Focus shifting (landscape)

[blackstar]

My next project (when the weather turns and all view clears) is to shoot a Yosemite scene from far (miles) away (see below screenshots from Google Earth).

4X zoom (clearly see the Half Dome in view)

Since the view is so far away, I plan to use Sigma 150-600mm lens to reach/catch the main object. Also don't want to miss view of the foreground and between. So it will be a focus-shifting project for landscape with a telephoto lens. Had researched quite a few online tutorials and done simple experiments trying to make it ready. Online tutorials are all too simplified without detailed camera settings and my experiments are limited with not too far away objects in view. So don't think it's ready.
Experiment results:
1. FS settings: # of shots=20, actual shots=11, focus step width=4, interval=0", 1st frame exp lock=ON, Silent photo=ON.
Camera settings: Focal length=150mm, f6.7, iso=200 (I think too high), speed=1/125, AF-S, single point (I think wide area S)

2. Same FS settings, actual shots=20, FL=600mm, same camera settings. Not stacked due to focus clearly not reaching the far end.
3. # of shots=40, actual shots=40, same camera settings. Stacking result:

4. # of shots=60, actual shots=58. same camera settings. Stacking result:

It's obvious for 600mm FL, it takes at least 60 shots to reach infinity in the experiment case (assuming camera stops shooting when whichever # of shots or infinity is reached earlier. In this experiment, 1 and 4 succeed. Still, other than FL other parameters could affect the "number of shots" setting. Mostly, aperture setting, focus step width, and focus distance. I wonder in the real case of Yosemite, it may take more than 100 shots or even 200 shots to succeed. I think it needs the correlation from all these parameters to decide the number of shots. Is there a correlation formula or rules for this matter? Any thoughts or comments?

I use Zerene stacker to merge the shots. It states sometimes it could make faults in the layer stacking as shown in 3 and 4: some foreground parts or objects become transparent. Though it can be corrected by the program's retouching tool (or other editors' layer blending tool).

 

[Fred Kingston_RIP]

Focus step calculator Here's a video

 

[blackstar]

Thanks, Fred. The Focus Step Calculator works more effectively and accurately for Canon's "focus increment" setting. But it contradicts Nikon's "focus step width" setting (see https://nps.nikonimaging.com/technical_solutions/d850_tips/useful/focus_stacking/ ). May need more experiments to validate the calculator with Nikon cameras. The video is for astrophotography which works within the micron range and is probably not practical for landscape.

 

[BF Hammer]

I have not checked my Sigma for focus breathing. As in the focal length changing as you move the focal distance. It could play into your project. I have had some zoom lenses that were very bad in this regard. The Tamron 18-270mm would only reach 270mm at infinity focus, and looked more like 210mm at close focus.

I've never fooled with focus-stack yet. All I know is that the in-camera solution to this is to use a tilt-shift lens to greatly increase the infinity focus depth.

 

[blackstar]

I have not checked my Sigma for focus breathing. As in the focal length changing as you move the focal distance. It could play into your project. I have had some zoom lenses that were very bad in this regard. The Tamron 18-270mm would only reach 270mm at infinity focus, and looked more like 210mm at close focus.

I've never fooled with focus-stack yet. All I know is that the in-camera solution to this is to use a tilt-shift lens to greatly increase the infinity focus depth.

I haven't considered the possible focus breathing issue in my experiment. If this issue affects the focus shifting with the Sigma lens, how do you verify it in my experiment results?

It seems that there are only prime T-S lenses, and no zoom lenses working on T-S. And all T-S lenses are MF, if not wrong.

 

[Fred Kingston_RIP]

I use camRanger 2 for focus shifting... tap the two focus points, (near and far) and camRanger does all the calculations and just goes to work...

 

[blackstar]

BTW is my Yosemite project realistic?

 

[blackstar]

I use camRanger 2 for focus shifting... tap the two focus points, (near and far) and camRanger does all the calculations and just goes to work...

That may well work for Macro. For landscape, I prefer the far point to be infinity, i.e., clouds without "pointing" it.

 

[BF Hammer]

I have not checked my Sigma for focus breathing. As in the focal length changing as you move the focal distance. It could play into your project. I have had some zoom lenses that were very bad in this regard. The Tamron 18-270mm would only reach 270mm at infinity focus, and looked more like 210mm at close focus.

I've never fooled with focus-stack yet. All I know is that the in-camera solution to this is to use a tilt-shift lens to greatly increase the infinity focus depth.

I haven't considered the possible focus breathing issue in my experiment. If this issue affects the focus shifting with the Sigma lens, how do you verify it in my experiment results?

