PC 19mm f/4 E ED

[Erik Lund]

For me the key feature would be shift for architectural purposes, as most of the samples in the Nikon presentation also show

There is a lot of loss of active sensor area while cropping from say a 14mm and not close to what the 19mm can do - But OK let's see when the reviews start to come out

[Airy]

Sure, but with 50 MP cameras becoming a commonplace, it is less of a problem, while not optimal.

Also, the quality loss at maximum recommended shift may still be problematic.

[chambeshi]

If you are asking for my opinion, I always consider longer lenses to be the better choice for landscape work. Thus the 85 PC is my preferred alternative. It combines nicely with the 24 PC as well.

If just a single PC lens is feasible, the 45 might do good service for some close-ups (it is labelled 'Micro-Nikkor' after all) and the occasional landscape scene as well.

Thank You :-)

[Airy]

Agree with Björn for landscapes - around 90mm (equiv.) was also mentioned as optimal one century ago.
For cityscapes, that's another story. 24-28 is fine.

[Roland Vink]

I wonder why there are not lenses with tilt-only function for controlling the focus DOF/focus plane. I almost never wish for a shift-lens, but a tilt-lens would be very useful for macro and landscape. A tilt-only lens does not require a larger image circle like shift lenses, so the optics could as compact and affordable as a regular lens.

[Bjørn Rørslett]

One usually has to use shift in conjunction with the tilt. Thus having both features available is not just convenient, it is basically a requirement.

Having only shift present makes for a different kind of approach, thus you can get away with shift only if the plane of focus does not require to be moved.

Still, modern DSLRs (or mirrorless) cameras are less convenient than the old technical view cameras, because the mirrorbox/shutter severely restricts movements by acting like a stop gate and cutting off light. The technical cameras used lenses with leaf shutters to provide many more degrees of freedom as to the kind and extent of movements usable.

[abergon]

One usually has to use shift in conjunction with the tilt. Thus having both features available is not just convenient, it is basically a requirement.

Having only shift present makes for a different kind of approach, thus you can get away with shift only if the plane of focus does not require to be moved.

The first Nikkor PC 35mm lenses offered shift only. However, it is not the easiest lens to work with.

[Bjørn Rørslett]

Many years ago, I literally hacked a 35 PC-Nikkor to get it to give tilt as well. This worked, but was restricted as to what could be achieve because tilt and shift could be properly separated. Therefore, the next project was modifying a 28 PC to allow shift and tilt concurrently with a degree of independence. (The literal scene with barnacles was shot with that lens). The image below shows the best I could do with the hacked 35.

[Roland Vink]

One usually has to use shift in conjunction with the tilt. Thus having both features available is not just convenient, it is basically a requirement.

Doesn't this depend on where the center of tilting rotation lies?

With the 85PC for example, the lens tilts around a point well in front of the image plane, so tilting the lens causes the image circle to move laterally, and shift is required to bring it back to the center of the sensor.

On the other hand, if the lens tilts around a point on the image plane, the image circle remains more or less stationary. All that is required here is a lens with sufficient focus travel so that any part of the sensor tilted back "beyond infinity" can be brought back into focus.

I could imagine a mirrorless camera designed for landscape work, where the sensor could be tilted for the same effect...

[Bjørn Rørslett]

I can only state that both tilt and shift are required, also for short focal length lenses. Probably even more in the latter case.

Even a mirrorless camera acts as a stop gate to light rays as it has finite thickness and there is a lens mount well in front of the shutter/sensor plane.

The idea of a tilted sensor is equivalent to rear tilt on a view camera. This makes exposure uneven over the film plane because light rays travel different distances from the exit pupil.

[Roland Vink]

Even a mirrorless camera acts as a stop gate to light rays as it has finite thickness and there is a lens mount well in front of the shutter/sensor plane.

The idea of a tilted sensor is equivalent to rear tilt on a view camera. This makes exposure uneven over the film plane because light rays travel different distances from the exit pupil.

This is true of any tilting system, whether the lens is tilted or the camera/sensor.
When the light path from the exit pupil to the sensor is relatively long, any difference in distance due tilting will be minor, in wide angle lenses with very short back-focus it could be significant, especially if the tilt angle is large.

Tilting the sensor on a mirrorless camera does away with problems of peripheral rays from a tilt-lens being cut off by the lens mount, or the  sensor conflicting with the reflex mirror of an SLR. One only needs to consider the focal plane shutter:
- either place the shutter far enough forward to clear a tilting sensor - not much space is required as tilts are typically small.
- or do away with the focal plane shutter entirely by utilising in-lens leaf-shutters, or global electronic shutter.

[Akira]

Another problem of the digital cameras is the on-chip micro lens which also narrows the effective angle of incident light.

[Bjørn Rørslett]

Rolkand: You might overlook the change in geometry caused by tilting in front or rear.  Any textbook on view cameras deal with these issues.

As long as there is a lens mount in the camera, it will act as a constraint for lens/camera movements. Only shutters at the aperture position can avoid such limitations.

Let us continue any further discussion on these matters outside this thread lest it be completely derailed.

[David H. Hartman]

The inverse square law works inside a camera just as it does out side a camera. The lens becomes the light source for the sensor. The distance matters.

Dave

[Erik Lund]

I hear what you are stating, of course both settings are ideal and since the 19mm has rotation between the two settings as well it doesn't get any better for FX,,,

Agree the price hurts even though it appears the lens is a mechanical marvel as well as optical,,,

Price is equal to buying an additional D810 and two samples of the 20mm 1.8 AFS for dedicated 'one shot' two images (Two cameras two lenses) stitching ready images,,, 
Then the nodal point is shifted,,, damn