tilt/shift lens vs focus stacking

Discussion in 'Nikon' started by kylebybee, Nov 12, 2013.

  1. My primary focus (no pun intended) on my photography is landscape. I currently only shoot with the D7000 and am looking to get another body in the next few months. I haven't tried either yet, but was interested in making the whole frame in focus using either tilt/shift lens or focus staking technique. I'm assuming that both would give me the approximate same outcome, but was curious as to pros/cons without stating the obvious of one taken in camera vs post processing and time to take the shot. I also know that the Nikon tilt shift lenses won't work well on the DX bodies, but maybe the Schneider lenses would?
    The caption on the picture should read D7000 not D700. Oops!
  2. Tilt won't help you at all with a scene like that. What it is useful for is when the subject is approximately a plane but is not perpendicular to the optical axis. E.g. when photographing ice or plants on the ground at an angle. Sometimes tilt can help you achieve a shot where everything or most of the interesting parts are in focus in that case without needing to use more than one exposure to achieve the final result. But a lot of the time the subject isn't a tilted plane and so stopping down and/or focus stacking may still be needed, but then the tilt can help reduce the number of exposures needed in the focus stacking process. Focus stacking does produce artifacts which you need to deal with usually by manually editing the image in photoshop, and a lot of the time it may not be possible to get a completely artifact free result. Sometimes also the scene doesn't stay still long enough to capture a sequence of shots, this is true of many macro situations. But for landscape I guess this isn't as much of a problem though three movement can create problems.
    I use tilt where it helps, usually it is for ice and flower shots; in particular it's possible to get the sea ice into focus from near range to infinity, also it can be used to get two flowers into focus even with a wide aperture so as to create nice out of focus blur. Focus stacking (either by hand drawn layer masks or using automated software) can be used to solve some situations where tilt won't help, for sure.
    Both techniques take a lot of practice and skill to perfect.
    Here is a shot which was made with a 220mm lens at a fairly small aperture (f/11 if I recall correclty); the subject (leaves on water) could have been gotten in focus using tilt if I had had a 200mm tilt lens but I don't. Actually I might be able to do it with a bellows but it's difficult as the 200mm lenses should be supported from the lens mount to be safe, and the bellows needs separate support so it gets complicated. But with tilt the shutter speed could be increased, leading to increased sharpness in the leaves that are floating with the water. Here focus stacking would not help at all as the movement of the leaves would put them completely in a different position from shot to shot.
    I posted that to illustrate the problematics of tilt and stacking. However, some cases are easy to handle with tilt:
    to higher resolution than would be possible by stopping down instead of tilting. I have some focus stacked landscapes but unfortunately I haven't posted them in a gallery. In general I think for landscapes where you have trees at different distances, focus stacking with manual blending is what would work usually, usually tilt would not be the right tool to use for that.
    mag_miksch likes this.
  3. I believe there are issues with shifting some lenses that clash with the camera housing on smaller cameras, but that may not be much of an issue for tilt. Live view is a significant help here, though a really solid tripod helps when adjusting things. I still wish (and will implement if I ever get to hack the BIOS) that there was a way to split live view into quadrants, positioned independently - that would allow my ideal solution of setting the tilt approximately and then waving the camera around to achieve focus. This is how I've used tilt lenses in the past - but on a camera with a pentamirror, it's a bit painful because of dimness.

    As Ilkka says, tilt lenses don't give infinite DoF, they just rotate the focal plane. This is often still very useful. Focus stacking is slow to expose - though some software will automatically rack the lens around for you, and in macro distances you can use a macro rail; shift lenses have a single exposure, but setting them up takes a bit longer.

