p.1 #1 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Most lens tests available online are performed at close focus distances and don't necessarily provide a good indication on how those lenses will perform @ infinity for landscapes, particularly at the edges of the frame. For those intending to use the 5DS/R for landscapes here's a running test I've been performing @ infinity on my lenses. I've tested 5 lenses so far and intend to add more later.
All images focused in LV at center of frame on far mountain and shot in LV EFCS+timer release. Processed in LR6 with sharpening set to 35/0.8/35. No lens corrections applied. Some exposure adjustments were needed due to the high DR of my chosen scene; in particular the pine trees at the rightmost edge of the frame have had their exposure lifted. Note that the pine needle bundles on those trees may appear unsharp to those not accustomed to photographing them at distance - that's the nature of the subject. Also note that decentering/tilted issues become much more obvious at 50MP - when evaluating these images be sure to check both edges and favor the sharper side when evaluating what the optical design for that lens is capable of.
Here are the full-sized 50MP images for the lenses I've tested so far:
p.1 #2 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Impressive effort!
I sometimes do tests like that — and they can be quite useful when trying to "come up to speed" with new gear quickly — but I haven't done anything that methodical with my 5DsR and my four L zooms.
However, I have now had the opportunity to do some serious field photography with the new body. Shooting entirely from the tripod, using live view for all photographs (except a few wildlife subjects that walked into my view), and doing things like waiting for the breeze to stop before activating the shutter, the results are very, very good.
I use four lenses for car-based landscape photography, including work that involves day hiking. (I use a slightly different setup sometimes when backpacking.)
16-35mm f/4L IS — I had no occasion to use this lens on the recent shoot. Stay tuned.
24-70mm f/2.8 L II — A few people "who know things" had expressed concern that this lens might not be able to resolve as much detail as some of the others. I did not believe this would be the case, and the many photographs I made with the lens confirm this. It is more than "up to the" (so-called) "challenge" of the 50MP sensor.
70-200mm f/2.8L IS II — I had no doubts about the ability of this lens to produce excellent resolution results, and no doubts are warranted. Results are excellent.
100-400mm f/4.5-5.6L IS II - Throughout most of its focal length range this lens is also top notch. In fact, in cases where I might have switched to the 70-200 for the best resolution, I usually ended up just using the 100-400 and the results are great. The lens becomes slightly softer at 400mm. It still resolves detail very well, but slightly less so at 400mm — though I don't hesitate to shoot it at 400, and I'll be happy to make very large prints from 400mm shots.
There has been a lot of talk about the need to shoot at the so-called diffraction limited aperture with the 50MP sensor. I'm more oriented to practical and experience-based decisions about such things. When no other particular aperture was called for I tended to shoot this camera at f/8 for landscapes. The results are excellent. I was interested in what I would see when stopping down a bit (I had no occasion to open up larger this time), which I had to do in a number cases for DOF reasons, particularly when shooting with longer focal lengths. I can report that results at f/11 are excellent, and I don't think that photographers should hesitate to shoot at f/11 over sharpness concerns. At f/16 and 100% magnification I can see that the images are slightly softer — but I won't hesitate to use f/16 if I need it, and sharpening in post produces very good image quality. I tended to steer clear of f/22 and smaller except in very rare situations in the past, and that won't change.
If anything, at this point the camera with these lenses is actually exceeding my expectations.
Once you take the files through post and apply CA corrections and typical sharpening... the output is truly excellent.
In other news, the 5Ds R seems to chew through batteries faster than my 5DII in live view mode. I'm not thrilled with that realization, but I'll pick up and extra battery or two.
p.1 #4 · 5DSR Mega Landscape Lens Test [Updated 7/27]
I'm seeing similar results: excellent (exceeding expectations) with 100 mm macro (non-L), very good with a 100-400 original but revealing of slight decentering I was barely aware of on 5D2.
