Can you "Focus Stack" using a lens with movements applied?

[wallpaperviking]

Very possibly a stoopid question but I could not find any answers online...

Just wondering if you can focus stack on either a Tech Cam with movements applied to a lens or a GFX and Canon TS-E combo?

Or is this something only achievable by the dedicated combo of a lens and its intended camera body ie GFX lens on a GFX body?

Also, is it mainly using the software that camera manufacturer provides or is it also a third part thing?

Thanks in advance!

 

[pegelli]

If you want to use the "automatic in camera AF focus stacking" you need an autofocus lens/rig to achieve that.

However "manual" focus stacking adjusting the focus of the lens manually (or using a macro rail for small subjects) is possible with any camera-lens combo

 

[wallpaperviking]

If you want to use the "automatic in camera AF focus stacking" you need an autofocus lens/rig to achieve that.

However "manual" focus stacking adjusting the focus of the lens manually (or using a macro rail for small subjects) is possible with any camera-lens combo

Thanks so much, much appreciated!

 

[dchew]

Shifts and rise fall won’t have any effect and focus stacking can be done as normal. Tilt would be extremely difficult to focus stack, but I can’t imagine the need to use both. You would just focus stack without tilt, or get everything sharp you want with tilt and not focus stack.

On technical cameras with helical systems, manual focus stacking is a breeze. Bellows systems require a bit more care but as long as you can rack focus in small enough increments it’s also pretty easy.

You just need to move the lens in (or out) the same linear distance each time. For example, most of Alpa’s helicals have a rotation pitch of ~0.03mm. I find at f/11, rotating the helical 5 degrees between images gets me acceptable results. That’s a lens movement of 0.15mm between shots. This is regardless of focal length or focus distance on a 54x40 sensor. Super easy to do with the HPF rings installed.

Dave

 

[MGrayson]

I don't see a problem focus stacking with tilt. The in-focus wedges will behave the same as the in-focus near-and-far planes in normal focus stacking. Near the sensor, DoF is a small neighborhood of the plane, and the angles of incidence of rays won't change much for small focus movements. (Probably wrong for extreme macro.) You'll get some image shift, the software should handle it.

Worth an experiment!

 

[rdeloe]

I have used focus stacking with tilt to solve some particularly difficult up-close problems. I don't do it a lot because I don't like relying on focus stacking, and because it's even less predictable than regular focus stacking. But it's there for emergencies

For the OP, here's a graphical demonstration so you can see what's happening: http://static.timparkin.co.uk/stati...ngth=50&N=11&coc=0.02&cameratilt=0&risefall=0

Try adjusting the Focus and Lens Tilt sliders to see what happens to the zone of acceptable sharpness. Basically, if you can adjust focus and tilt to get different parts of the scene in focus, you can stack the images. Whether it works well will depend on the scene, and all the factors that determine success in focus stacking without tilt.

 

[gerald.d]

I don't see a problem focus stacking with tilt. The in-focus wedges will behave the same as the in-focus near-and-far planes in normal focus stacking. Near the sensor, DoF is a small neighborhood of the plane, and the angles of incidence of rays won't change much for small focus movements. (Probably wrong for extreme macro.) You'll get some image shift, the software should handle it.

Worth an experiment!

The tilt and swing angles required to maintain the same focal plane angle change with focus distance.

Assuming you wish to stack perpendicular to the plane of focus, as long as you adjust the tilt and swing angles for each image accordingly, you can stack just fine.

 

[MGrayson]

Lines of the same color represent what is in focus outside the camera and where it is in focus inside. (Jargon suppressed.) The gray slanted line is the lens. The gray box inside the camera represents the sensor at some focus position. The quadrilateral outside the camera is the region in focus with the sensor in that position. As always, it's easier to keep the lens still and move the sensor, but the difference is small.

See, as far as the tilted lens is concerned, everything focus stacks just fine parallel to its plane. "Real" focus stacking will cut the sensor through those regions of sharp focus (between the colored lines above) at an angle, but will eventually cover everything inside the camera. Thus everything inside the slanted lines outside the camera will be in focus in some slice.

The even spacing of lines inside the camera corresponds to equal small turns of the helicoid. The Uneven spacing of lines outside the camera corresponds to the uneven distance markings on the lens barrel.

Upshot: You don't need to change ANY tilt or swing angles. Just stack as you normally would. You'll cover a big region with sharp focus. It won't be a rectangular box, but it will cover everything. If You WANT only a rectangular box in focus, then you have to get fancy.

Note: The thin lens equation was replaced with something stupidly simple, but qualitatively correct.

 

[gerald.d]

Ho hum. I simply stated two completely factually correct sentences.

I’ve probably done more tilted and swung focus stacks than anyone on the planet.

If you start with a requirement to tilt and swing your focal plane, you are doing that for a reason - typically because you want to align the focal plane with a surface that you want to get in focus that is not parallel to the sensor plane, and, the depth of the object whose surface you want to align your plane of focus with is deeper than the depth of field you get from a single shot.

