Home

Please support mflenses.com if you need any graphic related work order it from us, click on above banner to order!
SearchSearch MemberlistMemberlist RegisterRegister ProfileProfile Log in to check your private messagesLog in to check your private messages Log inLog in

Some bokeh from my Carl Zeiss Makro-Planar 60/2.8
Post new topic   Reply to topic
Home   Manual Focus Lenses
View previous topic :: View next topic  


PostPosted: Wed Jul 04, 2012 8:26 pm    Post subject: Reply with quote
iangreenhalgh1 wrote:
I think in both cases the backgrounds are painted.
...


Thanks for sharing this usefull information. I will have to research in this area. Never heard before about any extremely soviet sharp lenses. have to check.

As about Helios-44 - i had a lot of them. Helios-44M-4, Helios-44-2, Helios-44K-4 etc. They were not sharp and for sure they dont outresolve my 22Mpix sensor.

If we talk about your 4mpix sensor of NEX-3 we can talk about good sharpness of most lenses. 4mpix is too low resolution for testing lenses. Maybe its enough for your tasks but for testing is not.

You are telling such strange things i cant believe..

Quote:
Even the latest very dense sensors like the 36mp Sony full frame one are lower in resolution than a good film such as Fuji Velvia 50


I thought that battle digital vs film already ended with the victory of digital? Or i'm not right?

Than why all photo pro's moved to digital? They dont need resolution?

As for images made using film i seen i can say they were "poor resolution" and "very poor resolution". And have scanned a lot of films from 1980-90th and 2000th. Where is the truth?

Every new lens i get i try with my camera 5D mark II and if lens is good it shows more and more detail on image. So my sensor is far from outresolving i think. It's just fact - what i see.

Data sheets you shared are very interesting. Such a pity russian optical specialists were not publishing charts in same format as Zeiss or Leica or other Western companies do. So these charts can hardly be compared to charts of other companies without recalculation. I'm not sure i can do it correctly.

Do you have any prove about film defeating digital sensor? Is very interesting. I know only about Medium Format defeat 35mm digital. Or not defeat.


PostPosted: Wed Jul 04, 2012 8:33 pm    Post subject: Reply with quote
iangreenhalgh1 wrote:
To get back to the point about bokeh, I found my old post with 13 different 50-ish mm lenses shot wide open. I think this shows that there is less difference than most people think:

....


But i do see a big difference!

Please look at this strange flashlight behind a clock - Pancolar, Miranda and Industar shown absolutely different blurring of this area.

I suppose it was not very correct to shoot all lenses wide-open cause they have different speed. So they all will have absolutely different blurring of out-of-focus areas even if it was same lens (1.4 vs 1.8 vs 2.Cool

I know that closing down makes aperture shape affect bokeh but i dont see other way to compare more correctly.


PostPosted: Wed Jul 04, 2012 8:59 pm    Post subject: Reply with quote
NEX-3 is 14mp and very much adequate for testing.

5DMKII has 17% lower spatial resolution than my NEX-3, therefore my camera requires lenses of 17% higher resolution to produce the same apparent sharpness. It's all about pixel density. The NEX-3 packs 14mp into APS-C whereas the 5DMKII packs 22MP into Full Frame, if you do the maths (which I did a while ago) you can work out that this gives the NEX-3 a 17% denser sensor, the denser the sensor the higher the resolution it demands from a lens to give the maximum apparent sharpness.

This is a useful study of how the 14mp APS-C sensor of the NEX compares to a 24mp FF sensor:

http://www.robsphotography.co.nz/Sony-nex-A900.html

Helios-44s are sharp lenses, I don't have the resolution figures to hand but they all have over 40 lp/mm centre resolution, by the late models they had over 50. This compares very favourably with primes of the same focal length from other makers, so if the Helios isn't sharp then neither are most other lenses.

There are many reasons why professionals switched from digital to film, but resolution is certainly not one of them. The resolution of the 5DMKII is, on paper, a bit higher than Fuji Velvia 50, but there are other factors that affect the sensor such as the anti-aliasing filter. There is a lot of data out there about how film compares to digital in resolution, it would be good to study some.

