Home · Register · Join Upload & Sell

Moderated by: Fred Miranda
Username  

  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Canon Forum | Join Upload & Sell

1
       2       end
  

Archive 2012 · Why no tilt/shift sensor?

  
 
yauyi
Offline
• • •
Upload & Sell: Off
p.1 #1 · p.1 #1 · Why no tilt/shift sensor?


Just curious, is it possible to implement tilt/shift sensor onto our DSLR instead of relying on T/S lens? maybe some type of micro motor/servo to control the sensor movement? it would be sweet to retain AF on the lens and still tilt/shift via the sensor.


Jan 25, 2012 at 02:41 PM
Ian.Dobinson
Offline
• • • • • •
Upload & Sell: Off
p.1 #2 · p.1 #2 · Why no tilt/shift sensor?


Well there are probably quite a few reasons. But off the top of my head I would say you would need a lens that could make a much larger image circle to get any possible shift. Hell most thought it was pushing it for Sony to make a sensor is unit for a ff body


Jan 25, 2012 at 02:57 PM
retrofocus
Offline
• • • • • •
Upload & Sell: Off
p.1 #3 · p.1 #3 · Why no tilt/shift sensor?


Technically feasible I would say - yes, as Ian pointed out, the image circle of the lens must be larger, but this is already done in the T/S lens design for example. Practically I think that the demand for such kind of T/S sensor camera would be a niche market. Profit-wise it is also a lot better to charge at least $1K for a single T/S lens than offering a camera with T/S sensor.


Jan 25, 2012 at 03:22 PM
Ian.Dobinson
Offline
• • • • • •
Upload & Sell: Off
p.1 #4 · p.1 #4 · Why no tilt/shift sensor?


But that's part of my point. The TS lenses may through a larger circle but if it's to be done at the sensor plane then all the lenses would need to.

I guess you could do it using EF lenses on an EFS sized sensor



Jan 25, 2012 at 03:28 PM
dhphoto
Offline
• • • • • •
Upload & Sell: Off
p.1 #5 · p.1 #5 · Why no tilt/shift sensor?


All sorts of reasons - image circle, difficulty getting the image plane gearing precise enough, dust, probably loads more

But the biggest - nowhere near enough interest to warrant the R&D involved. This would never be a mass-market item IMO



Jan 25, 2012 at 03:32 PM
cohenxa
Offline
• • •
Upload & Sell: Off
p.1 #6 · p.1 #6 · Why no tilt/shift sensor?


Regaring the image circle, I agree with people comment on the S (shift: vertical and horizontal) but if you only focus on T (tilt), you should be able to tilt the sensor to get the same "effect" of controlled DoF, shouldn't you?


Jan 25, 2012 at 03:46 PM
jcolwell
Offline
• • • • • • •
Upload & Sell: On
p.1 #7 · p.1 #7 · Why no tilt/shift sensor?


Having tilt and shift on the sensor would be analagous to back tilt and shift on a large format camera. That plus tilt and shift of the lens would provide the "front" and "back" movements that large format cameras generally provide for people who still use large format film (whatever that is). Of course, many LF cameras also provide swing and rise movements, as well.

As David (dhphoto) mentioned, there's not sufficient interest to make this worthwhile. A quick reality check would be to compare the number of Canon T-S lenses sold in the past five years with the number of Canon DSLR cameras sold in the same time. I don't have these numbers, but I'm confident that they're different by many zeroes at the end. Few enough people do photography that requires tilt and/or shift in one axis at a time. Far fewer need multiple movements.

As far as image circle size goes, you don't need to have IC diameters greater than contemporary T-S lenses unless you want to combine lens + sensor movements. It's all about relative movement between the lens and sensor. Shifting the lens up 10mm is the same as shifting the sensor down 10mm. Of course, you won't any useful movements out of lenses with 'normal' IC diameters, whether you shift at the lens or at the sensor. They'll rapidly get into serious vignetting territory, in either case.

P.S. tilt doesn't control DOF, it controls the relative orientation of the "plane of focus". The DOF relationship with aperture still more-or-less applies.



Jan 25, 2012 at 03:50 PM
chrisdee
Offline
• •
Upload & Sell: Off
p.1 #8 · p.1 #8 · Why no tilt/shift sensor?


cohenxa wrote:
Regaring the image circle, I agree with people comment on the S (shift: vertical and horizontal) but if you only focus on T (tilt), you should be able to tilt the sensor to get the same "effect" of controlled DoF, shouldn't you?


No, the sensor tilt would also require a shift, in order to keep the composition, so you end up with the required larger image circle again.