The subject being shot might be something like a cereal packet, a perfume bottle, a computer, phone, or indeed anything in a box really, or a watch - typically the requirement to tilt and swing and focus stack will arise when you have an object offset to the plane of the sensor that has a relatively large surface area on two dimensions (let’s call them X and Y) to its depth on the third dimension (let’s call it Z).

The rationale for wanting to tilt and swing and focus stack will be to get the entirety of your subject in focus, but have focus fall off as you move away from the plane of focus up and down the Z axis. You do NOT want to simply “cover a big region with sharp focus”. If this was your aim, then why bother with the lens movements in the first place? Simply shoot the whole scene without any movements and just do a stack down the axis perpendicular to the sensor. It is simpler, faster, and easier to not bother with the movements.

And why would I want to focus stack along an axis perpendicular to the lens plane (your bolded comment)? Can you provide a practical example where this would be of benefit?

It’s all very well drawing pretty diagrams, but the practical applications for needing to do what was asked in first post are actually few and far between. If you want your subject in focus, and have the focus fall off down the subject’s Z axis satisfactorily, then you will need to adjust your lens movements as you stack. If you don’t want this, and just want to “cover a big region with sharp focus”, why bother stacking with lens movements in the first place?

 

[gerald.d]

There are three ways to shoot the above example, depending on the desired result.
1. Set the plane of focus on the face of the watch. Stack sufficiently to get the desired depth of the watch in focus (here, it was the LCD display to the top surface.
2. If you want the entire watch, including the band, in focus, set the focal plane perpendicular to what is seen in the image (basically parallel with that black line that goes around the bracelet), then stack through the “depth” of the watch head and bracelet.
3. If you want the entire scene in focus - here of course that scene is just the watch, but imagine the watch is set on a table, don’t bother with movements and just stack normally.

1 and 2 require the lens movements to change as you work through the stack.

 

[MGrayson]

And why would I want to focus stack along an axis perpendicular to the lens plane (your bolded comment)? Can you provide a practical example where this would be of benefit?

You wouldn't, of course. As you say, it would defeat the whole purpose of tilt. Nevertheless, that is what is going on inside your camera. Realizing that makes it clear what happens when you refocus without changing tilt. Focus is not a plane, it is a region, and that region is not a flat slab if there is tilt. If you WANT a flat slab in focus, of course you have to alter the tilt parameters as you stack. The question of what will happen if you DON'T alter the tilt is that stacking will work - you won't have OOF gaps - but you will get a larger wedge. That is the entire point of that diagram. I admit that one diagram isn't enough to make my point, but a 20 minute lecture won't fit.

And situation 2 above does not require altering tilt/swing parameters as you refocus. All you're doing is causing your in-focus wedges to overlap more at the far side of the band. Does it matter to the final image? Not a bit. BTW, if I'm wrong about that, please let me know why. I love being wrong. It's how I learn. (Yes, I'm opening myself to a cheap shot. Feel free.) Situation 1, requiring a flat slab only to be in focus in the final image, definitely DOES require adjusting parameters through the stack. I'm not implying otherwise.

I will not play the resumé game - I don't argue by reputation. I let the diagrams speak. It is common for expert practitioners to sneer at understanding. It's not even important that they understand - they're experts and produce great work. I am a specialist in understanding. Some expert practitioners still have an open mind to different ways of looking at their craft. Some are satisfied with their expertise. I am NOT discrediting expertise. If someone wants to know WHAT to do, it is the best. If someone wants to know WHY, that's where I come in.

Matt

 

[MGrayson]

Here's situations 2.
Adjusting the tilt parameters while stacking gives nicely overlapping zones of focus, all parallel to the centerline of the watch band.


Not adjusting the tilt parameters while stacking gets everything in focus, too, but spreads off the back. If there's nothing there but black cloth, the result is the same.


Yes, that's supposed to be a camera off to the right, and a watch with band in thick lines on the left.

 

[dchew]

I feel responsible for starting this debate with my not-too-well-thought-out comment: "Tilt would be extremely difficult to focus stack, but I can’t imagine the need to use both. You would just focus stack without tilt, or get everything sharp you want with tilt and not focus stack."

As usual in these scenarios, I was thinking myopically about landscape photography, completely forgetting about Gerald's wonderful watch images, where tilt and focus stack together are essential, and his equally fascinating CapCam. Sorry Gerald! I will say that without the CapCam, I think accomplishing what Gerald does manually would be pretty tough, at least for me.

I also agree with Matt that in more mundane landscape applications, it might not be as difficult as I originally stated. Here is what happens when you focus stack with a fixed amount of tilt applied. The red line is the plane of focus. Mr. Merklinger didn't add the DoF wedge to his movie, but you can image the cone spread out behind and in front of the red line:

[Thank you, Harold!]
If you can imagine a scene where you want everything in focus top to bottom in the frame, then you just need to know in what increments to move the lens / rotate the helical in order to rotate the DoF wedge down through the image in appropriate steps. I don't know what those increments are off the top of my head, but I suppose with a bit of testing you could figure that out.* It is the fact that your plane of focus rotates around the hinge line that made me think this would be difficult to manage in the field. But in most cases you don't have to worry about that. Again, all this is in the box of landscape applications, not the more precise situations required in studio and/or up close.

