Lenses for D800e

[Petrus]

I took the liberty of playing around with one of my own. This is shot at canon 5dmkII with canon 70-200 f4 L @160mm @f16. The red square marks the area of the two 100% crops. They are processed in RawDeveloper with just a slight WB and contrast adjustment. One has no sharpening, the other has been sharpened with R/L Deconvolution at radius .60 and 25 iterations. The third one has gotten additional sharpening in Photoshop.

The one without sharpening is definitely soft, and some softness is of course because of the AA filter in 5dmkII.

5DII starts do loose resolution at f:8 and smaller, so diffraction softens the picture much more than the AA filter. @f:16 5DII resolves only about 7 MPix (green, less at red, more at blue)

[jgbowerman]

I'm confused! My style is to shoot landscapes with outstanding DOF. I have a rough understanding of diffraction softening, but without focus stacking, DOF will suffer at f/8 and larger apertures in many landscape compositions (unless the scene is relatively flat and one is using a tilt-shift lens). I shoot f11 and even f16, depending on the composition (distance of nearest object to the lens), and when it comes to Cannon iPF prints, I am 100% pleased with apparent sharpness throughout the image. I shoot with a D700 and Nikon PC 45/2.8. I did a workshop with Alain Briot a couple of years ago, and he shoots f22 all the time, claiming he cannot see the difference in his prints as goes sharpness when compared to using larger apertures where he observes sacrificed DOF. He simply focuses 1/3 of the distance into any given composition and claims excellent DOF with outstanding print sharpness using the smallest aperture shooting with Hassy glass and a medium format digiback. Is my print quality suffering at f11 through f16 using high quality glass on an FX DSLR?

I have pre-ordered a D800E and I'm expecting to receive shipment on a Hartblei SuperRotator 40/4 in a few weeks.

[kers]

I am also going to get the Nikon d800e and am trying to find a great 50mm for it. I only need one lens and prefer this type of prime. What autofucos options do I have? Is nikon's 50mm 1.4 any good? It is not on Nikon's preferred lens list. Are there 3rd party options? are there converters for canon lens to fit on nikon? any contax AF lenses that would work? I am only interested in high quality lens fyi. If i cant find an AF lens I would go with the zeiss zf 50mm.

I wish Nikon had made a 50mm lens with the quality of their other 1,4 primes...
I just bought the nikon 1,8G 50mm to replace my 50mm 1,4G- That sounds odd and it is. Nikon made a mistake by making a 50mm1,4G too cheap. The other 1,4 primes are over 1000€ but the 50mm is 300€ that is the problem.
The 1,8g lens is cheaper has a faster autofocus and because of its asferic lens element has more clarity and a bit more sharpness wide open.
But both could be better i think. I hope they will come with a new 50mm noct version with nanocoating.
The alternative today is probably the zeiss d2 50mm;  the coating is better but not autofocus. The 1,4 Zeiss is also good at 5,6 but wide open it has an ugly unsharpness- I replaced it for my 50mm nikkor1,4G....

[erpman]

I'm confused! My style is to shoot landscapes with outstanding DOF. I have a rough understanding of diffraction softening, but without focus stacking, DOF will suffer at f/8 and larger apertures in many landscape compositions (unless the scene is relatively flat and one is using a tilt-shift lens). I shoot f11 and even f16, depending on the composition (distance of nearest object to the lens), and when it comes to Cannon iPF prints, I am 100% pleased with apparent sharpness throughout the image. I shoot with a D700 and Nikon PC 45/2.8. I did a workshop with Alain Briot a couple of years ago, and he shoots f22 all the time, claiming he cannot see the difference in his prints as goes sharpness when compared to using larger apertures where he observes sacrificed DOF. He simply focuses 1/3 of the distance into any given composition and claims excellent DOF with outstanding print sharpness using the smallest aperture shooting with Hassy glass and a medium format digiback. Is my print quality suffering at f11 through f16 using high quality glass on an FX DSLR?

I have pre-ordered a D800E and I'm expecting to receive shipment on a Hartblei SuperRotator 40/4 in a few weeks.

I say: If you want DoF you´re gonna have to live with diffraction. I will gladly sacrifice some resolution to avoid ugly out of focus areas where there isn´t supposed to be any. In print, it´s no big deal if you have a proper sharpening routine. Look at the examples I posted, whether they are 6mp or whatever people claim, I say the sharpened ones are pretty detailed. If this is actual resolution or perceived resolution, who cares???

I also contend that you loose a lot more sharpness and image quality if you have to interpolate a small file from a smaller resolution sensor say to 300% vs interpolating a large resolution sensor file to 150% (or not at all). Interpolation does much more damage to the image than a bit of diffraction.

That being said, I´m not saying stop down more than necessary, do proper calculations to figure out how far down you need to go, using for example a DOF calculator on your iPhone.

[mcbroomf]

Back to lens discussion for a moment.

You said you were concerned over a lens corner sharpness.  Why so if you always stitch (and corners alway overlap)? 

This may open up a few more lens options.. for example the ZF 25/2.8 Zeiss

[Tony Jay]

A couple of counterpoints need to be mentioned here:

Comparing diffraction issues between MFDB and 35mm equivalent sensors is not comparing apples with apples. Do your homework - sensel pitch is important in determining when diffraction may becomes an issue.

I don't know this for certain but, does this Hasselblad sensor have an anti-aliasing filter or not? If not, then the diffraction produced at f22 may still produce a very acceptable perceived sharpness that a 35 mm equivalent sensor, especially a 36MP sensor should it have the anti-aliasing filter, will not.