What you see is not fact, you must study the hard physical and mathematical facts, and one simple fact is that to obtain maximum perceived sharpness the resolution of the lens must be equal or greater to that of the sensor. Most lenses from respected makers have had resolutions far exceeding that of film for over 60 years, until we have cameras with sensors of very high pixel density, we don't have to worry about lenses not being sharp enough. Of course, compact cameras with tiny sensors have extremely high pixel densities so those do require very high resolution lenses, but if we are talking about interchangeable lens cameras with APS-C and FF sensors then we don't have anything to worry about as regards resolution of old manual lenses and that situation won't change for quite a while, not until we have much denser sensors than what is currently available.


PostPosted: Wed Jul 04, 2012 9:20 pm    Post subject: Reply with quote
Here is some info on the Zenitar, one Soviet lens that is exceptional in resolution:

http://forum.mflenses.com/viewtopic.php?t=9504

Quote:
Zenitar 2 series was developed as a successor of Helios 44 series. One year before launch of the last Helios 44 (44M-7 MC), S.B. Byshkin won selection procedure of new optical design and a few years later the first Zenitars (K2 nad M2) went to production. The first batch of K2 versions had resolution slightly over 70/45 (>double the 44-2!) and resolution of later batches had to be slightly lowered, because film didn't manage it very well. Resolution of final batches differs form 60/30 to 68/44.

E.Y.Rybnikova was responsible for the nice plastic design (it corresponded to plastic body - Zenit 412)


I have a Zenitar M2S 2/50 and can confirm it is a very sharp lens, both on film and NEX-3.

The resolution figures are very high indeed, I think the only higher ones I have seen for a 50mm SLR lens were for the Contarex 2/50 Planar. (I know Orio posted some statistics on that lens, maybe he can report them here, I seem to remember centre sharpness of 80 for the Planar).


PostPosted: Wed Jul 04, 2012 9:40 pm    Post subject: Reply with quote
***I know only about Medium Format defeat 35mm digital. ***


Indeed a MF film shot 70 years ago can crush a few digital cameras Wink
http://www.shorpy.com/node/12951?size=_original#caption


PostPosted: Wed Jul 04, 2012 10:01 pm    Post subject: Reply with quote
Actually, that's a large format 5x4.

Not fair to compare a larger film to a small sensor.

Compare APS film to APS-C digital or 35mm film to full frame digital, that would be fair.


PostPosted: Wed Jul 04, 2012 10:10 pm    Post subject: Reply with quote
iangreenhalgh1 wrote:
NEX-3 is 14mp and very much adequate for testing.

Ian, in your post further up (above the cat) you state the NEX-3 is 4mp. Clearly a typo but that's where Kenji got his information. Smile


PostPosted: Wed Jul 04, 2012 10:11 pm    Post subject: Reply with quote
SonicScot wrote:
iangreenhalgh1 wrote:
NEX-3 is 14mp and very much adequate for testing.

Ian, in your post further up (above the cat) you state the NEX-3 is 4mp. Clearly a typo but that's where Kenji got his information. Smile


Whoops, my bad. I blame my fat fingers Wink


PostPosted: Wed Jul 04, 2012 10:46 pm    Post subject: Reply with quote
iangreenhalgh1 wrote:
Actually, that's a large format 5x4.

Not fair to compare a larger film to a small sensor.

Compare APS film to APS-C digital or 35mm film to full frame digital, that would be fair.


....but you missed the important bit and it was:- "Kodachrome", it was probably the only 35mm film that could challenge 35mm digital, but now gone Sad


PostPosted: Wed Jul 04, 2012 10:56 pm    Post subject: Reply with quote
Plenty of films can challenge digital, in fact, most films, even the cheap ones have a higher dynamic range than most digital sensors. There are a good number of films that are superior to most digital sensors in resolution too. Then there's colour reproduction etc etc.

Kodachrome was pretty outdated and the modern film technologies far surpass it in many regards. The best Kodak films for 35mm stills photography like Portra and Ektra are actually way behind the current 35mm motion picture emulsions in performance. Even older motion picture stocks like Kodak Vision2 have better grain structure and dynamic range than any 35mm still photography stock, the lastest ones like Vision3 and Eterna are even further ahead, remember that all the money and r and d goes into motion picture stock development and the advances made only trickle down into the stills photography emulsions several years later, if at all.


PostPosted: Wed Jul 04, 2012 11:15 pm    Post subject: Reply with quote
iangreenhalgh1 wrote:
NEX-3 is 14mp and very much adequate for testing.

