider than f/8 or so

Oh, I beg to differ. A small opening will render background detail less unsharp than a very wide aperture. So for objects that are not separated by the use of a very smooth background, a wide aperture can be a boon because of the bokeh it creates. The bokeh is very different at f/8 compared to f/2, and stacking adds only depth of field but does not alter bokeh.

[DanK]

Oh, I beg to differ. A small opening will render background detail less unsharp than a very wide aperture. So for objects that are not separated by the use of a very smooth background, a wide aperture can be a boon because of the bokeh it creates. The bokeh is very different at f/8 compared to f/2, and stacking adds only depth of field but does not alter bokeh.

Yes, of course, that is so. But the point of stacking is usually to overcome shallow DOF. What I meant is "if you are trying to increase DOF, there is nothing to be gained by using an aperture wider than f/8 or so." DOF is so narrow at macro distances that I have very rarely had to worry about insufficient bokeh. If there is--if you have a background that is very close, or if you want to throw one part of the plant out of focus--then yes, it might be helpful to use a wider aperture. In my experience, I have very rarely had reason to go wider than f/8 when doing macro work that is anywhere close to 1:1. YMMV, but in the images that Stephen posted, I doubt this would be an issue, but I could be wrong.

[ajohnw]

Stephen,
A second suggestion is that you save or export the original images as TIFFs rather than jpegs. JPEGs are created by lossy compression, so they lose detail. Of course, with nearly a hundred images, TIFFs would eat up a huge amount of space, but with 3-15 images, even 16-bit TIFFs are manageable.
Dan

I just did a check an not at all unsurprisingly I find that jpg fine uses a compression level of about 99%. It's going to be similar on any camera with a decent pixel size. The losses are going to be virtually undetectable at this quality level. If some one did for some reason under the rather controlled conditions these sorts of shots are taken in decided to use raw they could in any case save 100% jpg's and finish up with a file of similar size to the raw file. In practice however far more will be lost when a shot is reduced to web size. On the other hand If a file is edited and saved at 99% and then re edited 3 times it finishes up at 97% hardly any different really but if saved at 100% it doesn't change at all.

Not that I have a bee in my bonnet about certain comments that often crop up concerning jpg's of course. It's just that they are a miss understood subject - caused by techniques used to minimise the storage requirements on web servers. I have always wondered why camera don't switch to png - loss less compression. I believe there is also a lossless jpg standard as well.

Last but not least I think you will find all stacking software works with 24bit colour anyway so maybe it's best to get that right before leaving it to the tender mercies of the stacker. That from a comment from one of the authors of Zerene if I remember correctly.

-

[DanK]

John,

Here is the recommendation from Zerene's author:

For highest quality, we recommend converting raw files to 16-bit TIFF using your favorite raw converter and whatever settings make it work the best. After stacking the 16-bit TIFFs, tell Zerene Stacker to save its output also as 16-bit TIFF.

Because 16-bit TIFF files are not compressed and also have deeper pixels than any current raw format, this process retains all of the image quality intrinsic to the raw formats.

In general, the best bet is to wait as long as possible before throwing out information. I don't know how you compared the jpeg to the tiff, but they should look the same on screen. However, that does not mean they will behave the same in editing.

Zerene ignores color profiles. DMap does not much affect color. PMax does, as Rik explains on the Zerene website.

Dan