2
\$\begingroup\$

When tilting a tilt-shift lens, does the framing actually change or does it just work similar to adjusting the focus ring (for a particular area of the image)? I have tried to search for videos where people tilt their lenses, but all I can find are before/after photos and not an actual video of tilting being (gradually) applied.

I want to try to apply the effect gradually to a time lapse (like this video), but I am not sure if the framing will be changed when I tilt the lens (the video in the link has added the effect in post production, so it does not prove anything).

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Quick answer: Yes there is a small frame change, as well as when you focus (this is called breathing), and small vignetting variations. The magnitude of change depends on the lens you're using. I got the 17 and 24 TSE Canon to tell. However, you can compensate the frame shift by reframing your video. Good luck \$\endgroup\$
    – Soleil
    Commented Dec 10, 2017 at 0:17

2 Answers 2

3
\$\begingroup\$

Very good question.

Short answer not really, framing doesn't change (for any tilt angle) if the tilt-shift lens uses axis tilt, but does change (for large tilt angles) if the lens uses base tilt (or some other kind of tilt).

So the answer depends on the particular lens you are interested in.

The longer answer is that even with axis tilt framing might change very slightly, as you might refocus, just like it can change when you adjust the focus in normal circumstances, but you already seem to consider that as not changing framing, so you're good.

Here's a video, at 4:14.

It's about a view camera, but tilt-shift lenses work the same way.

The Canon TS-E 90mm f/2.8 seems to have axis tilt, and functions like I described above, while the Canon TS-E 24mm f/3.5L seems to do some kind of asymmetrical tilt (neither axis nor base tilt). Since the wide angle is probably a retrofocus design, it's difficult to determine the effect of tilt on framing just from product pictures.

Looking at Nikon it also seems that the wider angle tilt-shifts use weird tilts while the longer focal length use axis tilt.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for clarifying this. I was looking at the Samyang TS 24mm, as it seems to be a decent lens for a budget more in-line with someone who lacks experience with these kinds of lenses. \$\endgroup\$
    – Jorn
    Commented Dec 9, 2017 at 18:53
1
\$\begingroup\$

I see nothing in this video that supports the theory that this is a post production effect and not real video from a camera with a Tilt-Shift Lens.

There's a distinction between Tilt-Shift (photo) and Cine-Tilt (video) lenses. For photos a bit of adjusting between each shot can be tolerated for a cost savings, in cinematography adjustments to one part of the lens (focus, tilt, zoom) should have no effect on the other adjustments; these lenses are far more expensive.

Look at: this video - the framing is altered by sliding, later (like at 0:47) the 'rolling in and out' is caused by changing the focus slightly; in neither of those cases, nor anywhere else in the video, does the Videographer adjust the tilt shift mechanism while taking the shot - he locks it and shoots.

Here is a video where zoom is used to alter the framing at the beginning. In this shot the Videographer has locked the mechanism and then mounted the camera on the dash (or hood?) of a vehicle and drives through the streets (speeding up in post), again the mechanism isn't adjusted while shooting.

Here's another video with a lot of movement, notice how the 'in-focus area' is the same whether moving or panning: drastic change in the framing but the mechanism is locked.

In this video you can see that if you adjust the mechanism you must reframe when Shifting or Tilting. There's also this tilt.

With a Cine-tilt lens like the Schneider-Kreuznach Cine-Tilt it is designed not to move the framing while shooting video - A photographic lens is not always designed to limit focus or zoom breathing and wobbling because each image can be adjusted prior to pressing the shutter.

A higher priced video lens can not be recomposed for each frame and ought to allow adjustments to be made without introducing unwanted changes in the other aspects of the image that are not being adjusted (example: adjusting focus or zoom shouldn't adjust the other, similarly with a Cine-Tilt the framing should be rock steady - you get what you pay for).

Here's a panning, adjusting the tilt-shift and attempting to hold the framing while moving around tutorial (a more exact answer to your question). In part one he talks at length about his setup and using tilt shift to create enormous resolution photos which can be zoomed and panned in post. In part two he talks about adjusting the mechanism while live shooting (this answers your question).

I am searching for another example when the fellow walks around with his camera near a fence (and a playground IIRC) and makes adjustments while hand holding the camera, shooting live video. I can't find the link but I'll keep looking and return to edit this answer later today.

As a point of interest, Keith Loutit made a video called The City of Samba - it's a great use of tilt shift but you can see he wasn't brave enough to adjust while he shot, it's extremely difficult.


Concrete example:

See this example of the Schneider-Kreuznach Cine-Tilt (watch the lower right corner) lens compared with the Canon Tilt-Shift photography lens.

With the cinematography lens large adjustments are smooth and flat-field, with the photography Tilt-Shift lens, when used for video, the framing bends and twists enough to make you seasick.

To use a photographic tilt lens for video you need to adjust slowly and gently, hoping that you can unwarp the video in post. Film makers don't have time and money to waste doing that, to get a marginal result.

\$\endgroup\$
5
  • \$\begingroup\$ "In this video you can see that if you adjust the mechanism you must reframe". He is shifting, not tilting, almost the whole video. The question was about adjusting tilt, not shift. When he is tilting he is not reframing. When he is tilting the framing changes just ever so slightly because the lens is not using axis tilt, but still it doesn't change so much he needs to reframe. \$\endgroup\$ Commented Dec 8, 2017 at 17:09
  • \$\begingroup\$ Good catch, that was the shift example, I added the two missing tilt examples to the end of the sentence (same video). \$\endgroup\$
    – Rob
    Commented Dec 8, 2017 at 17:59
  • \$\begingroup\$ of course it is entirely postprocessing, you would need a car-sized lens aperture to achieve this kind of DOF effects \$\endgroup\$
    – szulat
    Commented Dec 8, 2017 at 19:40
  • \$\begingroup\$ Thanks for the answer. The concrete examples were very useful. I guess I would ideally use a cine-tilt lens, but with the current budget the Samyang TS 24mm is what I have my eyes on. Regarding the example video I linked to, PetaPixel wrote "He [Keith Loutit] continues on to explain that the technique is often applied in post (so not true tilt-shift)", so I think it's safe to say that it was done in post. \$\endgroup\$
    – Jorn
    Commented Dec 9, 2017 at 19:07
  • \$\begingroup\$ @Jorn - He says "the technique is often applied in post", that doesn't mean he does that. In his interview at video-creativity.blogspot.ca/2009/09/… YouTube Creator's Corner asks him: "Tell us a fact about this video that no one would know by looking at it.". Keith Loutit: "That it's real :D.". -- That means he doesn't Photoshop, I'll stick with my opening line: "There's nothing in that video that supports the theory that it was done in post". The Xenon Cine-Tilt lenses start at $5.5K. You could also use split or especially strip diopters. \$\endgroup\$
    – Rob
    Commented Dec 10, 2017 at 1:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.