5DMKII has 17% lower spatial resolution than my NEX-3, therefore my camera requires lenses of 17% higher resolution to produce the same apparent sharpness.
..


It is very interesting idea but i have to disagree. Most of pixels of dSLR's are just interpolated. I cant give you now direct link to calculation of this but i can provide link to comparizon of 35mm 22mpix and Medium Format 22mpix.

And Medium Format being same megapixel should be defeated by 35mm 22mpix according to that theory you provided (i followed your link) cause it has less density of pixes. But from this test you easily can see that result is absolutely opposite. Canon 22mpix is defeated by 22mpix of twice bigger sensor with low pixel density.

So having more pixels on smaller sensor for today is not delivering any additional details on image. But if i agree to your theory i should consider that mobile phones provide best resolution ever and if i attach professional lens to my iphone i would get much better results than i get from my 5D mark II. Does it sound strange?

The reality is quite simple. Small sensor cant resolve so much pixels. To provide more pixel density camera firmware just surround one pixel with pixels having same colors. So image in "blown". It weight more but dont have any new details.

I own Olympus E-PL1 with a 13.1 Mpix and small sensor 17.3 x 13.0 мм. I will order an adapter C/Y->micro 4/3 and will represent results little bit later.
As for "Pixel density" theory my Olympus has best sensor ever made! Cause of a very high pixel density. but test will show..Smile

Quote:
Most lenses from respected makers have had resolutions far exceeding that of film for over 60 years,


Than why my Zeiss lenses (is Zeiss respected?) are not best performers being compared to Canon lenses? Bearing in mind that you said film has more resolution
Quote:
Even the latest very dense sensors like the 36mp Sony full frame one are lower in resolution than a good film such as Fuji Velvia 50


Even cheap Canon 50/2.5 Macro defeats well-known Zeiss Makro-Planar 60/2.8 in resolution. If it is not cause of lens resolution then you have to explain this phenomenon.

P.S. Please note - i like my CZ 60/2.8! But for commercial work i use Canon 50/2.5 cause i need resolution.


PostPosted: Wed Jul 04, 2012 11:41 pm    Post subject: Reply with quote
I'm sorry, but you have misunderstood what I wrote, of course the 22mp medium format sensor has a lower pixel density, and therefore a lower spatial resolution which means to achieve the same apparent sharpness as the 22mp 35mm sensor requires a lower resolution from the lens.

The higher the pixel density, the higher the resolution lens needed to give the maximum sharpness the sensor can achieve.

Pixel density is not a theory, and you have got it completely the wrong way round, your E-PL1 has a much higher density sensor than your 5DMKII, which therefore means that it is much more demanding of the lens in resolution terms.

This means that a lens that looks sharp on your 5DMKII may not look as sharp on the E-PL1.

As I wrote before, to achieve the maximum apparent sharpness, the resolution of the lens must be equal or greater than the resolution of the sensor.

Let's take some theoretical numbers and try to make it clearer for you.

Let's say your 5DMKII has a sensor resolution of 50, and the E-PL1, being a much denser sensor has a resolution of 100. Let's take a lens with a resolution of 75, it will give images with the maximum possible apparent sharpness on your 5DMKII because 75 is greater than 50. However it will not achieve the maximum possible sharpness on your E-PL1 because 75 is less than 100.

So please, understand that what I am saying is not theory in any way, it is simple physics and maths, and the key is the pixel density of the sensor and the resulting spatial resolution.

You have interpreted what I said in exactly the opposite way.

Seriously, you need to go and study the basics of sensors before you understand the issue of sharpness, at the moment you have got it all 100% wrong.

Here are some resources that you should study in order to understand this issue:

http://luminous-landscape.com/tutorials/resolution.shtml

http://www.clarkvision.com/articles/does.pixel.size.matter/

http://en.wikipedia.org/wiki/Image_resolution

http://dpbestflow.org/camera/sensor


PostPosted: Wed Jul 04, 2012 11:43 pm    Post subject: Re: Some bokeh from my Carl Zeiss Makro-Planar 60/2.8 Reply with quote
kenji wrote:

Why you mention only saggital curve? what about tangential which goes down to 0.3?


Because the tangential curve is afflicted by colour fringing and this is (one of) the reason(s) of the decline of the curve.
Sagittal curve is not influenced by the colour fringing and therefore is more useful to assess the performance with regards to contrast/resolvance.

Quote:
And when these curves a much apart new type of aberration occurs that lowers the resolution.


