5DSR 24-70II 2.8 vs 24-105 F4 resolution test at 50 and 70mm

- similar test as before with the longer lenses but much closer distance - 14.5feet away from target
- target on wall that you see blown up is 4 inches by 3 inches
- the crop used is basically equivalent to what a 3:1 crop is in LR (the files are just over 100k)

This is what it looks like at 50mm

IMAGE LINK: https://flic.kr/p/vENN​d2

IMAGE LINK: https://flic.kr/p/wBDW​w6

IMAGE LINK: https://flic.kr/p/wBDW​LK

IMAGE LINK: https://flic.kr/p/vEMa​5X

IMAGE LINK: https://flic.kr/p/wAX8​VN

IMAGE LINK: https://flic.kr/p/wk3i​W3

continued at 70mm

24-70 F2.8 II
- 24-70 @ F2.8 @70mm

IMAGE LINK: https://flic.kr/p/wk9V​tn

IMAGE LINK: https://flic.kr/p/wBDW​HZ

IMAGE LINK: https://flic.kr/p/wk2Q​3A

IMAGE LINK: https://flic.kr/p/wzkc​1s

IMAGE LINK: https://flic.kr/p/wBDW​wB

I was a bit surprised that the F5.6 70mm of the 24-105 seems to be worse than the F4 picture - the exif data shows it was at 1/20 versus 1/40 - all shots were tripod with 10 second delay so I think this shot highlights how sensitive a 'perfect shot' at this pixel density is hard to get if any vibration happens.

Because of the size of the crops, this test is only comparing the dead center of each lens - I thought a bigger difference would show up at this amount of magnification but when you look at the slrgear.com blur chart of the 24-105 it is a very sharp F5.6 lens in the lower half of it's zoom range - the results here show it to the very similar (in the center!!) as the $2k 24-70 at F4 / F5.6

That's what I see

Did you remember to turn off the IS if the camera was on a tripod?

It's always easy to sit on the sidelines and complain, but... the 24-105 @ F5.6 @70mm shot looks to have some motion blur, and most of the shots are down at a shutter speed that at least has a risk of some potential movement (even vibration in the tripod from someone moving in the room etc.)

When I do these sorts of tests I:

1. Use Liveview or mirror lock up to prevent mirror slap
2. Use a remote cable + 2s delay to try to prevent any of my movement affecting the camera
3. Use a flash (on which I change the power as I change the aperture)

The last point allows shooting at the flash sync speed (e.g. 1/200) and ensures a pretty constant (and consistent) light level for each shot. It also allows shooting at the lowest possible ISO so that sensor noise doesn't affect the results.

One of the things that also makes me slightly suspicious is that your 24-105 @ 50mm @ F4 shot looks if anything sharper than your 24-70 II @ 50mm @ F4 image. Any test I've ever done (or MTF chart I've seen) would put the 24-70 quite a bit ahead at those settings. In the centre at least it even tends to be sharper at f/2.8 than the 24-105 at f/4.

at such a low shutter speed make sure you use mirror lockup as well. Just to make sure that anything that could effect image is disabled.

Patman99 wrote in post #17650197
Did you remember to turn off the IS if the camera was on a tripod?

The 24-70 2.8 has no IS and no left it on which possible could have helped the 24-105 as I have seen it do with my longer lenses.

sploo wrote in post #17650446
It's always easy to sit on the sidelines and complain, but... the 24-105 @ F5.6 @70mm shot looks to have some motion blur, and most of the shots are down at a shutter speed that at least has a risk of some potential movement (even vibration in the tripod from someone moving in the room etc.)

When I do these sorts of tests I:

1. Use Liveview or mirror lock up to prevent mirror slap
2. Use a remote cable + 2s delay to try to prevent any of my movement affecting the camera
3. Use a flash (on which I change the power as I change the aperture)

The last point allows shooting at the flash sync speed (e.g. 1/200) and ensures a pretty constant (and consistent) light level for each shot. It also allows shooting at the lowest possible ISO so that sensor noise doesn't affect the results.

One of the things that also makes me slightly suspicious is that your 24-105 @ 50mm @ F4 shot looks if anything sharper than your 24-70 II @ 50mm @ F4 image. Any test I've ever done (or MTF chart I've seen) would put the 24-70 quite a bit ahead at those settings. In the centre at least it even tends to be sharper at f/2.8 than the 24-105 at f/4.

I appreciate all your comments. I not exactly happy with these particular results myself so I think k I may redo them. But the results themselves do show that when you are searching for the final 10% of details that the 5dsr can give you it is not easy.

I did use live view with 10 second delay.

Thanks for the suggestions.

wallstreetoneil wrote in post #17651125
I appreciate all your comments. I not exactly happy with these particular results myself so I think k I may redo them. But the results themselves do show that when you are searching for the final 10% of details that the 5dsr can give you it is not easy.

I did use live view with 10 second delay.

Thanks for the suggestions.

Absolutely - it does show how unforgiving the 5DS can be. I've seen a few posts from people saying that they've had to up their shutter speeds for handheld work.

Don't overlook the flash option though - you can get really consistent conditions that makes it much easier to compare shots at different apertures, or even different lenses.

wallstreetoneil wrote in post #17651120
The 24-70 2.8 has no IS and no left it on which possible could have helped the 24-105 as I have seen it do with my longer lenses.

Conventional wisdom is to disable IS when the camera is on a tripod, as IS is correcting for slight hand-held shaking, but when the camera is on a tripod, IS can cause some blur in photos when it is enabled. Go figure.

Patman99 wrote in post #17651328
Conventional wisdom is to disable IS when the camera is on a tripod, as IS is correcting for slight hand-held shaking, but when the camera is on a tripod, IS can cause some blur in photos when it is enabled. Go figure.

with all my IS lens, i get the complete opposite (70-200, 100-400, 24-105)

i sort of came upon this while doing a dot tune of my lenses and realized that with the IS on, DOT tune method would be able to achieve critical focus further from the center marker by a few clicks so I kept it on as I view it as more accurate (i.e. eliminating movement)