Pages: 1 2 [3]   Go Down

Author Topic: 45mm tilt-shift lenses, canon or nikon?  (Read 24721 times)

kers

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3875
    • Pieter Kers
Re: 45mm tilt-shift lenses, canon or nikon?
« Reply #40 on: February 21, 2015, 11:00:07 am »

I notice over and over again how difficult it is to do a valid test ...and to draw the right conclusions from it...
But indeed the TS mechanism of the Nikkors could be improved...
The lenses are made in the 12 MP era so doing still quite well...

Logged
Pieter Kers
www.beeld.nu/la

CptZar

  • Full Member
  • ***
  • Offline Offline
  • Posts: 157
Re: 45mm tilt-shift lenses, canon or nikon?
« Reply #41 on: February 21, 2015, 11:36:56 am »

Ben thanks for sharing...
It is clear the zeiss wins hands down...also more shift...
But there is some problem i think with the 45PC- the center images seems OK.
when shifted - some parts that were sharp in the center image have become unsharp..? (look at the grass between the houses with the snow)

Diglloyd writes in his review of the 40 Zeiss Hartblei about stitching:

Start Quote
While overall imaging performance is very high, with especially appealing color, attention must be made to subtle shifts in focus when shift is applied. This makes the Hartblei 40/4 less than ideal for quick shift-lens-stitching applications.
End Quote

As far as I understand it, there will always be a focus shift implied once a shift or tilt movement was applied. Therefore focusing can only be done after any movement. For stitching this means every single shot needs refocusing.

The Zeiss CFE 4/40 looks much better.

Cheers

Jan

ben730

  • Full Member
  • ***
  • Offline Offline
  • Posts: 217
    • www.benhuggler.com
Re: 45mm tilt-shift lenses, canon or nikon?
« Reply #42 on: February 21, 2015, 12:33:00 pm »

I notice over and over again how difficult it is to do a valid test ...and to draw the right conclusions from it...

My conclusion is: I do with both lenses professional work and none let me down, never (!) and I use both a lot.
All pictures were always by far sharp enough for their points of use (magazines + advertising).
I choose my equipment often due to desire, mood and motivation.
Normally I am the limiting factor, not my equipment. :-\

 

marc aurel

  • Full Member
  • ***
  • Offline Offline
  • Posts: 150
Re: 45mm tilt-shift lenses, canon or nikon?
« Reply #43 on: February 21, 2015, 12:37:42 pm »

Diglloyd writes in his review of the 40 Zeiss Hartblei about stitching:

Start Quote
While overall imaging performance is very high, with especially appealing color, attention must be made to subtle shifts in focus when shift is applied. This makes the Hartblei 40/4 less than ideal for quick shift-lens-stitching applications.
End Quote

As far as I understand it, there will always be a focus shift implied once a shift or tilt movement was applied. Therefore focusing can only be done after any movement. For stitching this means every single shot needs refocusing.

The Zeiss CFE 4/40 looks much better.

Cheers

Jan



@Jan:
I think that Diglloyd is not very clear in what he means with "focus shift". I suspect what he observed relates to field curvature.

As far as I understand it:
- the lens projects an image to the sensor.
- this projected image (and the focus in that image) does not change when you shift, you just use a different part of the image.
- if you stitch several images from different shift positions then you can capture a larger part of the projected image.
- but if a lens has field curvature, then the point of focus is different in the centre of the lens than further out
  (and most wideangle shift lenses with big image circles have very relevant field curvature, the Hartblei too)
- so in the shifted image, it SEEMS as if the focus point had moved (although you did not touch the focusing ring of the lens).
- the problem to get an image that is sharp across the whole stitched image is difficult because field curvature gets more relevant.

So if you refocus every shot - that would be something like focus stacking to overcome field curvature? Never tried it, but I think it is not easy to get the transitions done right. Maybe refocusing would slightly change magnification? Has any one of you tried that?

Marc
Logged

CptZar

  • Full Member
  • ***
  • Offline Offline
  • Posts: 157
Re: 45mm tilt-shift lenses, canon or nikon?
« Reply #44 on: February 21, 2015, 01:56:54 pm »

Agreed, field curvature might be the reason why the focus appears to move. And I have to admit, that I never saw the problem of "virtual focus stacking" by refocusing. Quiet interesting point though. And even more pronounced in architecture than in landscape. But even if you take only one picture, field curvature kicks in, if you focus first and then shift the lens.

However I have to admit, that I am not a real fan of stitching anyway. In the landscape light usually changes rather fast, and stitching always means you might not be able to catch the "decisive moment", because instead of waiting for it you will fiddle around with your gear. Also I don't really see the requirement of stitching with sensors larger than 36MP. A perfect panorama 3:1 might be shoot with one cropped 17mm shot instead of 3x40mm. The result will be big enough for 99% of the required prints. But that's just a personal preference. Again this might be another game in architecture photography.

Cheers

Jan


Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8901
Re: 45mm tilt-shift lenses, canon or nikon?
« Reply #45 on: February 21, 2015, 02:11:07 pm »

So if you refocus every shot - that would be something like focus stacking to overcome field curvature? Never tried it, but I think it is not easy to get the transitions done right. Maybe refocusing would slightly change magnification? Has any one of you tried that?

Hi Marc,

No problem for competent stitching applications that allow to optimize for 'focal length' per image. The focal length optimization allows to improve accuracy with respect to the nominal focal length (which only is valid for infinity focus). When the optimization is allowed to do that per individual image, the 'magnification' is applied variably.

But if one needs to do that per shifted image, it's almost easier to just use the center of the image and rotate camera and lens through the entrance pupil position, like with common stitching. That will describe a more circular in-focus path, but DOF is usually enough to deal with that. I wonder if DOF isn't enough with the stitched approach to cover field-curvature, unless one shoots almost wide open.

Cheers,
Bart
« Last Edit: February 21, 2015, 02:26:51 pm by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

marc aurel

  • Full Member
  • ***
  • Offline Offline
  • Posts: 150
Re: 45mm tilt-shift lenses, canon or nikon?
« Reply #46 on: February 21, 2015, 02:23:43 pm »

@Bart

Thanks, I never tried that with a stitching application. Did not take much time with these programs, maybe I should try that.

Stitching can give you amazing resolution - but like Jan I am not such a great fan of stitching. I love to see the final composition in the viewfinder. So I usually only stitch in the very rare cases where I have a wall against my back and the TS-E 17 is not wide enough.
I would love to have more resolution at my hands without stitching. But 36MP is a lot already. Let's see how things look when the 50MP cameras arrive...

Cheers - Marc
Logged
Pages: 1 2 [3]   Go Up