Tilting for DOF and focusing distance...

[ChristianRandwijk]

So, I have been saving up, and will be getting the new 24mm tilt/shift lens from Canon for my 5D2 next month. I have been reading a bit about tilting and controlling the plane of focus. I do have (at least) one question, as far as I know, the angle of the plane of focus is determined by BOTH tilt AND focus distance. Is this true, or am I mistaken? And if this is the case, what is a practical approach when tilting for larger DOF? Focusing at infinity or hyperfocal distance, and then applying tilt? 

[Tony Jay]

There are quite a few resources on the web detailing the Scheimphlug principle.

In practice the best way to decide on the the degree of tilt and the focussing distance is by trial and error.
Using live view is a true godsend here.
Set up your camera on a tripod and then use live view to fine tune the composition.
Once the composition is correct use the magnified live view to check focus in all areas of the composition.
Alter tilt by a degree or so and recheck focus using magnified live view again scanning all parts of the composition.
Slight alterations to focus using the focussing ring may also be necessary.

In general my experience with this lens using it to shoot landscapes (with maximum depth of field being the desirable outcome)
I have used f stops between f8.0 and f11.0.
Depth of field is great anyway depending on where one is focussing into the composition.
Without magnified live view I have found it almost impossible to accurately gauge focus just using the viewfinder.

Reviewing the principles governing tilt-shift lenses is helpful but in the field the trial and error approach described above will be
the way to go.


Regards

Tony Jay

[Schewe]

...as far as I know, the angle of the plane of focus is determined by BOTH tilt AND focus distance. Is this true, or am I mistaken?

You are mistaken. The plane of focus is the plane of focus regardless of the focus distance. You depth of field does reduce as you focus closer but that has nothing to do with the plane of focus. The reason you tilt is to adjust the plane of focus so closer objects and further objects can be brought into focus.


And yes, Google the Scheimphlug principle...the bottom line is the sensor plane is fixed (in a DSLR) and the lens plane can be tilted to get the focus plane to all intersect at a theoretical point. You can guestimate that by looking at the camera from the side and judge how much tilt would be required to have the plane of focus line up.

[mediumcool]

I always thought the spelling was Scheimpflug, starting with my photographic training in the ’60s. 

[Bart_van_der_Wolf]

So, I have been saving up, and will be getting the new 24mm tilt/shift lens from Canon for my 5D2 next month.

Hi Christian,

It's an amazing lens, I can tell you from personal experience.

I have been reading a bit about tilting and controlling the plane of focus. I do have (at least) one quon, as far as I know, the angle of the plane of focus is determined by BOTH tilt AND focus distance. Is this true, or am I mistaken?

You are correct. When in the traditional plan parallel optical distance versus film (or sensor) plane mode, common DOF principles are valid, the Scheimpflug principle will be better conceptualized as wedge (or rather hyperbolic) shaped DOF zones through the scene.

This means that the wedge shaped DOF zone takes on more of a vertial dimension with increased tilt, rather than a depth dimension only.

And if this is the case, what is a practical approach when tilting for larger DOF? Focusing at infinity or hyperfocal distance, and then applying tilt?

This means in practice that, for a significant enough tilt, the wedge shaped DOF zone is 'lower' at close distance, and 'higher' at larger distances.

This also means that for pre-visualisation, tethered shooting or Live View is essential for the best results (although for fixed distances to the focus plane, fixed degrees o til il work ).

Cheers,
Bart

[torger]

The 24mm TS-E is my favourite lens.

A basic method for tilt is to 1. estimate tilt you need (usually 1.5 - 2.5 degrees) and set that as starting point. 2. Focus using the focus ring on the nearest object you want in focus. 3. Adjust tilt so the far object gets sharp. 4. Reiterate 2 and 3 until sharp in both ends, making sure you adjust tilt less and less. Ideally use f/8 to f/11.

I suggest that you compare with hyperfocal distance too, sometimes the near object is so far away that you can just get whole scene sharp without tilt.

Sometimes the equation is "unsolvable", meaning you cannot get everything in the scene sharp, since the DOF is very thin close to the camera, and then you need to do a tasteful compromise. In cases when you have high objects like trees close to the camera (vertical) tilt is usually no good, since you won't get the tree sharp from top to bottom.

If you will be using shift, set shift first. Tilting heavily will affect framing a bit, so you may need to adjust that a little too.

After a while one gets quite good at estimating what tilt is needed so the tilt/focusing goes rather quickly. A good live view (like 5Dmk2 has) is key, so bring an extra battery or two since live view will drain power quickly.

[mediumcool]

Sometimes the equation is "unsolvable", meaning you cannot get everything in the scene sharp, since the DOF is very thin close to the camera, and then you need to do a tasteful compromise. In cases when you have high objects like trees close to the camera (vertical) tilt is usually no good, since you won't get the tree sharp from top to bottom.

This is where focus stacking could perhaps help in extremis; I don’t think that dedicated software would be able to easily cope with changes in image shape, not just scale.

If you will be using shift, set shift first. Tilting heavily will affect framing a bit, so you may need to adjust that a little too.

+1

[CptZar]

http://www.cambridgeincolour.com/tutorials/tilt-shift-lenses2.htm

[ChristianRandwijk]

Thanks all,
   I'll definately check out the links. My primary concern is the ability to shift (I'll be shooting both architecture and landscape), also, being a 24mm lens, DOF is pretty deep from the outset. But it would be great being able to get very deep depth of field already at f/8 (around this lens' optimum aperture I'm told, and still somewhat safe from visible diffraction). Also, something being within the DOF obviously isn't the same as that something being at, or very close to, the actual plane of focus. But, I really appreciate the info guys, I'll look into it presently, thanks again.