| Author | Thread |
|
|
10/05/2005 03:57:38 AM · #1 |
I'm pretty sure Robt will know, but anyone else's experience would be handy too:
What difference does focal length make (aside from the obvious) when using a tilt-shift lens? Does the equivalent amount of shift or degree of tilt produce a more dramatic effect at shorter, or longer focal lengths? Which is more useful generally, a short TS lens or a long one, and which is of more use in what situations?
|
|
|
|
10/05/2005 05:54:55 AM · #2 |
| Not sure if it will help answer your questions or not, but have you played around with the Tilt Lens Analyzer at all? (It doesn't take into account shifting, but it is still kinda cool.) |
|
|
|
10/05/2005 06:29:08 AM · #3 |
I have not used tilt-shift lenses on SLR's, but have extensively used a 4x5 camera, which has even more flexibility in terms of movements.
Typically, we would use camera movements in the studio to correct perspective on a product and to move the focal plane so that the label/name of the product was the sharpest. We normally would use a 210mm or 150mm lens for this. (roughly equivalent to a "normal" lens) Usually, people paying for product shots don't want any of the distortion associated with wide angle lenses.
For architectural interiors (and I'm sure Robert has more experience than me in this area), a wide angle lens (65mm or 90mm) is often necessary to capture the whole space. The camera movements were used not only to correct perspective, but, could also be used to take a shot of a mirror that looked to be straight on without the camera being in the reflection.
As with any piece of equipment, its usefulness to you will depend on what you want to use it for.
|
|
|
|
10/05/2005 07:25:36 AM · #4 |
ummmmmm.....
This is a VERY complex subject, sheesh.
There's a thing called the "Scheimpflug rule" that is second nature to working pros with view cameras. Basically, the rule states "if the plane of the subject, the plane of the lens, and the plane of the film all intersect at the same point, every point along the object plane will be in sharp focus."
To look at this in the simplest way, imagine your task is to photograph a 9x12 foot oriental carpet. You set up a shot looking down the length of the carpet, with a LOT of "foreground"; the carpet is keystoning away from you (it is "receding perspective", smaller at the far end than the near end) and the floor/wall juncture is your "horizon line" in the very top of the picture — the image is nearly all carpet.
Now, in this shot the carpet starts 3 feet from the camera and extends out to 15 feet away. You need tack-sharp DOF from 3 feet to 15 feet to satisfy your client. You can stop WAY down to get this DOF but when you do circle of confusion sets in for such small apertures and you don't really get that tack-sharp quality you want.
Now, DOF is always measured along the axis of the lens; if a lens has at a given f/stop and a particular focusing distance a DOF of 8 feet, it means that the 8 feet of sharp focus is measured on an imaginary line that extends right out the barrel of the lens, so to speak. The PLANE of sharp focus is at right angles to the axis of the lens. So if the camera is at a truly vertical orientation, the plane of focus is like a transparent, vertical wall. Anything that breaks the surface of that wall will be in focus at that point along its length that is actually intersected by the imaginary plane of sharp focus.
In the given example, the camera is pointed "down" at, say, 45 degrees to see the entire carpet from 3 feet to 15 feet. So the plane of focus is also at 45 degrees, and it intersects the plane of the carpet. The DOF is along the axis of the lens, so it is extending both above and beneath the carpet (into the empty space above the carpet and into the invisible space beneath the floor the carpet is on) where it's not doing you much good at all.
Using the Scheimpflug rule, if you tilt the lens so that the plane of the lens, the plane of the film/sensor, and the plane of the carpet all intersect at the same point, then the plane of focus is aligned with the plane of the carpet, and every point on the carpet will be in perfect focus regardless of the aperture used. The DOF is at right angles to this plane, so it now extends at right angles above and beneath the carpet, rather than at 45 degrees to the carpet.
To extend this a little further, let's suppose you were focused at 6 feet. Remember, the camera is pointed down at 45 degrees. Imagine that you placed on the carpet at the 6-foot mark a yardstick angled at 45 degrees so it is at right angles to the axis of the lens. Every mark on that yardstick will be in perfect focus, because it represents the PLANE of focus at right angles to the lens. The carpet, however, is less in focus the further in front of or behind this plane it is.
After applying the Scheimpflug rule by tilting the lens, you will find that the plane of focus and the plane of the carpet coinicde, and now the markings on the yardstick are drifing further out of focus the further from the surface the carpet they are.
This is what a tilt lens is for; to adjust the position of the plane of sharpest focus within the image. While you're doing that, it also makes dramatic changes in the orinetation of the DOF, so that you can get some very strange and unnatural results. If a PERSON is standing on the carpet, the shoes will be in focus, the ankles less so, the shins even less so, and the calves beginnning to look blurred, with the knees totally OOF — something like that.