    What I'd really like is a digital sensor with rear movements - it actually wouldn't be that hard to engineer, especially on a mirrorless system. Oh well.
  4. "I'm assuming that both would give me the approximate same outcome..." - Wrong assumption! Shifting a lens does nothing to alter the depth-of-field or plane of focus, so we're only talking about the tilt function.
    Tilting a lens will tilt or swing the plane of focus (obviously), and may, or may not, be able to accommodate all the parts of a subject that you want to be in focus. In a landscape, that's not always the case, because although a forward tilt of the lens may bring nearly everything at ground level into focus, vertical objects, such as trees or mountains, will tend to be progressively more defocused toward the top of the frame. Consequently IMHO, a tilt lens is often of very little use for lanscapes.
    Focus stacking is an entirely different beast, and its main purpose is to increase depth-of-field beyond what can be achieved through the use of aperture control alone. It only uses planes of focus parallel to the camera back. It also requires multiple exposures and hence a steady tripod and a non-moving subject, whereas a tilt/shift lens is a single-shot solution that can - if necessary - be used handheld. AFAIK, most focus stacking software can't be easily used to achieve the same effect as a tilted focal plane, although in theory it could.
    Edit: WRT any difference between a PC-E Nikkor and Schneider's tilt/shift offerings; as far as use on the DX format goes, there is no difference, because both makes of lens work in exactly the same way. In any case I'm not sure why you'd say they're of limited use. The maximum (say) 11mm of shift will actually have more apparent effect on DX than FX, due to the shift being a greater fraction of the frame width or height. The massive image circle also allows the DX frame to be shifted to the full mechanical extent of the lens without worrying about vignetting, and that's not the case with FX.
    Re-edit: Ah! I just realised that you're probably talking about a potential collision between the lens and the camera body. In that case I'm not sure whether Schneider's lenses are any more compact when shifted. I have an old Schneider 35mm PA-Curtagon, and the only thing I can comment on is that it's optically inferior to a 35mm PC-Nikkor of similar vintage.
  5. Kyle, I want to say first that your photo looks very nice the way that it is. You focused on the details of the objects close to you and let the distant objects blur slightly. This creates a very natural way of viewing the scene that appeals to our eyes. I would not bother with a tilt/shift lens or focus stacking in this case.
    I'm assuming that both would give me the approximate same outcome​
    Tilt works best when you are trying to focus on a plane that is not parallel to the plane of the camera's sensor. I have included an example below. This wouldn't work in the scene that you posted above. You could use focus stacking with that image, but I don't feel that it would improve your image.
    curious as to pros/cons​
    First or all, realize that it's not a direct comparison. The "tilt" movement of can give the impression of great depth of field if your subjects are arranged on a plane. But it's not really extending depth of field; it's just changing the orientation of the plane of focus. Further, the "shift" movement has no impact whatsoever on DOF, perceived or otherwise. It's used primarily for other reasons entirely. One reason is keeping parallel lines parallel as when looking upward at a building or a group of trees. The shift movement can keep trees, buildings, and other vertical lines from looking as though they are tipping over.

    Tilt be used to create special selective focus effects. Shift can be used as an aid to creating panoramic images in some cases.

    Focus stacking combines multiple images taken with the focus set to different distances.

    Focus Stacking Pros:
    • Creates a unique effect that can't be replicated.

    Focus Stacking Cons:
    • Requires specialized software and the skill to use to.
    • Multiple shots are not an option when there is movement in the frame.
    • Extra processing time and storage space required.
    • Tripod required.

    Tilt/Shift Lens Pros:
    • Can be used to manage parallel lines.
    • Can be used to approximate infinite depth of field under certain conditions.
    • Focus plane alterations work within a single image - stacking not required.
    • Can be used to assist in the creation of panoramic shots.
    • Can be used to create special effects with selective focus.

    Tilt/Shift Lens Cons
    • Expensive.
    • Slower, more methodical shooting style (might also be considered a 'pro').
    • Focusing effectively with tilt requires some practice.
    • Additional gear to carry.
    • Manual focus only - Live View with magnification is highly recommended for focusing.
    • Tripod recommended for best results.
  6. Thank you for the advice everyone. The question about tilt shift working on DX comes largely to a review/post from Shun Cheung, where he said there was a problem with the movement of the lens hitting the part of the camera where the prism is stored, that may have been with only the 24mm though. I do have a tipod and it sounds like the focus staking will be more work than I care to mess with for now. Has anyone used the PC-E 45mm for landscape?
  7. I use a medium format lens with an adapter to set the plane of focus. having focus peaking is very useful here - I get that from Magic Lantern on canon. The newer models have that build in (I use 6D).
  8. I'd have thought the FX Z6/7 has such a shallow flange depth and wide throat that you could use one as a tilt/shift 'back' without much modification at all.

    The DX Z50 with the same flange depth and big throat but a smaller, centered sensor would be even easier to work with and allow more movement.

    Infact, a few degrees of tilt on the Z50 with a 24mm (EQ) would work very well for landscape near-to-far use. If an FX lens is used the image circle would allow a-lot of lateral movement too without vignetting.

    The 'modern' Nikon DSLRs have in-body focus stacking (for the taking stage at least) so object movement of slowish things like clouds should be minimised.
  9. Were you thinking of strapping one to the back of a 5x4 camera Mike?