In all areas, 5DS meets or exceeds expectations, with the exception as others report that it uses up battery power quickly in Live View.
p.1 #5 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Excellent series. Now when is Adobe going to get up to snuff with Phase One and give us a lens sharpness falloff adjustment like C1 Pro? The 50mp files would really be shining with the ACR workflow usability.
p.1 #6 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Are they all supposed to be full res. I clicked on the 16-35 at f/4 and it was a full res file but the other three came down as low res 800 wide versions? I would love to see all of them at full res to pixel peep the crap out of them.
p.1 #7 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Nice test shot series.
What is your take on the "optimal" aperture for use with the 16-35/4 ? I didn't see anything (tired eyes ) that presented me with an obvious "best" ... suggesting, it was equally good @ all apertures. My general thought is to get off of WO, but then to use f/ 5.0 - 5.6 @ 16mm / 5.0 - 5.6 = 3.2mm - 2.85mm. My perspective is that once you get below, 2mm, diffraction starts to kick in a bit more, so I typically try to stay south of f/8 if I can (for 16mm) @ that wide.
p.1 #8 · 5DSR Mega Landscape Lens Test [Updated 7/27]
arbitrage wrote:
Are they all supposed to be full res. I clicked on the 16-35 at f/4 and it was a full res file but the other three came down as low res 800 wide versions? I would love to see all of them at full res to pixel peep the crap out of them.
Thanks for the effort.
They're all full-res. Your browser may be resizing the initial render to fit to your screen. You should be able to click on the image for the full-size view after it finishes downloading.
p.1 #9 · 5DSR Mega Landscape Lens Test [Updated 7/27]
RustyBug wrote:
Nice test shot series.
What is your take on the "optimal" aperture for use with the 16-35/4 ? I didn't see anything (tired eyes ) that presented me with an obvious "best" ... suggesting, it was equally good @ all apertures. My general thought is to get off of WO, but then to use f/ 5.0 - 5.6 @ 16mm / 5.0 - 5.6 = 3.2mm - 2.85mm. My perspective is that once you get below, 2mm, diffraction starts to kick in a bit more, so I typically try to stay south of f/8 if I can (for 16mm) @ that wide. ...Show more →
Generally speaking the 5DS/R will yield the highest resolution @ f/5.6 and larger. Then it's just a matter of what aperture is needed to get the most out of the corners for a particular lens. On my 16-35mm f/4 the corners approach their peak @ f/5.6 for 16mm, 20mm, 24mm, and 35mm. For 28mm the corners need a little more help and do better at f/8.
p.1 #10 · 5DSR Mega Landscape Lens Test [Updated 7/27]
snapsy wrote:
They're all full-res. Your browser may be resizing the initial render to fit to your screen. You should be able to click on the image for the full-size view after it finishes downloading.
Yes with the few full res ones I've found that is what I did but then all the others were these super small 800 wide ones. I'm using Safari on Retina Macbook Pro 15inch so not sure what is up. Only a few I've found to be full res after clicking on a bunch. Maybe I'll try Chrome and see how that goes...
Weird, now they are all downloading at full res....not sure what was up 30 mins ago. Thanks again for these shots.
p.1 #11 · 5DSR Mega Landscape Lens Test [Updated 7/27]
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.
p.1 #12 · 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 also wonder how many of the edge/corner issues are caused by field curvature?
Nobody seems to measure it anymore because cameras are used for testing lenses instead of proper instruments.
p.1 #13 · 5DSR Mega Landscape Lens Test [Updated 7/27]
That is a good question. My old Nikon wide angle lenses had terrific field curvature which on some scenes was great and on others horrible. I think that with the instant feedback it's easy enough for the average photographer to just go out and shoot and look at it right there. The other thing is that often more controlled testing in a lab is too often done at much closer distances than one would normally shoot at, especially with the ultra wide angle lenses, and that puts their floating elements into the equation and often times shows more or less field curvature and more or less distortion than you'd get at longer distances. This is why I printed a giant test chart for the wall in my studio that is something like 8'x13'. I'm still too close with ultra wides, but it's a start.
p.1 #14 · 5DSR Mega Landscape Lens Test [Updated 7/27]
It's all relative. As Roger at LR said, it's not like these lenses are going to be worse on a 5DsR than on a 5D3, it's just the difference between these zooms and a top level prime lens like a zeiss 21 or TS-E 24 will be even more evident.