It seems that there are only prime T-S lenses, and no zoom lenses working on T-S. And all T-S lenses are MF, if not wrong.

It should be simple to test. On tripod, shoot 2 photos. Minimum focal distance and infinity. If it looks like you zoomed the lens between the photos, there is a focus breathing issue. It is a common thing, kind of matter of how much zooming you get between the extremes.

 

[BF Hammer]

I use camRanger 2 for focus shifting... tap the two focus points, (near and far) and camRanger does all the calculations and just goes to work...

I wonder if Blackstar is missing the point of this? CamRanger is controlling the camera and doing all the work of moving the focal plane and taking the photos I presume. Not trying to manually move the focus ring by hand in a series of steps. Sometimes people forget you can connect that USB cable to a PC for control of the camera and not just for transferring photos. And that includes using the autofocus motor to shift the focus by precise amounts. Or wirelessly by Bluetooth or WiFi as an option with Z bodies.

 

[blackstar]

It should be simple to test. On tripod, shoot 2 photos. Minimum focal distance and infinity. If it looks like you zoomed the lens between the photos, there is a focus breathing issue. It is a common thing, kind of matter of how much zooming you get between the extremes.

Got it. I think this does not matter cause if FB affects my experiment much, the stacker would fail or render bad results.

 

[blackstar]

I wonder if Blackstar is missing the point of this? CamRanger is controlling the camera and doing all the work of moving the focal plane and taking the photos I presume. Not trying to manually move the focus ring by hand in a series of steps. Sometimes people forget you can connect that USB cable to a PC for control of the camera and not just for transferring photos. And that includes using the autofocus motor to shift the focus by precise amounts. Or wirelessly by Bluetooth or WiFi as an option with Z bodies.

No, I do appreciate Fred's sharing of using CamRanger. With Z camera's built-in focus-shifting function, there shouldn't be much difference from using CamRanger with DSLR. As I mentioned, for landscape I prefer to stop shooting until reaches infinity.

 

[Fred Kingston_RIP]

That may well work for Macro. For landscape, I prefer the far point to be infinity, i.e., clouds without "pointing" it.

Then you would simply select the clouds as the far point... When I say "select" you are using a cell phone or tablet as the remote tether/viewer, and you're actually "tapping" the screen... Tap the near point on the screen and then tap the far point on the screen... If the far point is infinity, then so be it...the device calculates the number of frames between the two and lets you set certain parameters... Regardless, for that type scene. I suggest you review your camera's Compressed RAW settings to reduce file sizes and your memory card's space...

 

[blackstar]

Then you would simply select the clouds as the far point... When I say "select" you are using a cell phone or tablet as the remote tether/viewer, and you're actually "tapping" the screen... Tap the near point on the screen and then tap the far point on the screen... If the far point is infinity, then so be it...the device calculates the number of frames between the two and lets you set certain parameters... Regardless, for that type scene. I suggest you review your camera's Compressed RAW settings to reduce file sizes and your memory card's space...

Thanks for the details, Fred. So the device calculates the number of shots... will it tell you the number before starting shooting? I just want to know the shot number. Also, it seems to be uncertain that you will be able to "tap" a point that represents infinity. Yes, file size, not file number, does pose a problem in my case: due to my old age Mac, the focus stacker limited file size to < 75% of fine size to work, otherwise it fails. Should I set small-size images or do batch-rescale images afterward?

 

[Fred Kingston_RIP]

Yes... Basically, you set all the parameters in camRanger... focal length, fstop, iso etc... then select the near and far points... and it displays the number of shots... you then hit the "start" button, and it starts taking the shots... Similar process to the Z camera's focus shifting...

You don't actually select infinity... after you select the near point, there is like an arrow key on the screen that you start tapping and the lens starts shifting focus farther out...you can see the farther areas start coming into focus and the near area start to get blurred... once you get to the farthest area you want in-focus, you tap the Set button for that... so that now camranger knows where the close area is and where the farthest are is... it then calculates the shots... If you're satisfied with the settings... you hit the Start button

 

[blackstar]

Thanks, Fred.