Another thing I thought would be difficult but after a quick test found not to be a problem: All the helical-based technical cameras tilt behind the lens to some degree. That changes framing as you add tilt. I mistakenly thought it would also change framing with changes in focus, but that is not the case. Magnification changes as you rack focus in and out, but that is no different than without tilt and the software deals with that just fine.

Dave

*Edit Note: Because the DoF wedge gets smaller (more narrow) as you increase the degrees of tilt, you would need to know how wide the DoF wedge is given the angle of tilt applied. That makes things a bit more complicated but not "extremely difficult."

 

[gerald.d]

You wouldn't, of course.

Thanks for clarifying that. How bizarre that you felt the need to use a bold font to stress a scenario that you admit is “of course” something you would never actually have a need for.


Nothing else that follows is relevant to creating actual photographs of actual subjects, least of all the barely veiled insult that I have no comprehension of this subject from a purely technical perspective.

You may well think that’s where you come in, but from where I’m standing, it’s where you’re shown the door.

 

[gerald.d]

I feel responsible for starting this debate with my not-too-well-thought-out comment: "Tilt would be extremely difficult to focus stack, but I can’t imagine the need to use both. You would just focus stack without tilt, or get everything sharp you want with tilt and not focus stack."

Certainly no need for any apologies, because you are totally correct - the scenarios where one would actually want to be able to focus stack a tilted focal plane are very few and far between.

It really does only make sense when one wants to take a photo of, and control the focus fall off from, a subject as I described - typically something with a depth small compared to the width and height dimensions that is presented at an angle to the sensor and against which you establish your plane of focus.

If you are shooting a cube, at any angle of presentation to the sensor, there is little to no benefit in making the effort to align your plane of focus (and it IS a plane at the center of the wedge) to one of the cube surfaces, and then stacking as you move that plane perpendicular to itself. It is far far simpler (and just as efficient) to discard movements, focus on the closest point of the cube, and then rack and shoot until you land on the furthest point. (Obviously if you desire some kind of “artistic” control over the focus in relation to an orientation of the subject cube, this is not the case.)

There are two potential benefits in play to stacking images with lens movements.

Firstly, you have that “artistic” control over focus fall off from the subject.

But the second potential benefit is more practical. It can introduce significant efficiency benefits, because it can dramatically reduce the number of images required in the stack to get your subject nailed. In some instances by around a factor of 10.

At the end of the day, what we are missing here is “why” the OP felt the need to ask the question in the first place.

But I stand by my original two sentence response. It IS a fact that as you change focus distance, you need to adjust your lens tilt and swing to keep the lens plane of focus parallel. And there are no diagrams that can dispute that. Any attempt to do so is pure obfuscation.

 

[MGrayson]

But I stand by my original two sentence response. It IS a fact that as you change focus distance, you need to adjust your lens tilt and swing to keep the lens plane of focus parallel. And there are no diagrams that can dispute that. Any attempt to do so is pure obfuscation.

And that statement is 100% correct, and no diagram is disputing that. In a landscape, however, one does not want parallel planes of focus unless one has no tilt at all. You are not going to capture a mountain top by moving the plane of focus up 5,000 feet. You rotate the wedge of focus around the hinge, and that stacks nicely.

 

[pegelli]

At the end of the day, what we are missing here is “why” the OP felt the need to ask the question in the first place.

I'm speculating a bit but my reading between the lines of the question was if you could do the focus stacking with the in-camera automated focus step feature using a tech cam or CANON tilt-shift lens. Obviously this can only be achieved with an AF lens, which neither a tech-cam nor the CANON TS-E is.

 

[alajuela]

Shifts and rise fall won’t have any effect and focus stacking can be done as normal. Tilt would be extremely difficult to focus stack, but I can’t imagine the need to use both. You would just focus stack without tilt, or get everything sharp you want with tilt and not focus stack.

On technical cameras with helical systems, manual focus stacking is a breeze. Bellows systems require a bit more care but as long as you can rack focus in small enough increments it’s also pretty easy.

You just need to move the lens in (or out) the same linear distance each time. For example, most of Alpa’s helicals have a rotation pitch of ~0.03mm. I find at f/11, rotating the helical 5 degrees between images gets me acceptable results. That’s a lens movement of 0.15mm between shots. This is regardless of focal length or focus distance on a 54x40 sensor. Super easy to do with the HPF rings installed.

Dave

On a bellows camera you have the option of moving the rear standard, This is much preferred to moving the front standard as the perspective does not change moving the ear standard and does moving the front

Thanks

Phil