p.3 #1 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Merci snapsy. That's a great job!
I'd like to know how the 5DS/R compares to the 1Dx/5D3/6D about rendering distant small details like distant foliage. Sort of a bigger resolution vs bigger photosite match. I remember having been surprised in the past when I discovered the "winner" of a similar match. It's was not the one I would have expected.
p.3 #2 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Just added the Zeiss Distagon T* 35mm f/2 ZE
Notes on Zeiss Distagon T* 35mm f/2 ZE
Excellent sharpness across at f/8. Lens exhibits field curvature but fortunately focusing at edge doesn't noticeably reduce sharpness in the center, indicating the field curvature can be focused around. See f/8 vs f/8 edge-focus in images for a comparison. Interesting Note: The lense's infinity hard-stop corresponds to the best edge focus(on my copy anyway), making focus a breeze.
p.3 #3 · 5DSR Mega Landscape Lens Test [Updated 7/27]
snapsy wrote:
Just added the Zeiss Distagon T* 35mm f/2 ZE
Notes on Zeiss Distagon T* 35mm f/2 ZE
Excellent sharpness across at f/8. Lens exhibits field curvature but fortunately focusing at edge doesn't noticeably reduce sharpness in the center, indicating the field curvature can be focused around. See f/8 vs f/8 edge-focus in images for a comparison. Interesting Note: The lense's infinity hard-stop corresponds to the best edge focus(on my copy anyway), making focus a breeze.
Both the Zeiss ZE 50 f/2 MP and ZE 100 f/2 MP perform beautifully on the 5DS. My ZE 50MP has a hard infinity stop. Put it on f/8, turn the focus ring all the way, and you're there. The ZE 100 infinity is a hair short of the hard stop.
p.3 #5 · 5DSR Mega Landscape Lens Test [Updated 7/27]
The 24-105 truly sucks at the corners both at the wider and the longer ends. It has always been a ho-hum lens albeit a very useful one but on 5DS it is beginning to look like a real dog. This is the one lens I wish Canon would update.
p.3 #6 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Peter Figen wrote:
It's interesting to see how good some of the lenses are and how some just never come up at the edges, let alone the corners. The new 16-35 looks to be about the best of your bunch here overall, and especially at the wider apertures, while the old in the tooth 24-105, well, let's just say that it's never been one of my faves.
I wonder it focusing half way between the center and the edge might have improved the overall performance of some of the lenses that were really soft on the edges.
Interesting test nevertheless.
I did a check of both field curvature and focus shift on the Canon 24-70 f/4 IS, 24-105 f/4 IS, and Tamron 24-70 VC for all previously tested focal lengths. I tested by performing 3 exposures at f/8 using three focus methods: 1) Focus at center wide-open, 2) Focus at center stopped-down, and 3) Focus at edge stopped down. All focus performed manually using 16x LV magnification and a Zacuto magnifying LCD loupe.
p.3 #8 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Wearing my hat as a commercial retoucher, I've had to work on more than my share of 24-105 files, and all I have to say is that's a lens I would never consider owning. Not sharp in the corners and an appalling amount of distortion at both ends of the zoom range. Yeah, you can correct that, but often at the expense of some of the resolution you're already left wanting for.
p.3 #9 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Added Samyang 14mm f/2.8 (Bower-branded, F-Mount with F-Mount to EF adapter). This is a golden copy of the lens that is sharp edge-to-edge; I handled 10 copies of this design on the Canon, Nikon, and Sony FE mounts and this is the only copy that I've had that is well-centered. The sharpness across the frame at f/8 is mind-boggling for a $300 14mm lens.
The lens has significant field curvature and focusing for edge sharpness noticeably reduces center sharpness. It also has some mid-field softness and all kinds of crazy wavy fields for those familiar with the design. Due to the significant curvature I decided to use edge focus for the entire aperture series; I also provide a center vs edge focus sample at f/8. Lastly, I've include a post-processed sample at f/8 which includes additional contrast and sharpening, for those who would like to see what the lens can produced with typical processing.
p.3 #10 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Peter Figen wrote:
Wearing my hat as a commercial retoucher, I've had to work on more than my share of 24-105 files, and all I have to say is that's a lens I would never consider owning. Not sharp in the corners and an appalling amount of distortion at both ends of the zoom range. Yeah, you can correct that, but often at the expense of some of the resolution you're already left wanting for.
And yet, I would wager that I've made more money from images taken with that lens than from all of the other lenses I've owned (which is an awful lot...) combined.
I do like the 24-70 f4L IS though, and no longer own a 24-105L.
p.3 #11 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Added Zeiss Makro-Planar T* 50mm f/2 (F-Mount with F-Mount to EF adapter). Wow, just wow - sharpest lens I've tested thus far across the frame at f/5.6 I'm surprised because I remember the lens having edge issues on the D800 - maybe the D800's horrible LV made focusing more difficult. No noticeable focus shift or field curvature. Unreal how a lens optimized for macro performance can be so good at infinity too.
p.3 #12 · 5DSR Mega Landscape Lens Test [Updated 7/27]
snapsy wrote:
Added Samyang 14mm f/2.8 (Bower-branded, F-Mount with F-Mount to EF adapter). This is a golden copy of the lens that is sharp edge-to-edge; I handled 10 copies of this design on the Canon, Nikon, and Sony FE mounts and this is the only copy that I've had that is well-centered. The sharpness across the frame at f/8 is mind-boggling for a $300 14mm lens.