**********
Now, back to the original question: "does the degree of shift produce a more dramatic effect at shorter or longer focal lengths?" The short answer is, "There's nothing "dramatic" about this effect"; it's a shifting in focus, not a shifting in perspective, so geometrically speaking the images would look the same with or without tilt. However, in any given relationship of subject plane to film plane, the longer the focal length of the lens the more tilt is required to get the correct convergence. You can draw this out on paper to prove it to yourself, or take my word for it.
Related to this is the fact that basically, the wider the lens the less tilt you have available to you before you run out of image to play with. To understand why, think about how a given lens is "wider" on a full-frame camera than it is on a 1.6 crop camera like mine. When you think about that, you'll realize that the lens is projecting a larger image than my sensor is covering, so my sensor is in effect "cropping" a smaller rectangle out of that larger image, ok?
The way it works is, the lens throwns a circular image and the sensor/film crops a rectangle out of that. If a lens is "optimized" for a certain format, say for 35mm film, it throws an image circle that's barely big enough to fit a 35mm film in the center of without vignetting. When the same lens, optimized for 35mm film, is used on my 20D, the image circle is larger than the sensor, so in theory I could place the sensor anywhere within that image circle and get a good shot.
In fact, this is exactly what a "shift" lens does (and most, if not all, "tilt" lenses also "shift"; in fact, there were originally shift lenses that did not tilt — tilting is much more complext to engineer than shifting); the shift lens allows you to position the sensor anywhere within the image circle thrown by the lens.
In any case, back to tilting; basically, the longer the focal length of the lens the easier it is to throw an oversize image circle. With extreme wide angle lenses, they have a hard enough time engineering those to cover the actual sensor area without distortion, let alone any excess area. And the amount of tilt that you can use depends on how much excess area is available for your use. It also depends on the physical limitations of the lens mount, the mirror placement, the shape of the "box" inside the camera and how much the lens projects into that box, etc, but that's a whole other issue.
*******
Whoo, I'm losing track here, writing this on the fly. Bottom line; the shorter the focal length, the less tilt you need to cover the same object plane edge-to-edge, but the less tilt you have available to you (in theory) so it tends to average out.
Robt.
|
|
|
|
10/05/2005 08:35:32 AM · #5 |
As an addendum, here's a scanned reproduction of an older print of mine where tilt was used to maximize DPF along a a dominant plane; Tijuana Border Crossing, 1976 — SinarP 4x5 camera, Schneider Super-Angulon 90mm lens, Plus-X film, f/22 at 1/4 second. (90mm on 4x5 is about the same as 30mm on a full-frame dSLR or 20mm on my 20D. The scans nowhere near as sharp as the pringt, but you can see the zone of focus; look at the tree upper right.
Robt.
|
|
|
|
10/05/2005 09:17:45 AM · #6 |
Some examples from my portfolio. Ditto everything that Robert said. He knows his stuff. As far as my photos, I'm still learning but the lens is great!
  |
|
|
|
10/05/2005 09:24:11 AM · #7 |
Originally posted by rikki11:
Some examples from my portfolio. Ditto everything that Robert said. He knows his stuff. As far as my photos, I'm still learning but the lens is great! |
But you're showing us shift, not tilt, right?
R.
|
|
|
|
10/05/2005 09:27:40 AM · #8 |
|
|
|
10/05/2005 09:42:20 AM · #9 |
On a 35mm camera the 24mm canon tilt-shift is said to have an effective focal length of 21mm when tilted. If your doing interiors ot landsapes the wider the better. I hear the 90 TS works great for portraits. That's about all the detail I'll be going into after Robert's post. ;o)
|
|
|
|
10/05/2005 09:44:20 AM · #10 |
| On reading Bear's explanation my eyes began to shift around and my head tilted to the side. I eventually had to slap myself out of the trance..... |
|
|
|
10/05/2005 09:46:10 AM · #11 |
Originally posted by jbsmithana:
On reading Bear's explanation my eyes began to shift around and my head tilted to the side. I eventually had to slap myself out of the trance..... |
I DID warn you it was complex... And what's worse, that's a ridiculously simplified overview.
Robt.
|
|
|
|
10/05/2005 09:48:41 AM · #12 |
Originally posted by bear_music:
And what's worse, that's a ridiculously simplified overview.
Robt. |
I agree! I thought you did a pretty good job explaining it.
|
|
|
|
10/05/2005 09:50:51 AM · #13 |
Originally posted by MeThoS:
On a 35mm camera the 24mm canon tilt-shift is said to have an effective focal length of 21mm when tilted. If your doing interiors ot landsapes the wider the better. I hear the 90 TS works great for portraits. That's about all the detail I'll be going into after Robert's post. ;o) |
Yes to the wider angle; you're essentially using a diagonal and it stretches things a tad. Not a lot though.