 

[Fred Kingston_RIP]

My typical workflow is to add all the NEF images to a separate folder... and then use a RAW viewer to look at the images... for Focus stacking I typically select an area in front of the near point, and an area beyond the far point... that results in images in front of and behind where I want focus... so I discard those images... I then typically run DXO's PureRaw on all the images... That cleans up any noise and runs lens correction on all the images... and converts the NEF files to DNG... I then Import the DNGs to Lightroom... If I need to make any adjustments there, I only adjust the first image in the series...and then create a pre-set for that image, and then use that pre-set and sync it with all the other images... I then EXport the images for StarStax... and bring the final image back to Lightroom for final editing

 

[Paliswe]

I have tried to read all the experiments you have described here, but since I'm not native english speaker I might have missed some details.
I will however described my experiences with focus shifting. For me the number of pictures I have to take from "before" the subject to "after" is not dependent on the camera but on the lens.
How big is a focus step? I have five lenses with slightly different focal lengths (and different apertures) that were mounted on my Z6. I have laid out nine rubber cubes with 10 cm intervals (0-80 cm) on the floor and for a number of different focal lengths the camera will take 10 pictures with an offset of 2 steps for each picture, according to the scale found in the camera's menu. The camera thus sends out 20 steps to the lens for each series. On the last image in each measurement series, I note which cube that is in focus.
To get the same size of the cubes at different focal lengths, the camera is moved out.

The lenses I have used are
24-70 f/4
24-200 f4-6.3
70-200 f/2.8
105mm f/2.8 MC
60mm f/2.8 Micro F-mount, mounted on an FTZ adapter.

With three zoom lenses and two primes, I can get a total of 13 measurement series, two measurement series each for 24mm and 200mm and three measurement series each for 60mm, 70mm and 105mm.

The distance for the five measurement series was calculated by setting the smallest focal length, 24 mm, to 1 and the other focal lengths were calculated in relation to this. The shortest focus distance for the two 24-mm lenses is approx. 0.5 m. Then the distance for 60mm == 1.25m, 70mm == 1.46, 105mm == 2.19m, and 200mm == 4.15m. Then the cubes will be the same size on all images. The camera was placed on a Manfrotto Pixi.
Tape on the floor marked where the camera would be placed for the different focal lengths. The camera was placed so that the sensor itself ended up directly above the pieces of tape.

Results
The last image in a focus series shows which cube the lens' focus ends up on. It varied between the second and eighth cube, ie the focus shifted between 20 and 80 cm.
For 24-70 and 24-200 with focal length up to 70mm it was 5-6 cubes, for 24-200 at 105 and 200 the displacement was 8 cubes
For the other lenses, 60mm Makro F-mount, 105mm MC Z and 70-200, the displacement was 2-3 cubes.
The attached pictures shows the last ones for 24-200mm and 70-200mm at 200mm.

What is interesting is that the cheaper 24-200mm lens goes "longer" at 20 microsteps than the 70-200mm does. That means that the more expensive 70-200mm lens have up to 4 times more steps to focus on than the cheaper 24-200. Thus, the 70-200mm lens can focus much more accurately than the 24-200 can within the same focal length.
I don't know if this has any effect on focusing accuracy in real life when taking a single photo.

.

 

[blackstar]

In response/appreciation to Paliswe's post, give my thoughts on 2c:
First, my bad if wrong I think for z cameras' settings of focus shifting, "focus step width" is related to the "depth of field" of the system. And the "DoF" is set by focal length, aperture, and subject distance. The "step width" is then set to be a fraction of the "DoF" (e.g., step width=4= 4x1/10xDoF=4/10 DoF, the number "10" is just my guess because it is the max number you can set for step width)
Second, I think the number of steps is the number of shots. The complicated thing here is, say, the first shot is taken at the subject distance determined by the photographer (the nearest point), and then the camera moves FP to 4/10 of the first DoF (example), takes the second shot, then moves FP to 4/10 of the SECOND DoF (because the subject distance changes, so does DoF), continues...
Third, if my above second thought is valid, for my project it seems to be too much work needed to know the exact number of shots for reaching the infinity point (beyond the farest-away subject).
p.s., For Macro or close-up (like Paliswe's experiment), it will need much less work to figure all this out (I think or maybe not?).

 

[Paliswe]

Determining the number of shots when doing focus shift in landscape photography is unnecessary as the camera stops when you reach infinity. Just set the number to 100 or 200 or a value that far exceeds what is needed. If you reach infinity after 12 shots, the camera stops.
Should you not reach infinity, when the camera stops, just continue with one more shooting, the lens will continue from the last focus distance.
The number of steps is the number of shots if you set the step to 1. In my experiment I have set the focus step to 2. Sometimes (before I became aware of the differences between the lenses) I have used 5 steps, resulting in blurry areas when I finally stacked the pictures.
The problem is that you don't know how big a focus step is for a certain lens. You want it to be 4/10 of DOF and that's a good assumption, but different lenses can give anything between 2/10 and 8/10 of DOF (or more).
Since information about this cannot be obtained from calculators based on the camera, you have to test what number of focus step you shall enter in the camera's focus shift menu, for each lens.
In my case I now know that for my Z 24-200 I should always use 1, but for my Z 70-200 (and Z 105) I can happily use 5.