Specularist wrote:
The MTF curves can't be directly compared, since Zeiss curves are based on measured (achieved) results, while Canon's are calculated based on perfect construction and, importantly, a world without diffraction. Perhaps we can all agree that Canon and Zeiss lenses are equally subject to diffraction.
What you can estimate from the curves, though, is that the Zeiss has less lateral chromatic aberration than the Canon. This is also clear from the TDP results — flawed as they are — and sample photos I've seen elsewhere, not to mention most other reviews such as the one at Photozone.de, etc.
No, this is wrong. You cannot infer from MTF curves anything about chromatic aberration of ANY sort, because the published data are not specified by wavelength. CA is not a Seidel aberration! That is to say, if the MTF decreases as a function of image height, it could just as easily be due to field curvature, astigmatism, or coma, and it cannot be assumed that CA is the cause.
Furthermore, to say on the one hand that the MTF charts between two manufacturers generally cannot be directly compared--which is absolutely true (but not necessarily for the reasons you state), and then to say that you can pick out chromatic aberration as a cause of the relative MTF behaviors, is self-contradictory.
Your apparent lack of understanding on this matter calls into question your ability to correctly interpret an MTF curve.
gdanmitchell wrote:
Wait. They would be "equally attractive to [your] eye" but one would look better than the other? Isn't that sort of like "they both taste just as good but one tastes better?"
Dan
nah... more like I like chocolate icecream, and I like vanilla icecream. Which I prefer on the day is dependent on the mood I'm in.
wickerprints wrote:
You just don't get it, do you? We are NOT comparing other pairs of lenses here. We are saying, show everybody a situation in which the 100 MP will take an image with this desired "look" but the 100L, under the same conditions, will not.
Whoever has that time and patience can provide that evidence for you. Probably be a big waste of time for them as you wouldn't accept it anyways.
wickerprints wrote:
You just don't get it, do you? We are NOT comparing other pairs of lenses here. We are saying, show everybody a situation in which the 100 MP will take an image with this desired "look" but the 100L, under the same conditions, will not.
SKumar25 wrote:
Whoever has that time and patience can provide that evidence for you. Probably be a big waste of time for them as you wouldn't accept it anyways.
Aravind already posted the side-by-side comparisons. There was no reason to dispute them. I find it rich that you imply I would be selective in accepting the results, yet you ignore the results that have shown there is no significant difference. You're the one making the claim. The fact that you refuse to back it up with a simple controlled test shows your true colors.
I am not the only one asking for this evidence, and it would not be for my sole benefit, but for anyone else interested the matter to witness. It's certainly your prerogative to not provide the support to your claim--it's also our prerogative to dismiss your claim for the lack of such support. In other words, put up, or shut up.
88% correct identification of a Zeiss Vs Canon lens. Blind test, read the post from the beginning. In general, Zeiss lenses have noticeably more local contrast. But the differences aren't always complimentary, ie the Zeiss 21/2.8 vignettes more than the Canon 16-35 II at f4. Of course since its a blind test you don't know which Zeiss or Canon is being tested.
Mike K
gdanmitchell wrote:
Can you define "per pixel sharpness?" I do not understand how a pixel can have "sharpness."
Sorry, stupid use of words. Disregard. What I mean to say is that if you view the images at full size (which is what I linked to, but on some browsers may get squished down to fit on the page) the image is sharp even though the reported size is only 1200 px wide. This leads me to conclude that some kind of sharpening was applied on these images after downsampling.
But more contrast (which is something you totally can't do in post) is just less dynamic range. Why do you want that? Also, if I want sharpness, I turn up sharpening. I don't want my pictures to be Ken Rockwell status all the damn time when I don't want to.
dudemanppl wrote:
But more contrast (which is something you totally can't do in post) is just less dynamic range. Why do you want that? Also, if I want sharpness, I turn up sharpening. I don't want my pictures to be Ken Rockwell status all the damn time when I don't want to.
Big difference between local contrast and overall contrast range. The local contrast impacts the acutance. The following description of acutance comes from Luminous-landscape: "Acutance isn't about resolving detail, it's about the transition between edges. In other words when an edge changes from one brightness level to another. This is what Sharpening in digital parlance is all about. Scanning and digital capture softens acutance and so we apply a (ill-named) process called an Unsharp Mask to increase edge sharpness back to what it should be. Remember, this has nothing to do with resolution, the other aspect of sharpness."
Do you feel the Zeiss images are overly sharpened without adding any additional USM? See if the crop examples between Zeiss and Canon lenses in the FM post "a small experiment" which I referenced my previous post bears this out?