Aberrations do not affect the resolvance of a lens (which is the lens' ability to record the details).
They can affect the sharpness, i.e. the perception which is determined by the way our brains combine contrast and resolvance of an image.

Quote:
And in the same time you prefer 20lp/mm than 40lp/mm that is official resolution curve (10lp/mm is for contrast).


Laughing "official"? And in what formal document would this be stated? Laughing
Curves are not "official" and they are not "made for" something.
The issue of optical perception is the subject of scientifical psychological studies and can surely not be summarized in pseudo-scholastic compendium definitions.

For instance, it is now generally accepted, by scientifical tests, that the human perception of sharpness is much more dependent
on the micro-contrast of an image (the transition between different values in image details) than on it's actual resolution.
The 40mm cycles/mm curve of a lens MTF graph does tell us much about how many fine details a lens is able to resolve. But it tells us little of how these details will appear in the actual photograph.
This because the human eye is not able to resolve successfully such minuscule detail in an image. Nature did not make our eyes efficient enough for that task, because it matters little to our living.
Instead, nature gave us an excellent ability to distinguish contrast between image elements. This because contrast tells us much more immediate information for our brain to decode.
And vision, as survival tool for animals, depends a lot on the speed of information processing.
To put it in a simple but clear example: for our survival in the jungle, it is much more important, for our vision, to tell immediately the contrast between the eye of a tiger and it's fur,
than to be able to count the number of hairs on a tiger's face.
This because the contrast is decisive to create the shape and lets us know where the head is, and if the tiger is looking at us. Which is what really matters to us.
Getting back to our MTFs: what you perceive as sharpness in an image captured by a lens, is much better described by a curve of about 20 c/mm, than by a curve of 40 c/mm.
Not because the 20 c/mm is "better" or "made for" anything special - simply because it is a value more realistic for the way the human perception works.
40 c/mm may be interesting for the records, but not much significant photographically wise.
For instance, in most photographic conditions, including home made tests, the amount of disturb generated by the film, or by the image sensor, or the digital processing, or even
by the atmospheric elements such as haze, is enough to make the 40 c/mm curve virtually meaningless even if you look at the image with a magnifying glass.

Quote:
If you look 40lp/mm you'll see not a good result especially at wide-open (0.65->0.4? 0.1?).
But look at MTF charts for 100/2.8
You see - it is much sharper. 40lp/mm goes 0.6->0.6 Almost absolute even sharpness along frame.


Again: the 40 c/mm curve indicates the resolvance at 40 c/mm (in perfect studio situation - as I said above, everyday use is a different song), not the sharpness.
The sharpness is a subjective quality of the human perception.

Quote:
Here i compared 100mm lenses from famous brands (Canon, Nikon, Zeiss, Leica)


For your information, Zeiss is the only company today which makes real MTF tests on actual production samples.
All the other companies, Canon and Nikon first, do not make real MTF tests on actual production lenses.
What they make are computer simulated MTF "tests": they put the project data of the lens (not the real lens: only the
data of the project, what the lens is supposed to be), and run computer simulation of virtual light rays passing through them.
In other words, nothing is real in those "tests". The lens is not real, it's the platonic idea of the lens.
But people who buy the lens, buy a real one, which can differ even greatly from the project design, due to materials used,
precision of build, possible defects.
And not even the light of those "tests" is real. It is simulated light. All computer games.
With those conditions, the "tests" look unavoidably much better than what the actual samples of the lens, that are bought in the shops,
do really perform.

Another information: Zeiss runs all their MTF tests with lenses set to infinity focus.
The performance of the macro lenses is however optimized to close distance focus. It is easy to verify with your 2.8/60 M-P:
if you take a photo of a landscape at infinity, you can see even at first look that the quality of the image is not as good as in the closeup shots.
This is to say that the Makro-Planar 2.8/60 used close-up as in your flower photo performs better (probably also much better) than what the Zeiss' MTF would suggest.


PostPosted: Wed Jul 04, 2012 11:55 pm    Post subject: Reply with quote
Nice post Orio, I learnt quite a bit from reading that Smile

I read that the way lens resolution was tested in the pre-computer days was to shoot a pattern of parallel lines using a very high resolution microfilm, then examine the resulting negative under a high power microscope. That, to me, being a real world test, is far more based in reality in terms of how the lens actually performs than any of the modern computer simulations you describe.