    Because with no dedicated Z mount tilt/shift lenses on Nikon's 'roadmap', that might be the only way to make use of the wider mount opening.
  10. Yup. I've got an old mono-rail somewhere in bits....;)

    ...and of course if you use a DX F mount lens on a Z mount body you've got the FTZ 'depth' to play with and the extra coverage for a novel T&S adapter....:cool:
  11. Experiment and try new things like focus stacking etc. If you want to keep it simple, shoot at f22 and focus on different items in the frame and see what images you think look the best.
  12. According to the Scheimpflug Principle, the plane of sharpest focus tilts roughly twice the angle the lens is tilted. Depth of field limits tilt proportionately, so objects above and below the plane of sharpest focus are increasingly affected. In practice, you find that the angle setting is very sensitive. My Sinar view camera has indicators which help get the best effect for tabletops and landscapes.

    If you are taking "calendar" photos with flowers in the foreground and mountains in the distance, you can get both in sharp focus. However trees and buildings in the mid-ground may be out of focus at the top.

    In its simplest form, focus stacking moves the plane of sharpest focus, keeping it parallel to the film or sensor. You could combine it with the Scheimpflug effect, but by trial and error. Focus stacking software automates the rendering of multiple images by selectively masking areas of sharpest focus (edge contrast) in each layer. Sensitivity of the masking can be adjusted, but is usually applied only to larger objects. It can be fine-tuned to accommodate blowing grass, for example, or moving vehicles or persons. It's better to wait for a pause in the wind or moving things to exit the field of view.

    In a modern, FF digital camera, nothing is in focus at f/22. Many lenses are diffraction-limited between f/5.6 and f/8. The better the lens and greater acuity of the sensor, the lower the limit.
    Last edited: Jun 10, 2020
    mike_halliwell likes this.
  13. Just a short note: I use a D7100 and before that a D3200 and have found that the 85/2.8D tilt shift lens works pretty well on these. I rarely use the tilt, but at least experimentally it seemed quite all right. I have used it in macro work, where it definitely helped, but this lens is pretty hard to handle without a tripod, so I don't use it as much as others for my more common hobby of chasing insects around the back yard.

    I regularly use a shift-only 35/2.8 PC as my normal lens, and that also works just fine, with the shift used occasionally, more often to shift the camera out of reflected views than to correct perspective.

    There may be some optical disadvantages to using these lenses on DX format, but there is no physical conflict.

    My interest in the 85 is more at the macro end than the scenic, but I would note that it goes to an insanely small aperture, with less noticeable diffraction than one might expect.

    Given that those things are usually breathtakingly expensive, it might be worth while to try renting something to see whether it meets your needs.
  14. You can pick up the non-E version of the Nikkor 85mm PC for about £600 these days.....;)

    ...and yes it is a handful off tripod, esp. if tilt is involved.

    I found my 28mm f3.5 PC smeared at the 'top' on full shift even on my DX D300.
  15. Just to say... huh? Surely the relationship between angles depends on the focus distance. That's one reason I tend not to try to work it out, and just use focus peaking (or at least focus confirmation), and want the four-way version of split-location live view (although the 2-way helps a little), which would make tripod use a lot less tedious. Canon at least used to have a mode where all the "in focus" AF points would light up, which was useful for placing the image plane; Nikon won't do that with a manual lens (for... reasons?) but might if it's chipped and thinks it's in the right AF mode?

    I used my 65mm Hartblei (which is originally a medium format lens with a dedicated mount adaptor) a moderate amount on my Canon, but still haven't got around to blacking the mount I substituted on it when I went Nikon, which puts me off. I also have a 35mm Kiev tilt-shift, with locked tilt/shift angles and an iffy mount that I worry might hook on the aperture lever. I was thinking of getting rid of them, but then did a test and found that actually neither is all that bad optically (for their cost). I did look at the 24mm Samyang, but the corners were clearly iffy; the 19mm is on my "things to save for" list.

    I suspect the original poster has taken the time since 2013 to form an independent opinion.
  16. When I bought by PC-E 24, the camera body was a D300. A very close fit and difficult to use the controls. As I recall, the lens would not fit on a D7x00. For practical use, shifting for perspective control is easier to perform in post. Many of my images using PC shift used the outer image circle where IQ deteriorates and using a wider angle lens + post achieved better results. Tilt for plane of focus requires a tripod and focus aids although I had a default tilt angle for most near-far compositions. For landscapes, I prefer DOF or stacking these days. There are a few 35mm applications where tilt achieves better images, but my hope for a 35mm view camera kit didn't work well in practice. It's just not as enjoyable as getting under the hood of a large format view camera.
  17. The Sinar method of tilt calculation is cited in this relic from the PNET archives...
    Sinar Tilt Calculator Help

    The ground glass is marked about 1/3 from each edge of the screen. You focus on an object near the upper line (near), set a sliding scale on the focusing knob, focus on an object near the lower line (far), and read the estimate angle from the scale on the focusing knob. That's often pretty close if you stop down, but it may take a couple of iterations to nail it. It helps that Sinar swing and tilt adjustments are yaw-free - optically centered on the lens. It helps to use a 5x loupe on the ground glass.