Mike K
wickerprints wrote:
Aravind already posted the side-by-side comparisons. There was no reason to dispute them. I find it rich that you imply I would be selective in accepting the results, yet you ignore the results that have shown there is no significant difference. You're the one making the claim. The fact that you refuse to back it up with a simple controlled test shows your true colors.
I am not the only one asking for this evidence, and it would not be for my sole benefit, but for anyone else interested the matter to witness. It's certainly your prerogative to not provide the support to your claim--it's also our prerogative to dismiss your claim for the lack of such support. In other words, put up, or shut up....Show more →
As I already said, the Zeiss look is not there in every picture. Aravind's comparison is good to compare resolution, however even to me there is not much difference in rendering.
Interestingly Aravind is the first guy I've come across who uses Zeisses and does not see a difference in rendering between Zeiss and Canon.
Others have linked other reviews but you already dismissed those as biased.
Rather than put it to me to provide the evidence, I'd suggest you try a few Zeisses out. I am familiar with both Canon and Zeiss lenses as almost all people who shoot Zeiss ZEs are. If people think what we're on about is rubbish then why not try it out. You might be surprised.
88% correct identification of a Zeiss Vs Canon lens. Blind test, read the post from the beginning. In general, Zeiss lenses have noticeably more local contrast. But the differences aren't always complimentary, ie the Zeiss 21/2.8 vignettes more than the Canon 16-35 II at f4. Of course since its a blind test you don't know which Zeiss or Canon is being tested.
Mike K
"Unsharp mask" is actually not "ill named" at all. A "mask" was created using an "unsharp" version of the image, which combined with the original negative increased local contrast near edges.
It isn't that it is a "mask that makes things unsharp." It is the use of a mask that is itself unsharp" to increase local contrast.
Dan
Mike K wrote:
Big difference between local contrast and overall contrast range. The local contrast impacts the accutance. The following description of accutance comes from Luminous-landscape: "Acutance isn't about resolving detail, it's about the transition between edges. In other words when an edge changes from one brightness level to another. This is what Sharpening in digital parlance is all about. Scanning and digital capture softens acutance and so we apply a (ill-named) process called an Unsharp Mask to increase edge sharpness back to what it should be. Remember, this has nothing to do with resolution, the other aspect of sharpness."
Do you feel the Zeiss images are overly sharpened without adding any additional USM? See if the crop examples between Zeiss and Canon lenses in the FM post "a small experiment" which I referenced my previous post bears this out?
Mike K
As I wrote earlier: "Several powerful forces are at work here. In the lineage of the "emperor's new clothes," few want to admit that they don't see the difference that everyone else sees, much less the difference that the really smart people claim to see. Sometimes though, especially in areas like wine and sound equipment and photographic equipment, where a certain percentage of people use the field as a way to demonstrate their superiority by acquiring "the best stuff," certain people are really, really invested in having the best stuff. And another issue is that once one invests (mentally and financially) in "the best," it is very, very difficult to see rationally when owning the best makes no discernible difference in the final product. Making things more complex is the fact that the quality of the stuff does matter - though not quite in the way that these folks imagine."...Show more →
That assumes that Zeiss lenses are people's most expensive and exclusive lenses. It is not always the case. Many people own Leica or expensive/exclusive Canon lenses as well and still talk about the special rendering of Zeiss.
thedigitalbean wrote:
Fascinating. The only comparison in that series that is at all relevant to this thread is "D" the comparison between the Zeiss 100 MP and the Canon 100 Macro.
What I find curious is this, here is the Zeiss image:
Here is the Canon image:
The difference in per pixel sharpness is great, so great that I have to wonder if its real. It looks like the Zeiss image has had some kind of post resizing sharpening applied and the Canon one hasn't.
Regardless its not something I've been able to reproduce try as I might.
This makes the discussion very interesting to me, because I believe those images (I removed them from the quote) are processed exactly the same way.
I also believe that your side-by-side comparisons are as accurate as can be.
I accept the different results, and I accept that we do not have a plausible explanation at the moment. To me, it is not technique, it is not lens quality copy variation, and it is not confirmation bias. But it has to be something.
gdanmitchell wrote:
"Unsharp mask" is actually not "ill named" at all. A "mask" was created using an "unsharp" version of the image, which combined with the original negative increased local contrast near edges.
It isn't that it is a "mask that makes things unsharp." It is the use of a mask that is itself unsharp" to increase local contrast.
Dan
Agreed, that is the derivation of the term. Wording is Reichmann's and perhaps a little out of context. I was trying to distinguish between overall image contrast and perceived detail sharpness.