The lens has significant field curvature and focusing for edge sharpness noticeably reduces center sharpness. It also has some mid-field softness and all kinds of crazy wavy fields for those familiar with the design. Due to the significant curvature I decided to use edge focus for the entire aperture series; I also provide a center vs edge focus sample at f/8. Lastly, I've include a post-processed sample at f/8 which includes additional contrast and sharpening, for those who would like to see what the lens can produced with typical processing....Show more →
snapsy, great work.
that is one lens that I set to f8/f11, and set my focus ring to 1.5 meters (which is really 3 meters in reality) and get very good center and corner sharpness out of it.
the other thing is, I wish we could make this entire thread a sticky. Its one of the most useful gear threads I have seen.
p.3 #13 · 5DSR Mega Landscape Lens Test [Updated 7/27]
snapsy wrote:
Added Zeiss Makro-Planar T* 50mm f/2 (F-Mount with F-Mount to EF adapter). Wow, just wow - sharpest lens I've tested thus far across the frame at f/5.6 ...
p.3 #14 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Snapsy -- Awesome work! Great effort, and truly appreciated by all.
I hate to cast myself as the 24-105L Apologist, but I've had really great results with mine. I'm going to try to provide links to two versions of an image shot last weekend in RAW and converted with minimal PP in LR6 to jpgs and hosted on Photobucket (haven't used that account in forever!). We'll see if it works.
The shot was a "warm-up", so no special attention paid except to exposure and general framing/focal length. Please ignore the people's expressions and tilting of the camera!
The first version is with the lens profile turned off and the CA correction turned off. Yes, there is some CA, but overall sharpness at f/5.6 is pretty astounding, and fringing only shows up (mildly, but notably) on high contrast edges, particularly on the image edges.
The second version is with the lens profile turned on and CA set to "1". This eliminates all CA fringing for my purposes.
I think in tests, along the lines of those you have provided, it is good to show results, or remind readers, that the CA is quite correctable in most lenses and doesn't significantly impact edge sharpness -- certainly improves over the CA smearing. I'm a fan of showing images with normal PP and in context of final use.
Let's see if this works where you all can see the images at 100% image. If so, please note high contrast edge details and center sharpness (somewhat behind the driftwood horse sculpture).
(PS -- Well, that didn't do the 100% I'd hoped. Let me see if I can pull some 100% crops of details. )
p.3 #15 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Maybe now. . . ?
100% crops with no profile or CA correction vs. crops with profile and CA turned on. LR6 2015CC for PP.
Please look to the area of the window dividers on the right crop, and the trees/sky-tent-light globes on the left crop.
I don't need to show the center, do I? It's even sharper and less CA in the uncorrected and corrected versions.
The caveat to introducing corrections is as Peter stated, resolution suffers slightly, as can be seen in comparing the lettering on the sign from crop on the far right. Less correction could be applied, or more deftly applied -- these are fast and dirty crops for illustration of my experience with the 24-105L.
I don't doubt there is wide variation in production of this lens, as many experienced owners have stated, but I overcame my own reluctance and finally purchased about two years ago, and now it is my most used, go-to lens, unless specific other needs present themselves (TSE, UWA, etc.).
I'm open to someone questioning my sanity or powers of observation! To me, this shot looks pretty good un-corrected, and great corrected, considering shot on the fly, hand held for a simple "get-in-the-groove-of-the-shoot" shot. I'm open to improving my work if anyone has a more practical suggestion at reasonable price for higher IQ with the zoom range.
p.3 #16 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Added Canon 85mm f/1.8. Another fantastic performance from a Canon prime. Sharp across the frame at f/5.6...gets just a tiny bit sharper at f/8. No noticeable focus shift or field curvature.
p.3 #19 · 5DSR Mega Landscape Lens Test [Updated 7/27]
I think there is a big missing factor for both the initial studies and the above discussion. Lenses have some individual variation. That is likely to be worse in inexpensive lenses but much less for high quality, L, lenses. Lenses also are very complex and have compromises and known performance limitations. Canon claims that their DLO, digital lens optimization, software can correct many of these deficiencies and will even correct softening due to diffraction. Since I do not have a 5DSR I will not comment further except that for my limited camera and lenses DLO does seem to make substantial improvements. Has anyone tried DLO with the 5DSR?
p.3 #20 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Any chance of posting the 16-35 f2.8 L II? They say it's virtually the same as f4 L IS at f8 or f11. But mine is pretty unsatisfactory even on 1.3x 1D IV between 16-20mm.
Anyway, this is an interesting and helpful thread, thanks for all that hard work!
)