PostPosted: Thu Jul 05, 2012 1:43 am    Post subject: Re: Some bokeh from my Carl Zeiss Makro-Planar 60/2.8 Reply with quote
Orio wrote:

For your information, Zeiss is the only company today which makes real MTF tests on actual production samples.
All the other companies, Canon and Nikon first, do not make real MTF tests on actual production lenses.
What they make are computer simulated MTF "tests": they put the project data of the lens (not the real lens: only the
data of the project, what the lens is supposed to be), and run computer simulation of virtual light rays passing through them.
In other words, nothing is real in those "tests". The lens is not real, it's the platonic idea of the lens.
But people who buy the lens, buy a real one, which can differ even greatly from the project design, due to materials used,
precision of build, possible defects.
And not even the light of those "tests" is real. It is simulated light. All computer games.
With those conditions, the "tests" look unavoidably much better than what the actual samples of the lens, that are bought in the shops,
do really perform.

Another information: Zeiss runs all their MTF tests with lenses set to infinity focus.
The performance of the macro lenses is however optimized to close distance focus. It is easy to verify with your 2.8/60 M-P:
if you take a photo of a landscape at infinity, you can see even at first look that the quality of the image is not as good as in the closeup shots.
This is to say that the Makro-Planar 2.8/60 used close-up as in your flower photo performs better (probably also much better) than what the Zeiss' MTF would suggest.


It's a good information, too bad it's coming straight from Zeiss marketing brochure. As far as I understand OP has both Zeiss and Canon and his Canon is sharper. To me this is it, open and shut case.

Now the "tests" you are referring to are manufacturer's published MTF curves. Kudos to Zeiss for testing real lenses, but to me Zeiss MTF curves are no more reliable than Canon's for the simple fact that they are coming from Zeiss and Zeiss is not an impartial tester. There are various subtle ways to make the results look better than reality. For example you can test 100 lenses and publish the results from the best sample. I am sure Zeiss management is smart enough to realize that sharper lens will sell better Wink. So the only tests that can be trusted to some extent must come from independent review sites (such as photozone) and users reports.


PostPosted: Thu Jul 05, 2012 6:00 am    Post subject: Reply with quote
iangreenhalgh1 wrote:

As I wrote before, to achieve the maximum apparent sharpness, the resolution of the lens must be equal or greater than the resolution of the sensor.

Let's take some theoretical numbers and try to make it clearer for you.

Let's say your 5DMKII has a sensor resolution of 50, and the E-PL1, being a much denser sensor has a resolution of 100. Let's take a lens with a resolution of 75, it will give images with the maximum possible apparent sharpness on your 5DMKII because 75 is greater than 50.


Assuming that you are equating sharpness solely with resolution then your arguments are wrong.

You are assuming that, in an imaging system made up of two components, the lens and the camera, the overall resolution of the system is determined by the component with the lowest resolution. This is not true. It's a common myth. It's found all over the place... The camera outresolves the lens ... The lens outresolves the camera. But it's still not true.

In order to calculate the MTF of the system you need to multiply the MTFs of the two componenrts ...

MTF(lens+camera) = MTF(lens) x MTF(camera)

So if, in your example, we take a lens with a resolution of 80, instead of one with a resolution of 75, the resolution (and by implication the sharpness) will increase. The resolution and sharpness are not determined solely by the resolution of the camera.

Quote:
However it will not achieve the maximum possible sharpness on your E-PL1 because 75 is less than 100.


Here the resolution of the lens is less than that of the camera but we can increase the overall resolution by increasing the resoltion of either of them.

Quote:
So please, understand that what I am saying is not theory in any way, it is simple physics and maths ....


Unfortunately some people use a different mathematical model. If you do a search for "MTF multiplication" you will find out who they are.

Quote:
Seriously, you need to go and study the basics of sensors before you understand the issue of sharpness...


Don't we all.

Quote:
...at the moment you have got it all 100% wrong.


???


PostPosted: Thu Jul 05, 2012 10:35 am    Post subject: Re: Some bokeh from my Carl Zeiss Makro-Planar 60/2.8 Reply with quote
fermy wrote:

It's a good information, too bad it's coming straight from Zeiss marketing brochure. As far as I understand OP has both Zeiss and Canon and his Canon is sharper. To me this is it, open and shut case.


