Carsten, they are listed on top of post #2 on this page.
Delatant wrote:
Secondly, while I throughly enjoyed the task (thanks for spending the time to put it together), almost none of these pictures would I have taken with a Zeiss lens - for the most part they were fairly monochrome - the advantage of Zeiss is the color rendition - my Zeiss goes on when color starts to splash! Zeiss is color! While other lens may approach (or at times exceed) Zeiss optically, no other lens can produce the vibrance, rendition, or saturation - ie 3-D effect or "pop" that Zeiss does. Now reproduce this test, with lots of color and I think the percentage of correctness will zoom....Show more →
I would disagree with you there - Zeiss is about contrast - local and global. The colors are a consequence (higher contrast in RGB space = higher saturation) but the differences are subtle. Canon and Zeiss are relatively close color wise - much closer than for instance Zeiss and Minolta. The camera itself plays a much larger role there. 5DII + Zeiss is much closer to 5DII + Canon in colors than 5DII + Zeiss is to M9 + Zeiss.
A less monochromatic environment would have helped to show differences in bokeh.
Anyway, I have a color test that I did not post. I was actually surprised that the difference was so big, but still, it's far more subtle than the first series. There are few hard edges where the Zeiss could do its high local contrast thing.
I had to do something a bit different though. When the images were processed exactly the same way, there was a very significant difference in global contrast between the lenses. So I applied an "Auto Contrast" in Photoshop on all the images before resizing them.
This test is a bit different - there are five images total, two where shot with Zeiss lenses. Can guess which two?
denoir wrote:
Carsten, they are listed on top of post #2 on this page.
Ah, so they are. I thought I had read the post in detail, but somehow my eyes must have moved quickly to the table. Oooooh table!
So I switched the 35 and 50 in my guess, partly because I thought you had the 50 Planar, not the Makro-Planar, and I switched the 85 and 100. I should have picked up on the fact that the second one was taken from the street, which brings it somewhat closer, but I didn't notice it yesterday. Two apertures sort of correctly guessed (implicit in my first guess was wide open), two one stop off, which I am pretty happy with. What I am not happy with is that I thought the 21 Distagon shot was at f/8 or f/11, but it was only at f/4. Apparently I need to spend more time with this lens
From the new set, I would say that #1 is for sure a Zeiss, #2 and #5 are not that great lenses (i.e. Canon) and I am guessing that #3 is the other Zeiss, although I would not be totally surprised if it were #4.
Aw crap, there's camera shake blur from Lens #2. I have removed it from the list.
I must have pushed the tripod. I was very surprised by the results as #2 is possibly the best Canon lens I've used (it's the 135L). It's ultra sharp and produces superb colors that have nothing in common with the typical Canon colors. So I was surprised to see the weak results. I just checked the full resolution image and it's blurry, which explains the lousy result.
I've removed the lens from the list of choices. Too bad. I wanted it to be the wildcard - something with a unique signature that you would not associate with Canon or Zeiss.
--
I took the shot again with the 135L but it's not usable as I had natural light when I shot the series and it's dark outside now. Anyway here are 100% crops of I2 first the one I took just now and then the one that had the camera shake:
By the way, A-E is pretty obvious. And what's the response of the people in the Canon board?
F-H is surprisingly close and difficult to tell though, I would expect MP 100/2 to be much better... Perhaps the indoor lighting doesn't bring out the best of this lens?
Yes, well done Aleksander! You are the only one that identified both sets correctly.
I didn't post it on the Canon board in the end as it looked for a while that the task A-E was way too simple.
As for F-H, I think the 100 MP did not have an opportunity to show its strength in those shots. There was nothing with fine nuanced texture. Flip between at these shots for instance - both at f/2.8:
Quite a difference, right? The F-H shots all had a lot of high contrast stuff in the plane of focus but no gradual texture or color shifts where the greater micro contrast of the 100 MP would be noticed.
denoir wrote:
Yes, well done Aleksander! You are the only one that identified both sets correctly.
I didn't post it on the Canon board in the end as it looked for a while that the task A-E was way too simple.
i didn't play because i've never shot with any of the lenses used and doubt i ever will but i'd be interested to see how the canon forum does with the test. they'd be a sort of data point halfway between alt lens nuts and people that just like to look at photos not take them (i'd love to hear how this group does as well). you'd have to re-randomize which photos are which to prevent cheating though.
Yes, OK. I'll post it on a Canon forum. Not at FM though as believe it or not there are some people that frequent both the Canon and Alt forum. I'll randomize the order of the images again and temporarily remove the original ones.
we can see that for 25 samples and with a 95% the results are statistically significant for set 1 (for that number of data points the error margin is 20%). There is one big issue though - that people can see what other people guessed. It does not have to be conscious copying of the results but it can certainly influence. This is especially the case when the choice is explained. A good example are how three people consecutively missed on D & E after a theory was put forward that one should look at diffraction stars in the images.
Set 2 was a great control experiment though. It showed that where there was objectively very little difference between the images, the results were poor. And they were not systematically poor, but randomly so. That would indicate that people are in fact not just copying the answers from previous posters. The results were truly statistically random. Had I thought of it initially I would have made a true control set, but this is good enough.
So a reasonable conclusion is that one can consistently tell Zeiss and Canon rendering styles apart (set 1) for a variety of types of shots but that it is also possible to take photos where it is very difficult to tell them apart ( set 2).
And actually it's better than that Rusty, For example I jokingly reversed all mine. But even a noob like me could tell all of them - 100% with A, B, C, E correct and D a tung-in-cheeker...
What happened with F, G, and H is what I wanna know. Why the difference?
Bifurcator wrote:
What happened with F, G, and H is what I wanna know. Why the difference?
That's easy enough. Put a lens cap on a Canon lens and a Zeiss lens and take a shot. Will you see a difference? No. So there are circumstances where you can't tell the lenses apart.
F, G, H were picked exactly because I could not see a substantial difference. I posted them after some people said the A, B,C D & E were too easy. For those particular circumstances it wouldn't matter if you used a Zeiss 100 MP or a Canon 100 Macro. Marco lenses are typically superbly sharp when doing closeups and if there are no textures with subtle color variations in the scene there won't be much of a difference.
The idea for the test was to see that for scenes where there was a substantial difference if that difference could consistently identified due to it being a Zeiss or a Canon lens.
I deliberately posted my answers before reading anyone else's. There is nothing interesting in it if I am not really trying, is there?
The last comparison shot of the ZE100 and the Canon 100 Macro is more what I was looking for in the first three sets, but didn't find. I liked the Canon macro lens initially when I had it, but with time I came to find it somewhat lacking. The 100L is meant to be better.
Excellent, and fun, but I was too late. I managed to get the first set right but not the second one. I guessed them all but won't tell how that turned out... it reminded me about certain moments back in school.
Thank you denoir!