Using a tilt lens for portraits can be VERY effective, yes, but the model has to sit damned still and the camera has to be on a tripod. Here's the principle; we all like portraits where everything-but-the-face is going soft, and we control our DOF religiously to attain that. But if the subject is looking anywhere other than straight at the camera, the plane of the face is not parallel to the plane of the sensor/film, so you have to stop down quite a bit to keep the near ear and the far eye both in focus. This will render the BG sharper than you wish, say.
So you tilt the lens to bring the defined plane of the face into complete focus, and then stop down just enough that the tip of the nose and the back of the near ear are in focus, and you got it.
Robt.
|
|
|
|
10/05/2005 09:56:14 AM · #14 |
Big thanks for the tilt-shift explanation, Rob!
Been looking for some info about tilt-shift lenses for a while.
By the way. Is it possible to make a tilt-shift adapter to fit a normal lens? I have a Soligor 28mm f/2.8 in T-Mount. And the T-Mount has 55mm of distance between the film plane and the rear element. Wondering if there might be tilt-shift adapter for T-mount lenses..
Message edited by author 2005-10-05 14:01:15. |
|
|
|
10/05/2005 10:01:37 AM · #15 |
Just when I think I am getting somewhat out of the dark photography confusion I've been in since I bought my DSLR, there's another eclipse!
Can someone please pass the Excedrin, I have a headache. You guys are talking English... I can read the words... I just don't understand!
Maybe I should take up needlepoint or underwater basketweaving!
Edit: <--- Shuffles off to rant section to start a fight! Slippy, where are you buddy!!!
Message edited by author 2005-10-05 14:03:26.
|
|
|
|
10/05/2005 10:09:07 AM · #16 |
Originally posted by Palmetto_Pixels:
Just when I think I am getting somewhat out of the dark photography confusion I've been in since I bought my DSLR, there's another eclipse!
Can someone please pass the Excedrin, I have a headache. You guys are talking English... I can read the words... I just don't understand!
Maybe I should take up needlepoint or underwater basketweaving!
Edit: <--- Shuffles off to rant section to start a fight! Slippy, where are you buddy!!! |
If you don't already know what we're talking about, none of this will make anys ense. It's a very arcane topic, and they are very specialized lenses. I'd be surprised if more than a dozen photgraphers in DPC have one; I certainly don't, though I used to in my architectural photograohy days.
Pay us no heed. Return to your regular haunts. This is not information you need... jejejeâ„¢
Robt.
|
|
|
|
10/05/2005 10:10:39 AM · #17 |
Originally posted by Uusilehto:
By the way. Is it possible to make a tilt-shift adapter to fit a normal lens? I have a Soligor 28mm f/2.8 in T-Mount. And the T-Mount has 55mm of distance between the film plane and the rear element. Wondering if there might be tilt-shift adapter for T-mount lenses.. |
I'd be astonished if it were possible. These are not "normal" lens reverse-engineered to some sort of tilt/shift mount. The lens are optically different to allow this to happen.
Robt.
|
|
|
|
10/05/2005 10:16:49 AM · #18 |
Originally posted by bear_music:
Pay us no heed. Return to your regular haunts. This is not information you need... jejejeâ„¢
Robt. |
Probably very true, I just find myself with a maddening desire to understand all the details of any subject I have an interest in (photography in this case). I'm the guy that took apart all of his toys as a kid to see how they work. Usually I can get it eventually, but when I find something I can't wrap my brain around, I get frustrated with my own ignorance. This thread is the deep end of the pool... I'll be heading back over to the kiddie pool for now to thrash around some more... Maybe I'll be able to dive back in over here eventually! I'm bookmarking this thread for further reading at a later date.
This seems to be a very good description Robert, just a little (OK, way) over my head right now!
|
|
|
|
10/05/2005 10:23:37 AM · #19 |
Mamiya made one for the RZ, but I don't know of one for 35mm.
|
|
|
|
10/05/2005 10:27:25 AM · #20 |
Originally posted by Palmetto_Pixels:
Usually I can get it eventually, but when I find something I can't wrap my brain around, I get frustrated with my own ignorance. This thread is the deep end of the pool... I'll be heading back over to the kiddie pool for now to thrash around some more... Maybe I'll be able to dive back in over here eventually! I'm bookmarking this thread for further reading at a later date.