But the point was not to decide which is sharper.
The point is that he stated that the M-P 2.8/60 is not a sharp lens, and I think it's not correct to say that.
The original poster invoked the MTF, so I took the MTF and showed that a correct reading of it shows that the lens is sharp, even wide open.
Which of course does not mean that there can't be sharper lenses!
It was beyond my purpose to discuss any lens comparison, because I don't own the Canon lens, so I have no direct experience.
My comment about Canon's MTF was meant to say that it's not correct to put the two graphs (Zeiss' and Canon's) in direct comparison,
because they are two completely different things.
I have no way to prove that Zeiss is a correct company in publishing their MTFs. At the same time let me state that you have no proof in your accusation that they are biased.
I can only guess that if Zeiss tests were biased, the competition would have already raised the issue, which did not happen to my knowledge.
In any case, comparing real MTF tests with computer simulations is like comparing apples with oranges. It seems that the original poster did not know
about the origin of Canon's graphs. Now he knows, so he can better make evaluations by taking into consideration that Canon's are
computer simulations and not real world tests.


PostPosted: Thu Jul 05, 2012 12:04 pm    Post subject: Reply with quote
Yeah, I agree that it's incorrect to compare Zeiss MTF with Canon's. I am just going one step further and saying that one should not put too much stock in manufacturer's MTF at all. I am not accusing Zeiss of rigging anything, but it's a simple common sense that their impartiality in this case is questionable.


PostPosted: Fri Jul 13, 2012 10:04 pm    Post subject: Re: Some bokeh from my Carl Zeiss Makro-Planar 60/2.8 Reply with quote
Orio wrote:
kenji wrote:

Why you mention only saggital curve? what about tangential which goes down to 0.3?


Because the tangential curve is afflicted by colour fringing and this is (one of) the reason(s) of the decline of the curve.
Sagittal curve is not influenced by the colour fringing and therefore is more useful to assess the performance with regards to contrast/resolvance.

Quote:
And when these curves a much apart new type of aberration occurs that lowers the resolution.


Aberrations do not affect the resolvance of a lens (which is the lens' ability to record the details).
They can affect the sharpness, i.e. the perception which is determined by the way our brains combine contrast and resolvance of an image.

Quote:
And in the same time you prefer 20lp/mm than 40lp/mm that is official resolution curve (10lp/mm is for contrast).


Laughing "official"? And in what formal document would this be stated? Laughing
Curves are not "official" and they are not "made for" something.
The issue of optical perception is the subject of scientifical psychological studies and can surely not be summarized in pseudo-scholastic compendium definitions.

For instance, it is now generally accepted, by scientifical tests, that the human perception of sharpness is much more dependent
on the micro-contrast of an image (the transition between different values in image details) than on it's actual resolution.
The 40mm cycles/mm curve of a lens MTF graph does tell us much about how many fine details a lens is able to resolve. But it tells us little of how these details will appear in the actual photograph.
This because the human eye is not able to resolve successfully such minuscule detail in an image. Nature did not make our eyes efficient enough for that task, because it matters little to our living.
Instead, nature gave us an excellent ability to distinguish contrast between image elements. This because contrast tells us much more immediate information for our brain to decode.
And vision, as survival tool for animals, depends a lot on the speed of information processing.
To put it in a simple but clear example: for our survival in the jungle, it is much more important, for our vision, to tell immediately the contrast between the eye of a tiger and it's fur,
than to be able to count the number of hairs on a tiger's face.
This because the contrast is decisive to create the shape and lets us know where the head is, and if the tiger is looking at us. Which is what really matters to us.
Getting back to our MTFs: what you perceive as sharpness in an image captured by a lens, is much better described by a curve of about 20 c/mm, than by a curve of 40 c/mm.
Not because the 20 c/mm is "better" or "made for" anything special - simply because it is a value more realistic for the way the human perception works.
40 c/mm may be interesting for the records, but not much significant photographically wise.
For instance, in most photographic conditions, including home made tests, the amount of disturb generated by the film, or by the image sensor, or the digital processing, or even
by the atmospheric elements such as haze, is enough to make the 40 c/mm curve virtually meaningless even if you look at the image with a magnifying glass.

Quote:
If you look 40lp/mm you'll see not a good result especially at wide-open (0.65->0.4? 0.1?).
But look at MTF charts for 100/2.8
You see - it is much sharper. 40lp/mm goes 0.6->0.6 Almost absolute even sharpness along frame.