    Since focusing points for a mirrorless camera are embedded in the sensor, they work equally well with manual lenses. With my Sony, the image can be magnified 6-12x, and the focusing position can be adjusted with the joy stick. Needless to say the settings are fussy and time-consuming, strictly on a tripod if you want any sort of precision.

    Actually, I wouldn't mind working at a 19th century pace with 21st century equipment. However the Scheimpflug Principle works only if your subjects are in or close to a single plane. Focus stacking is faster snd handles as many planes as you wish. Two or three points are good for landscapes, but you might use dozens in a macro of a large 3D object (e.g., a sculpture). You can manually focus on one plane or object at a time, much like the Sinar method, or use a stepping servo to make successive changes to the distance or focusing helix.
  18. For what it's worth, I think there's a difference between tripod-mounted image composition using a 10x8 sheet of film that'll cost you £5 to expose and develop using an optical work flow, and a DSLR capture. You absolutely can still mount everything on a tripod, carefully examine all corners of your frame by panning around at 100% zoom (although in my experience this is much more likely to shift the tripod than the actual lens adjustments are, which is why I wanted the split live view...), get out your tilt calculator, shoot at f/64 with a multi-second exposure, and so on. But you can also set approximately the tilt that you might want, use focus highlights (or multi-point AF confirmation), and wave the camera around by hand until the focal plane hits the things you care about - then crop the result in post. The latter wastes more pixels, but it does get you ten times the number of images. An actual 5x4 has been vaguely on my wish list for a while, but I can't really imagine not trying to focus one (at least approximately) by grabbing one of the standards and wiggling it about. Lensbaby have an entire line based on the premise that this kind of works if it doesn't cost you a fortune in film to mess things up.

    It's slightly unfortunate that all the PC lenses are manual focus. There's no real reason they have to be, if you choose the right AF point. I really think there's a chance for a functional step in photography by offering autofocus and automatic focal plane orientation - I still see a number of TV shows where the focal plane is clearly not parallel to the sensor, and I don't think they're all applying a post-processing effect to emulate it. I've long had the thought that in live view you can approximate this by allowing the shutter and sensor to tilt (mostly back, because otherwise you'll run out of image circle even if you don't hit anything); involving a mirror in that is an exercise in insanity, but it ought to be an option for the Z series.

    My faith in focus stacking except for static subjects and using a tripod (which hurts some of the sub-pixel resolution enhancement) has taken a hit from seeing rolling shutter issues, so I'll be a little unsure until we get proper global shutters that don't impact image quality too much. Capturing a 3D light field in one step solves some problems (partly, if you're Canon and have dual pixels), but current solutions have tended to get too hit by static image quality issues to justify the expense. Maybe in the future. For now, I think a decent PC lens still has its place, even in casual shooting.
  19. Name one!
    Every shift lens I've looked at is dogged by inferior IQ. Due either to sharpness being sacrificed to coverage, or to colour fringing from presenting an angle to sensor micro-lenses/filters.

    I really don't know what the optimum solution would be. For me, and in monochrome, it's to hang onto my 5x4 kit and use it sparingly where Tilt/Shift is really needed.

    Which reminds me. There's a Victorian building facade that deserves documenting. It's trapped between two totally unsympathetic modern buildings in a narrow street that doesn't allow backing the camera away from it more than a few metres. Still scratching my head over that one as well.
  20. I've heard only good things about the 19mm, but can't say I've used it (donations gratefully accepted). I agree that the quality/price trade-off has been questionable on many cases - not helped by Nikon's fixed orientation arrangement; my first T/S, back to my Canon days, was the Hartblei, and it's fully flexible like the 19mm. I get the impression that some larger format lenses are a bit iffy too (it helps that the relative aperture tends to be slow, although at some point diffraction kicks in), but I hope to try at some point. Time and money...

    Unless you can get above ground level in the building opposite and use an ultrawide, could you talk to the neighbors and get permission to do a vertical stitch with a drone?
    Ed_Ingold likes this.

Share This Page