Nagging Question #1

[rtbaum]

As I was reading a post in the Nature Photographers Network, I ran across a post regarding focus stacking of landscape photography. http://www.naturephotographers.net/i...990bb2&u=14158. My question would be; Does the focal length of the lens and curvature of the lens elements of a landscape type of lens make this a lesson in futility. More specifically, my 90mm macro is designed to capture an image on a flat plane with very little distortion on the edges. A 28mm landscape type of lens has much greater curvature, causing distortion around the edges. Even if one were to design a 30 ft focus rail, would it be possible to stack the photos into quality result?

[herbert]

Hi Randy,

In this case you would not use a focus rail since it is not possible to change the position of the camera through the entire scene. You would simply change the focus point of your lens. Ideally you would want a lens that exhibits very little focus breathing. This means it can change focus without changing the angle of view. If breathing does occur then you may have to resize images so that they are in register. This should be possible with good focus stacking software.

Note that lenses with very little focus breathing are very expensive. This is a key feature of movie lenses and their extraordinary cost. For example in a movie it is very desirable to change the focus point from two actors in a scene without the room changing shape. This makes the viewer feel sick.

I do not think that the curvature of the lens will be an issue since each shot will be taken with the same curvature. Perspective correction software should be able to distort each image to be the same. However if the curvature changes with focal length then it will be a problem. In this case it is possible that the objects near the lens, or close to the edge of the frame, will change shape due to an altered perspective distortion not exactly corrected by the software.

In addition when blending multiple images of a landscape any movement in the scene is impossible to control since you cannot lock down the entire landscape. This typically occurs on windy days with trees. This is also a problem for HDR software since the images do not overlap and you can get 'Ghosting', or apparent appearance/disappearance of parts of the image. This can be solved by cloning back pixels from a single image.

I have not tried focus stacking for landscapes. In general I find a F11-22 will adequately cover my desired depth of field anyway. With macro the depth of field is so small that focus stacking is more useful.

Focus stacking should allow you to combine landscape images all shot at below the diffraction limit for your sensor. However combining images form a stack may require scaling the images and so you immediately throw away any advantages of a sub-diffraction image since it is resampled during scaling.

You could try this out and see how it works for you. Helicon Focus and Zerene Stacker both have free trials where you can test them out. I would be interested to find out how you get on.

Alex

[FrankMi]

Hi Randy. Although it is always wise to find out as much about a subject as you can through research, this really needs to be augmented with your own experimentation to get a better feel for what is really happening.

Your own testing can provide you with answers such as for a particular set of conditions the result will be acceptable (to you) but for another, the results will be unacceptable. In addition you will learn exactly where and how the results will differ. Armed with this information you can more easily seek out alternatives.

"Experience is what you get when you don't get what you wanted."

[DanK]

I've never tried, but I suspect it will work.

I do all of my macro work without a rail, by changing focus. When I stack images, each image has a different FOV. The stacking software I use (Zerene) handles that fine.

Re curvature: I'm guessing it won't matter, but I haven't tried stacking images taken with any lenses other than macro lenses. I think the software should simply pick from each image the parts that are in focus and ignore the rest. With one of Zerene's algorithms (DMap), you can actually watch this happening in real time on the screen.

But by all means, test it out and post your results.

[Geoff F]

I have done it with CS5 using 7D and 24-105 lens.

Obviously a tripod and manual focus work best. The greatest potential problem is movement in the scene between shots. Wind movement of trees, water or people etc.

I found that I tended to prefer simply Auto Aligning the shots as layers then selecting which areas to use with masks. But Auto Blending should work in most cases.

[rtbaum]

I don't actually have a wide angle, but just got an itch as to whether it could be done. In the winter, I install Christmas lighting and like to document the installation for future installs. I figure that the shots should at least be nice to look at as well.