Again: the 40 c/mm curve indicates the resolvance at 40 c/mm (in perfect studio situation - as I said above, everyday use is a different song), not the sharpness.
The sharpness is a subjective quality of the human perception.

Quote:
Here i compared 100mm lenses from famous brands (Canon, Nikon, Zeiss, Leica)


For your information, Zeiss is the only company today which makes real MTF tests on actual production samples.
All the other companies, Canon and Nikon first, do not make real MTF tests on actual production lenses.
What they make are computer simulated MTF "tests": they put the project data of the lens (not the real lens: only the
data of the project, what the lens is supposed to be), and run computer simulation of virtual light rays passing through them.
In other words, nothing is real in those "tests". The lens is not real, it's the platonic idea of the lens.
But people who buy the lens, buy a real one, which can differ even greatly from the project design, due to materials used,
precision of build, possible defects.
And not even the light of those "tests" is real. It is simulated light. All computer games.
With those conditions, the "tests" look unavoidably much better than what the actual samples of the lens, that are bought in the shops,
do really perform.

Another information: Zeiss runs all their MTF tests with lenses set to infinity focus.
The performance of the macro lenses is however optimized to close distance focus. It is easy to verify with your 2.8/60 M-P:
if you take a photo of a landscape at infinity, you can see even at first look that the quality of the image is not as good as in the closeup shots.
This is to say that the Makro-Planar 2.8/60 used close-up as in your flower photo performs better (probably also much better) than what the Zeiss' MTF would suggest.


Orio, first of all i have to say big thanks to you for your detailed answer. And only after this i will comment..Smile

All curves affected by a color fringing. Even at center you can easily see color fringing. I suppose you heard about LoCa which affects center of image.



But when curves are far away one from other astigmatism aberration appears.

"If the difference between
S and M (astigmatic difference) is large, a
point will not be formed as a point and
the image will smear. Moreover, the blur
image in front of and behind the image
formation plane will be unnatural."
(c) Canon LensWork

About official document..(same documents have other companies)

"The closer the 10-line/mm
curve is to 1, the better the contrast and
separation ability of the lens, and the
closer the 30-line/mm curve is to 1, the
better the resolving power and sharpness
of the lens. Additionally, the closer the
characteristics of M and S are, the more
natural the background blur becomes."

(c) Canon LensWork

As about sharpness and resolvance. If you cant see detailes when looking at the whole image then i suppose noone stops you from scaling it on computer and crop a part you prefer more.
I know that it would be very nice to shoot a picture where your main object occupy 100% of frame. But usually you have some excess image information that you have to crop (al least i usually have to. maybe i'm the only one and all others are snipers Smile ).

Sensors that we have now easily resolve 40lp/mm. So we need lenses to be adequate. And we have to bear in mind that in future sensors will be developed much faster. So we (i talk about lens manufacturers) have to develope lenses which will correspond to sensors that are not developed yet.

As for "real tests" i know at least one more company that do them. It's Leica. I just dont have information how Canon or Nikon do their tests or just calculate synthetic tests on computer.
Btw where you got information that Canon and Nikon do only synthetic tests?

If you look at the image of a test chart i placed before you will see that right image is more sharp (at least sharp) than left. Left image was made with CZ makro-planar 100/2.8 and right was made with Canon 100/2.8L IS USM. As official MTF charts tell us - Canon is sharper (or resolve more? I dont know how to say correctly here cause i check results by my weak eyes that are only 20lp/mm -joke)

I know that test process is much more complicated than my "home-test". But as you mentioned we live in real world not synthetic one and i have to check sharpness of my lens with my photo gear. And computer helps me to check any resolvance i need by scaling.

As for Zeiss vs Canon. Here is test CZ 60/2.8 (left) vs Canon 50/2.5 (right). All on hyperfocal, camera on a tripod, sunny day = short exposure.

I had to place only small crop but i suppose you will easily see what lens is sharper. Please bear in mind that it is 50mm vs 60mm so CZ have a handicap.



I dont remember if i have such images for a close-up but i remeber that Canon was also better.

Btw Zeiss test their lenses at infinity.

P.S. Btw Canon 50/2.5 is an old lens wo any flyourite or UD glass inside.

How would you explain such a phenomenon?