The tilt lens uses the Scheimpflug effect to achieve the "control of DOF" without requiring larger image circle, but when you move the tilt to the sensor, the movement required is a complex tilt+shift to achieve the same results.

EDIT: as Jim pointed out, "control of DOF" above is incorrect, should read control of focus plane



Jan 25, 2012 at 04:01 PM
RCicala
Offline
• • • •
Upload & Sell: Off
p.1 #9 · p.1 #9 · Why no tilt/shift sensor?


Simple reality is the manufacturers are struggling to keep the sensor plane perfectly parallel to the lens mount now (well, not perfectly, but reasonably close). Put that sensor on a mechanical tilt-shift arrangement and imagine what it would be like.


Jan 25, 2012 at 04:17 PM
Roland W
Offline
• • • •
Upload & Sell: On
p.1 #10 · p.1 #10 · Why no tilt/shift sensor?


There is not enough room where the image comes in to the camera to provide for sensor tilt and shift, and still have a mirror big enough, and still have a metering system. Even if you had lenses that had big enough image circles, they just could not work with a tilt shift sensor with a mirror. You would likely need to give up the reflex viewing that a DSLR normaly has, and for the same reasons you would not have exposure metering via a system that depends on the mirror to get its light from the image. Also, the current style of autofocus would go away when the mirror does.

A dedicated sensor only style camera, with good live view, could likely be designed, but as mentioned, the market is very limited. But the live view could still do exposure metering and some form of auto focus or focus confirm at full aperture, and then could stop down for exposure. If focus information was made available from the live view sensor data in some form, it could be specific to different parts of the image based on user preference or settings. And the camera could have automatic cycles included to evaluate focus as it stops the lens down in increments, to show what aperature is required to get all points in focus. Lots of possibilities, but not too much need or demand overall.

Once you think in terms of removing all the mirror related stuff from a camera, you revert to something much like a view camera. That then means that there is no reason to use the retrofocus design for wide angle lenses, making it easier to build high quality lenses with large image circles at lower cost with less distortion issues. But that would mean that you would need a whole new family of lenses to get nice performance, which is also unlikely because of the limited market.

Be glad that Canon has built two very nice tilt shift lenses with modern designs for us to use. They project the tilted and shifted image into the center of the current camera geometry, although the exposure metering gets messed up. Most things can be done with the current TS-E lenses, now that the alignment axis between tilt and shift is user selectable to any angle. This all reminds me that I need to get out more and use my gear, including my TS-E lenses, to create images.

Edited on Jan 25, 2012 at 04:29 PM · View previous versions



Jan 25, 2012 at 04:25 PM
eosfun
Offline
• • • •
Upload & Sell: Off
p.1 #11 · p.1 #11 · Why no tilt/shift sensor?


In camera settings of all kind is definitely an interesting route to reduce cost while adding creative possibilities. An electronically controlled tilt and shift sensor is one of them. In camera image stabilization is already a fact for some brands. In camera autofocus would be one of the most interesting innovations that would simplify lens production a lot and it would make the fit of almost every lens brand AF possible on that kind of camerabody. Contax once did that with it's AX, a film camerabody which made all Zeiss manual focus lenses autofocus. I think an EOS body with some in camera focus system would be the best kind of EOSfun for our alt-gear friends in that other board Not that I believe this is gonna happen, but there is definitely more room for innovation in camera construction then we have seen so far.


Jan 25, 2012 at 04:29 PM
eosfun
Offline
• • • •
Upload & Sell: Off
p.1 #12 · p.1 #12 · Why no tilt/shift sensor?


@Roland: There is more possible than you might think with a Canon D-SLR and current view camera for digital "back" solution from our Dutch photographic equipment producer Cambo Cambo Ultimate 35] This is a "low cost" view camera solution that a few colleagues use in their studio instead of high cost medium formats or digital backs on a Sinar or Linhof view camera. True, you need different glass that mounts to the Cambo. But with some of the great Schneider or Rodenstock lenses in retrofocus design you have great creative possibilities. Especialle live view and ISO performance - noise treatment of the Canon bodies are what some photographers like about this great professional source of EOSfun I guess this is a kind of application of Canon that not everyone at this board here was aware of, that's why I share. I have no commercial connections with Cambo, but my patriotism for this 'made in Holland' makes me feel proud


Jan 25, 2012 at 04:43 PM
Richard Nye
Offline
• • • •
Upload & Sell: Off
p.1 #13 · p.1 #13 · Why no tilt/shift sensor?