Mike K
Mirek Elsner wrote:
That assumes that Zeiss lenses are people's most expensive and exclusive lenses. It is not always the case. Many people own Leica or expensive/exclusive Canon lenses as well and still talk about the special rendering of Zeiss.
Yes it's funny that people always assume that. My five Zeiss ZE lenses has a much lower average price than the average Canon lens that I own. My ZE 100 MP lens is a low priced lens compared to many Canon lenses that I own
Can you PLEASE stop injecting other people's photographs into this thread without their consent? It is especially inappropriate because these images are not taken with the specific lens(es) in question, which is clearly stated in the title of the thread. If you continue with this behavior I will report you.
These images you are posting, taken with other Zeiss lenses, are not germane to the discussion, given that some of these don't even have a Canon counterpart, such as the 21/2.8 Distagon. And then you make comparisons between the 50/2 MP against the 50/1.2L, which are completely different types of lenses--one is a semi-macro at f/2, the other is a portrait lens that is more than twice as fast. And you just keep repeating "from my experience," as if we are supposed to simply accept this without qualification. Nobody cares about your "experience." If you haven't figured it out by now, we are talking about two lenses: The 100 MP and the 100L.
I post here against better judgement, but I have been asked by PM to comment and as I was the executor of the blind Canon/Zeiss test I'll answer some of the questions raised.
The purpose of the test was not to make a quantitative comparison but a qualitative one - i.e to see if there was such a thing as Zeiss/Canon drawing style. The 1200px images were developed from RAW using Lightroom's default settings with the "Canon Standard" camera profile. The images were then resized and sharpened using the same script. The shots were taken from a tripod. Focus was done with live view and a 10 second timer was used for the shutter release.
The reason for doing it at two forums is because the people that frequent the alt forum tend to have a Zeiss bias and the people that post at POTN tend to have a Canon bias. Also the people in the alt forum tend to have a solid experiences with a variety of lenses while the POTN population mostly consists of people just beginning with photography.
I also made use of control samples. Here at FM I arranged a set of images shot with both lenses that were exceptionally difficult to distinguish (by design). How good that was as a control set can be debated but it was an afterthought. Over at POTN I did it properly with a Leica control sample. The reason for using controls was to see that people weren't simply copying each others' results.
Without getting into a discussion which is better, it is clear that there is a difference between Canon and Zeiss and that people with varying photographic experience can tell them apart with statistical certainty. Also the control samples showed that people weren't cheating in either forum (by copying the answers from other people).
The FM control set also shows that you can take pictures with both brands that are indistinguishable to the same population of test subjects. This should of course be obvious to anyone. Just shoot both lenses with the lens cap on and you'll get identical results, if you want an extreme example. With real world images you'll find a large spread ranging from indistinguishable to very different.
What the experiment showed was that when given a set of photos where there was some visible difference, a population of people interested in photography could with a very high degree of statistical certainty tell Zeiss and Canon rendering apart.
wickerprints wrote:
You cannot infer from MTF curves anything about chromatic aberration of ANY sort, because the published data are not specified by wavelength. CA is not a Seidel aberration! That is to say, if the MTF decreases as a function of image height, it could just as easily be due to field curvature, astigmatism, or coma, and it cannot be assumed that CA is the cause.
In this case we know field curvature, astigmatism, etc., are low, so we can estimate from the MTF curves that the Canon has more lateral chromatic aberration. LCA reduces the contrast of tangential structures but not sagittal structures, worsens with image height, and isn't affected by stopping down the aperture. And as the MTF charts of the Canon show a poorer improvement in tangential structures in the field when the lens is stopped down, yet the sagittal structures improve strongly, LCA is likely to blame.
If someone with the two lenses does a test for lateral chromatic aberration, the differences will be obvious to all.
wickerprints wrote:
Can you PLEASE stop injecting other people's photographs into this thread without their consent? It is especially inappropriate because these images are not taken with the specific lens(es) in question, which is clearly stated in the title of the thread. If you continue with this behavior I will report you.
These images you are posting, taken with other Zeiss lenses, are not germane to the discussion, given that some of these don't even have a Canon counterpart, such as the 21/2.8 Distagon. And then you make comparisons between the 50/2 MP against the 50/1.2L, which are completely different types of lenses--one is a semi-macro at f/2, the other is a portrait lens that is more than twice as fast. And you just keep repeating "from my experience," as if we are supposed to simply accept this without qualification. Nobody cares about your "experience." If you haven't figured it out by now, we are talking about two lenses: The 100 MP and the 100L....Show more →
Many (even recently) have posted blanket statements ridiculing rendering differences others have pointed out between z and c. How can we go further if we can't even agree on this? That is the significance of the post.
I have not come across any rule related to linking. Go ahead and report me...