p.1 #15 · 5DSR Mega Landscape Lens Test [Updated 7/27]
I think what you end up seeing, and especially in a lens like the 24-105, is that on a lower res body, there's less of a difference between the center and edge, but on these new cameras, then center is sharper but the edges are still pretty much where they always were, creating an exaggerated representation between the center and the edges. Already, on a 5D3, the 24 T/S-E falls off quite a bit toward the edges, but nothing like the aforementioned zoom. If you are going to make large prints, then this may indeed be something that needs to be considered, especially for landscape people who tend to want the whole thing crispy.
p.1 #16 · 5DSR Mega Landscape Lens Test [Updated 7/27]
gdanmitchell wrote: 100-400mm f/4.5-5.6L IS II - Throughout most of its focal length range this lens is also top notch. In fact, in cases where I might have switched to the 70-200 for the best resolution, I usually ended up just using the 100-400 and the results are great. The lens becomes slightly softer at 400mm. It still resolves detail very well, but slightly less so at 400mm — though I don't hesitate to shoot it at 400, and I'll be happy to make very large prints from 400mm shots.
Dan, it is really good to read your report about the 100-400L II. I am thinking about not having a 70-200mm and just going with the 100-400L II. For scenarios where it is fitting and when optimal resolution is desired, I can utilize my Zeiss 2/135mm.
p.1 #17 · 5DSR Mega Landscape Lens Test [Updated 7/27]
Peter Figen wrote:
I think what you end up seeing, and especially in a lens like the 24-105, is that on a lower res body, there's less of a difference between the center and edge, but on these new cameras, then center is sharper but the edges are still pretty much where they always were, creating an exaggerated representation between the center and the edges. Already, on a 5D3, the 24 T/S-E falls off quite a bit toward the edges, but nothing like the aforementioned zoom. If you are going to make large prints, then this may indeed be something that needs to be considered, especially for landscape people who tend to want the whole thing crispy....Show more →
I agree - I've written about center vs edge delta before in higher MP sensors an it's not just a problem with marginal lenses although naturally it's more obvious with those. For example, Roger's initial 5DSR resolution tests show that corners on the 85mm Otus trail the center by a whopping 40% @ f/2.8 and the Zeiss corners near their peak at that aperture (lenstip review).
Good point about field curvature. I'll integrate that check into other lenses I plan to test for this thread and I'll also do a spot retest of lenses I've already posted.
p.1 #18 · 5DSR Mega Landscape Lens Test [Updated 7/27]
darbo wrote:
Dan, it is really good to read your report about the 100-400L II. I am thinking about not having a 70-200mm and just going with the 100-400L II. For scenarios where it is fitting and when optimal resolution is desired, I can utilize my Zeiss 2/135mm.
I just sold my 70-200 f2.8 IS v1 and got a 100-400 f2 for the same reason. I figured the gap from 70-100 is pretty small.
p.1 #19 · 5DSR Mega Landscape Lens Test [Updated 7/27]
darbo wrote:
Dan, it is really good to read your report about the 100-400L II. I am thinking about not having a 70-200mm and just going with the 100-400L II. For scenarios where it is fitting and when optimal resolution is desired, I can utilize my Zeiss 2/135mm.
There are situations in which not having a 70-200 and having a 100-400 could be very viable. While the 70-200 lenses are among the best that Canon makes — true workhorse lenses — the new 100-400 seems so good in the 100-200 range that you don't give up much at all my using it instead in many situations, plus you get that added range.
A friend who made his first trip to Antarctic this past year to teach in a couple of workshops did something similar, though using the Nikon 80-400, and he used such a lens a lot with excellent results. When I photograph migratory birds, I combine this shooting with a sort of landscape photography... and I can do most of both with the 100-400.
) that presented me with an obvious "best" ... suggesting, it was equally good @ all apertures. My general thought is to get off of WO, but then to use f/ 5.0 - 5.6 @ 16mm / 5.0 - 5.6 = 3.2mm - 2.85mm. My perspective is that once you get below, 2mm, diffraction starts to kick in a bit more, so I typically try to stay south of f/8 if I can (for 16mm) @ that wide.