Would not a tilted sensor distort the image aspect ratio? Take it to extremes, if the image sensor was tilted 45 degrees from the plane of focus, only a portion of the image would be projected onto the sensor. When the pixels are printed, the aspect ratio would be restored and the image distorted.

I'm not sure I'm making myself very clear, but I'm certain the image would be distorted (stretched vertically or horizontally depending on the tilt direction).



Jan 25, 2012 at 04:57 PM
eosfun
Offline
• • • •
Upload & Sell: Off
p.1 #14 · p.1 #14 · Why no tilt/shift sensor?


Right, that is why movements of focal plane, sensor plane and optical axis in relation to the subject in general need to be more complex than just a simple tilt of the lens (or the back). Read on Scheimpflug's rule: http://en.wikipedia.org/wiki/Scheimpflug_principle Shifted or on Tilt Have EOSfun


Jan 25, 2012 at 05:11 PM
AJSJones
Offline
• • • •
Upload & Sell: Off
p.1 #15 · p.1 #15 · Why no tilt/shift sensor?


Richard Nye wrote:
Would not a tilted sensor distort the image aspect ratio? Take it to extremes, if the image sensor was tilted 45 degrees from the plane of focus, only a portion of the image would be projected onto the sensor. When the pixels are printed, the aspect ratio would be restored and the image distorted.

I'm not sure I'm making myself very clear, but I'm certain the image would be distorted (stretched vertically or horizontally depending on the tilt direction).

You would get the same kind of distortion as experienced when the camera itself is tilted away from vertical (most obvious with parallel lines and wide-angle lenses - and often blamed on the lens and not the camera angle ). The plane of the rectangle in the subject (e.g building facade) is not parallel to the plane of the sensor). Tilting the sensor and not the camera causes the same distortion but I don't think anyone is suggesting large tilts like 45 degrees.A tilt of ±5-10 degrees of sensor would not require much increase in image circle...



Jan 25, 2012 at 05:21 PM
Photon
Offline
• • • • • •
Upload & Sell: Off
p.1 #16 · p.1 #16 · Why no tilt/shift sensor?


Distortion resulting from sensor tilt would be no different from what occurred with view cameras using back adjustments, and lens tilt introduces distortion as well. In practice, tilt and shift at just one place (the lens, or the back/sensor) can provide all of the perspective and focus control available through moving at both places. The limitations are the optical properties of the lens and the mechanical limits of the mounts.

One nice thing about a movable sensor (if shift were part of it) would be the ease of shooting a stitched panorama with the lens fixed in place. No more need for Nodal Ninja or other fancy tripod accessories!



Jan 25, 2012 at 05:40 PM
chrisdee
Offline
• •
Upload & Sell: Off
p.1 #17 · p.1 #17 · Why no tilt/shift sensor?


AJSJones wrote:
You would get the same kind of distortion as experienced when the camera itself is tilted away from vertical (most obvious with parallel lines and wide-angle lenses - and often blamed on the lens and not the camera angle ). The plane of the rectangle in the subject (e.g building facade) is not parallel to the plane of the sensor). Tilting the sensor and not the camera causes the same distortion but I don't think anyone is suggesting large tilts like 45 degrees.A tilt of ±5-10 degrees of sensor would not require much increase in image circle...


Tilting the sensor 10 degrees would require a sensor shift of approx. 8mm on EOS mount, which is very significant.



Jan 25, 2012 at 06:12 PM
Sp12
Offline
• • •
Upload & Sell: Off
p.1 #18 · p.1 #18 · Why no tilt/shift sensor?


Pentax has this. It works very well IMO.


Jan 25, 2012 at 08:25 PM
jcolwell
Offline
• • • • • • •
Upload & Sell: On
p.1 #19 · p.1 #19 · Why no tilt/shift sensor?


Pentax has in-camera IS. That's not the same as tilt-shift.


Jan 25, 2012 at 09:43 PM
Hammy
Offline
• • • •
Upload & Sell: Off
p.1 #20 · p.1 #20 · Why no tilt/shift sensor?


Actually,

I have a Canon body that does this.

After a fall, my wife and the 200/1.8 attached are ok, but what took some time to notice is that all photos taken with that body exhibit tilt behavior.

As in where the image was focused is dead on in focus, but depending on the range, the subject's feet are out of focus - shifting to back focus.

It's actually a neat effect which works for 98% of our photos taken in portrait mode. But I'll prolly have to send it in for service to make it easier on those 2% of photos that I forgot about the issue. ;(



Jan 25, 2012 at 09:45 PM
1
       2       end




FM Forums | Canon Forum | Join Upload & Sell

1
       2       end
    
 

You are not logged in. Login or Register

Username       Or Reset password



This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.