I fully agree that how one perceives the print, or whatever end-point mode of presentation is used, should be the final arbitrater of sharpness but assuming that what Alan Briot can achieve with his equipment will be the same with everyone else and their equipment requires a little more scrutiny.

Regards

Tony Jay

[erpman]

Back to lens discussion for a moment.

You said you were concerned over a lens corner sharpness.  Why so if you always stitch (and corners alway overlap)? 

This may open up a few more lens options.. for example the ZF 25/2.8 Zeiss

I see your point, and I certainly wouldn´t mind shooting with the mighty 25/2.8 Zeiss. There are however two problems: 1) I want to reduce the number of images in a stitch, and do single-row panoramas with the camera mounted in portrait mode. Then the upper and lower parts of the image will be the ones closest to the edges of the image circle, and therefore, sub-par areas of the lens will still be making up relatively large parts of the image. 2) Since stitchers can´t discern between sharp and blurry areas, weird artifacts can occur in the joints between images if the difference between center and corner sharpness is too large.

Problem 1) might be mitigated by allowing for a little cropping, since it´s usually just the extreme corners that suffer. But is there a way around problem 2)? Are there stitchers that can select only the sharpest areas of the images with consistency?

Looking at this mtf-graph the zeiss does have a good balance between center and corner at f/8, which is mostly where I´ll be using a lens that wide, this makes it very interesting. Thanks for the tip!

[mcbroomf]

I don't know how stitching software works so I can't answer 2), but if you overlap enough then corners are simply eliminated from the final pic.
How much do you overlap?  I typically go for 25% or so, so with a lens stopped down to F11 or even more any residual corner softness (due to lens issues.. depending on the lens of course) should be eliminated, just leaving DOF softness.

[JohnBrew]

erpman, don't get your lenses mixed up. You say you wouldn't mind shooting a Zeiss 25/2.8 but the MTF graph you refer to is for a 25/2.0. They are very different lenses with different characteristics. But maybe you already know that and just made a typo  

[OldRoy]

I don't know how stitching software works so I can't answer 2), but if you overlap enough then corners are simply eliminated from the final pic.
How much do you overlap?  I typically go for 25% or so, so with a lens stopped down to F11 or even more any residual corner softness (due to lens issues.. depending on the lens of course) should be eliminated, just leaving DOF softness.

In any case, assuming you have sufficient overlap, you can always use alpha-channel masking to prevent the stitching software from including any areas that aren't desirable.
Roy

[jgbowerman]

Thanks for the feedback, erpman and Tony

I assume Briot's digiback did not have an AA filter, and the 45MP digiback he used likely had a pixel size of ~6 microns. The D800 pixel size is ~4.9 microns. I appreciate the medium format systems will likely perform better at smaller apertures given the larger pixel size when compared to something in the ballpark of a DSLR FX like the D800. I once read the ideal pixel size is around 6 microns based on work done by Leica when the S2 was released. I have not read much about pixel size relative to diffraction softening. Pixel size discussions are more often about high ISO noise reduction. With the D800 pushing pixel size limitations to the extreme, diffraction issues could become more evident based on what I am reading and the reason why I was compelled to join this discussion. From what I am gathering, for purposes of printing and maintaining DOF, I'll be fine shooting apertures as small as f16, but I'll do my best to shoot wider when appropriate.

This brings me back to the subject of lenses, I also appreciate DOF is something much easier to accomplish with wider lenses. I don't like getting too wide... over the years I have found my "sweet spot" is shooting around 40 to 45MM focal lengths. Subsequently, I began getting rid of my zoom lenses for landscape work in an effort towards improving print quality... no more ugly chromatic aberrations and serious lens flare issues! I stuck with Zeiss glass for Nikon (even if it is made in Japan, the quality of build is superior to Nikon's). Then I tried Nikon's 45MM PC-E and I was hooked with the convenience of fine tuning landscape compositions without having to move the tripod in difficult terrain by using the shift movements. I also gained the opportunity to produce better files for stitching if desired, although I have yet to take much advantage of stitching and most of my landscape work does no lend itself to tilting. Another advantage or the tilt-shift lenses is they serve reasonably well for macro. For me, tilt-shift lenses offer the most in terms creative possibilities, I'll never outgrow all of the options one can explore.

Although the Nikon 45MM PC-E glass is outstanding, the build quality has something to be desired... it is too loose with tilt and shift movements, subsequently I desired a better build quality. Nikon brags about how their PC-E lenses have rubber O-rings, but what good is the O-ring with loose fitting tilt and shift movements? While cost is a serious limiting consideration for 99% of us, it is clearly evident that we get what we pay for when it comes to lens quality.

Eventually, my search for the perfect lens settled on Hartblei Superrotators. Their 40MM model is perfect for my style of shooting, the glass is German made ("TRUE" Carl Zeiss glass) even if the barrel and assembly is Ukrainian. I put the Hartblei 4/40 IF TS Superrotator on my wish list two years ago. When the D800E was released, I figured it was time to bite the bullet and buy one. Stefan has my hard-earned money and I'll soon have one of his lenses.

[erpman]

JohnBrew: Thanks for that reminder, I did have them mixed up  Checked the other zeiss on dxoMark and it does seem better than the 24-70 from f8 onwards.

Mcbroomf: Yeah, I go for something like 25% and looking at the field maps at dxoMark it´s clear that this is more than enough to avoid blurry corners. I assumed that blurriness would be more even spread out along the short sides of the frame.

OldRoy: I guess since the idea is to just stitch 3-7 frames, there won´t be such a huge amount of overlaps and the job of fine-tuning the stitch should be within a reasonable workload. Exactly how do you use apha-channel masking when stitching in PS?