This seems to be a very good description Robert, just a little (OK, way) over my head right now! |
Let me try to simplify it:
1. DOF is always perdendicular to the long axis of the lens. Shoot stroght down a ruler, the ruler is parallel to the lens axis. DOF is from, say, 3 inches to 9 inches; perpindicular to the lens axis. Got that?
2. The PLANE of focus is the imaginary plane perdendicular to the lens axis that runs through the POINT of focus: if you focus at 6 inches, anything on an imaginary plane through that 6-inch point parallel to the sensor plane/perpendicular to the lens axis will be in sharp focus.
3. If 3 planes not parallel are extended indefinitely, each of them will intersect with the other at some point; basic geometry.
4. (Take this one on faith) if the 3 extended planes meet at the same point, everything on the central, subject plane will be in focus.
I don't suppose that helps much, sorry...
R.
|
|
|
|
10/05/2005 10:32:15 AM · #21 |
Originally posted by bear_music:
..I don't suppose that helps much, sorry... |
Actually it does help... Honestly that's about as much as I understood, the rest is what I was having difficulty with.
Edit: Thanks for the explanations!
Message edited by author 2005-10-05 14:33:06.
|
|
|
|
10/05/2005 10:55:58 AM · #22 |
Originally posted by Uusilehto:
By the way. Is it possible to make a tilt-shift adapter to fit a normal lens? I have a Soligor 28mm f/2.8 in T-Mount. And the T-Mount has 55mm of distance between the film plane and the rear element. Wondering if there might be tilt-shift adapter for T-mount lenses.. |
Speaking from an engineering standpoint, it would be possible. That's basically what a Lensbaby is, but it would probably not be too practical. For one thing, the image circle on most lenses for 35mm format have too small of an image circle to allow much in the way of camera movements.
It would be much more practical to adapt a medium format or large format lens to the task, but, you'd have to custom engineer and fabricate an adapter with movements built in. That would not be an easy task.
If you want a T/S lens, it might be better to just rent one or pick up a Russian made one from Kiev Camera. These are really not L lenses, but for as little as $300, they are pretty damn good.
|
|
|
|
10/05/2005 11:29:49 AM · #23 |
Originally posted by Spazmo99:
Originally posted by Uusilehto:
By the way. Is it possible to make a tilt-shift adapter to fit a normal lens? I have a Soligor 28mm f/2.8 in T-Mount. And the T-Mount has 55mm of distance between the film plane and the rear element. Wondering if there might be tilt-shift adapter for T-mount lenses.. |
Speaking from an engineering standpoint, it would be possible. That's basically what a Lensbaby is, but it would probably not be too practical. For one thing, the image circle on most lenses for 35mm format have too small of an image circle to allow much in the way of camera movements.
It would be much more practical to adapt a medium format or large format lens to the task, but, you'd have to custom engineer and fabricate an adapter with movements built in. That would not be an easy task.
If you want a T/S lens, it might be better to just rent one or pick up a Russian made one from Kiev Camera. These are really not L lenses, but for as little as $300, they are pretty damn good. |
Arr.. The Arsat seems to go for around $250-350
I did a bit of checking on eBay and there seem to be a few MF -> EOS/F/etc.. shift adapters. //cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=7550733260 |
|
|
|
10/05/2005 11:48:58 AM · #24 |
Originally posted by Uusilehto:
Originally posted by Spazmo99:
Originally posted by Uusilehto:
By the way. Is it possible to make a tilt-shift adapter to fit a normal lens? I have a Soligor 28mm f/2.8 in T-Mount. And the T-Mount has 55mm of distance between the film plane and the rear element. Wondering if there might be tilt-shift adapter for T-mount lenses.. |
Speaking from an engineering standpoint, it would be possible. That's basically what a Lensbaby is, but it would probably not be too practical. For one thing, the image circle on most lenses for 35mm format have too small of an image circle to allow much in the way of camera movements.
It would be much more practical to adapt a medium format or large format lens to the task, but, you'd have to custom engineer and fabricate an adapter with movements built in. That would not be an easy task.
If you want a T/S lens, it might be better to just rent one or pick up a Russian made one from Kiev Camera. These are really not L lenses, but for as little as $300, they are pretty damn good. |
Arr.. The Arsat seems to go for around $250-350
I did a bit of checking on eBay and there seem to be a few MF -> EOS/F/etc.. shift adapters. //cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=7550733260 |
Too bad it doesn't have tilt.
|
|
|
|
10/05/2005 12:02:18 PM · #25 |
Originally posted by bear_music:
If you don't already know what we're talking about, none of this will make anys ense...... |
As sad as it sounds, your discussion makes sense to me (or maybe I misunderstood enough to just think it did). I just find it an interesting topic even though I am not likely to buy a lens for it :-) Damn, something else to read about - I can see how this is useful in some types of photos. |
|