PostPosted: Fri Jul 13, 2012 10:07 pm    Post subject: Reply with quote
Hmm I suspect your focusing is out for the Zeiss shot. I have never seen a Zeiss prime shoot that badly at infinity. Could you use liveview to focus instead?


PostPosted: Fri Jul 13, 2012 10:25 pm    Post subject: Re: Some bokeh from my Carl Zeiss Makro-Planar 60/2.8 Reply with quote
Orio wrote:
fermy wrote:

It's a good information, too bad it's coming straight from Zeiss marketing brochure. As far as I understand OP has both Zeiss and Canon and his Canon is sharper. To me this is it, open and shut case.


But the point was not to decide which is sharper.
The point is that he stated that the M-P 2.8/60 is not a sharp lens, and I think it's not correct to say that.
The original poster invoked the MTF, so I took the MTF and showed that a correct reading of it shows that the lens is sharp, even wide open.
Which of course does not mean that there can't be sharper lenses!
It was beyond my purpose to discuss any lens comparison, because I don't own the Canon lens, so I have no direct experience.
My comment about Canon's MTF was meant to say that it's not correct to put the two graphs (Zeiss' and Canon's) in direct comparison,
because they are two completely different things.
I have no way to prove that Zeiss is a correct company in publishing their MTFs. At the same time let me state that you have no proof in your accusation that they are biased.
I can only guess that if Zeiss tests were biased, the competition would have already raised the issue, which did not happen to my knowledge.
In any case, comparing real MTF tests with computer simulations is like comparing apples with oranges. It seems that the original poster did not know
about the origin of Canon's graphs. Now he knows, so he can better make evaluations by taking into consideration that Canon's are
computer simulations and not real world tests.


Omg.. Orio, where you got information that Canon MTF are only computer simulations? I've seen a video of Canon test and they used projector gear for this test. It's a known advertising clip btw.

If it's not correct to put MTF graphs together than it's not correct also to say that CZ 60/2.8 is sharp. We have different cameras and different eyes and all of them (cameras and eyes) have their own MTF. So maybe i have my eyes MTF like an eagle and the best camera ever and you (just for example) have weak eyes and old 4mpix camera. So i will say that CZ60 is a weak lens and you will say it's very sharp. Where is the truth?

Or all that we are going to say will be wrong?

My idea was to compare these charts assuming that they are real lens tests. Just take it as axiom.
To prove this axiom i compared sharpness of these lenses and checked if it correspond to the MTF.

This method is not 100% correct i know. Maybe it's only 60-70-80% correct. But from now i can say that Canon make graphs that approximately correspond to real life.

I dont say without any proof that Canon or Nikon or smbdy produce fake tests. I try to prove them for myself and others who will be interested.

If you have better idea than to place camera on tripod and make some shoots with both lenses - you are welcome. I will be highly interested if i will be able to realize it in my studio.


PostPosted: Fri Jul 13, 2012 10:26 pm    Post subject: Reply with quote
kenji wrote:
Here is test CZ 60/2.8 (left) vs Canon 50/2.5 (right)

I had a long time to see a so funny test, I will need half an hour to stop laughing Laughing


PostPosted: Fri Jul 13, 2012 10:36 pm    Post subject: Reply with quote
ManualFocus-G wrote:
Hmm I suspect your focusing is out for the Zeiss shot. I have never seen a Zeiss prime shoot that badly at infinity. Could you use liveview to focus instead?


Lens was focused to infinity (hyperfocal) so no focus needed. All buildings had to be sharp as they can be in ideal focusing. But please note this shot made at F4.

If it was not test i will set F8 or F11 for the best results.

Here at F5.6 difference is smaller but still noticable (Zeiss in the left)



PostPosted: Fri Jul 13, 2012 10:40 pm    Post subject: Reply with quote
poilu wrote:
kenji wrote:
Here is test CZ 60/2.8 (left) vs Canon 50/2.5 (right)

I had a long time to see a so funny test, I will need half an hour to stop laughing Laughing


Keep laughing - i heard it will extend your life Smile

Then please try to post some useful info.


PostPosted: Fri Jul 13, 2012 10:44 pm    Post subject: Reply with quote
Just because the lens indicated infinity, it doesn't mean you actually have infinity Wink You may have passed it. Even at f/4, the shot looks way too soft in my opinion.
Post new topic   Reply to topic
Home   Manual Focus Lenses
Goto page Previous  1, 2, 3  Next
Page 2 of 3