wayne seltzer wrote:
I guess nobody has comments about the low-light tendering differences.
Well, the scenes are different so it is hard to tell. Further I am not sure what you are saying about the dynamic range. If you are saying that it is a desirable property of a lens to scatter light from highlights into the shadow zones, I disagree.
No, I am talking about different global contrast levels of lenses so that shadows are rendered darker than you may want and hightlights brighter than you want too. Is there such a thing as sensitivity to light or amount of light transferred by the lens to the sensor, t-stop??
The transfer characteristics of glass is AFAIK linear all the way to where the internal losses would cause heat buildup and permanent structural change in the material ( AKA melting).
All differences in global contrast is due to light scattering.
The last sentence on that page says it all: What a mess.
Anyway, if I understand your explanation correctly you are not using the same definition as Zeiss so comparisons between your results and Zeiss data sheets are not meaningful. The weakness of your approach is that your reference value (the initial radius) is already distorted. You compare a corner point, displaced by distortion, with a border point, also displaced by distortion. Thus you find smaller distortion values than Zeiss, who use the image point in the absence of distortion as the reference value (i.e. according to this figure from the link you provided).
I know that there is a deviation at lower but not is my intention to give a date approved match with any rule, to begin I use much longer distances than the normal in other tests. Doing this test with a chart of 120 cms. x 80 cms. marked with a pattern I would make it easier to calculate the exact value but at that time I´ll have the problem of focusing distance.
When I wanted to do these tests (2,005) chose the way of informing a mix of graphic and numeric but always giving more emphasis to the graph, that is why I always put thumbnails of all. I can not find representative calculate distortions and sharpness at distances of little use, unusual or where the lens does not develop well (MTF).
I recalculated the value of the approach to 25 cms. reference to the projected size and the accumulated value is -2,07%. The value of the zone corner is -6.16%. I put the original file downloadable from the link in the same picture.
About the sections of lab. I give more value to the sequence of charts and curiosly is the least that look the readers ;-)
Wayne, I would indeed comment, as I think your tests are most interesting, but I am traveling, and viewing the site on an iPad, not the best tool to form an educated opinion. This week-end, I should be home, and will conduct tests of my own, which could complement yours IMHO, such as ZE 25 Vs ZE 21 VS ZE 35.
Philippe, look forward to seeing your vacation shots as well as your testing with new ZE 25/2!
I am interested in how the 25/2 color rendering and low CA compare with the new ZE 35/1.4.
Also how it compares sharpness wise with 35/1.4.
I just read in PopPhoto review that their DXO testing found the 25/2 to not reach the Outstanding A+ level for 11x14 relative print size at any aperture, whereas it did for both ZE 35/1.4 and for 24Lmk2.
Since photozone hasn't posted ZE35/1.4 test results but has 25/2 results I am not sure if PopPhoto results make sense.
Alundeb and Toothwalker, light scattering sounds bad. Not sure what you mean by that. Some lenses are more contrasty than others. Olympus lenses seem to always be less contrasty than zeiss lenses. Does that mean the Olympus lenses scatter light more than Zeiss lenses??
To me the Zeiss 25/2 compared to my 24G is just like turning up the contrast slider in LR. If I try to expose the dimly lit scene to not blowout the highlights near light sources or objects illuminated directly by the direct lighting, then the dark areas of the scene become much darker than what the 24G renders and even if I add quite a bit of Fill in LR the shadows are still murky. The 24G renders the various light levels in the scene more like what I saw, where as the 25/2 seems to brighten the highlights more and darken the shadows more.
wayne seltzer wrote:
Alundeb and Toothwalker, light scattering sounds bad. Not sure what you mean by that. Some lenses are more contrasty than others. Olympus lenses seem to always be less contrasty than zeiss lenses. Does that mean the Olympus lenses scatter light more than Zeiss lenses??
Yes.
A reduction of microcontrast is due to lens aberrations (related to refraction).
A reduction of global contrast is due to reflection and scattering.
Toothwalker wrote:
For the Y/C 2.8/21 there is a third version, in the book "Only Zeiss".
I tried to buy that book last year, but couldn't find it at a price I was willing to pay.
Toothwalker wrote:
The Zeiss focus criterion is to optimize the MTF at 20 lp/mm, in the image center at full aperture.
There is no refocus for measurements at reduced apertures.
Good information.
wayne seltzer wrote:
No, I am talking about different global contrast levels of lenses so that shadows are rendered darker than you may want and hightlights brighter than you want too.
And:
wayne seltzer wrote:
The 24G renders the various light levels in the scene more like what I saw, where as the 25/2 seems to brighten the highlights more and darken the shadows more.
This suggests the Zeiss has greater flare resistance, which wouldn't surprise me in the least. While better flare resistance can in theory make the highlights brighter, I doubt the tiny difference would be visible there; rather, any differences would be visible in the shadows, where even a small amount of non-image-forming light can dramatically brighten things.
My own Zeiss lenses have noticeably better flare resistance, in general, than my other lenses. I think this difference is caused by coatings more than anything else, though possibly some types of glass/polymer used for lens elements have non-negligible internal scattering. Maybe Zeiss pays more attention to the effect of element shapes on ghosting.
Veiling flare and ghosting can add mystery, drama, or straightforward visual interest to a photo, so they're not always bad. But flare also reduces colour saturation and colour accuracy, since it pollutes strong colours like blues skies with non-imaging-forming light (which is often dirty white).
S, do you think Wayne's Nikkor so relatively inferior in those characteristics as to be responsible for the differences in the images he posted, and the luminosity handling he notes? It's a modern, high end Nikkor we are talking about...
TW, your comment on p11:
'For the Y/C 2.8/21 there is a third version, in the book "Only Zeiss".'
Your comment also leads the reader to the fact that there are three versions of the CY 21/2.8. Assuming you are mis-speaking here, and refer to only one other CY version, the question arises:
Was this some obscure limited release version you refer to, perhaps a prototype, from the Zeiss monograph? I have seen no mention of such a lens version anywhere else. I ask because I do not favour obscurantism ;~)
Nasse makes no such distinction in the Distagon white paper. He is clearly referring to the familiar marketed version of the lens:
'..thus it is no wonder that its price on the pre-owned market exceeded the original price after it was no longer produced.'
'For the Y/C 2.8/21 there is a third version, in the book "Only Zeiss".'
Your comment also leads the reader to the fact that there are three versions of the CY 21/2.8. Assuming you are mis-speaking here, and refer to only one other CY version, the question arises:
Was this some obscure limited release version you refer to, perhaps a prototype, from the Zeiss monograph? I have seen no mention of such a lens version anywhere else. I ask because I do not favour obscurantism ;~)
Nasse makes no such distinction in the Distagon white paper. He is clearly referring to the familiar marketed version of the lens:
'..thus it is no wonder that its price on the pre-owned market exceeded the original price after it was no longer produced.' ...Show more →
As far as I know there is only one version of the Y/C 2.8/21. I could have been more clear; I meant that there are (at least) three different MTF charts for that lens.
From Lloyd Chambers recent blog entries:
"Zeiss 25mm f/2 Distagon — perhaps the best 24/25mm lens available, but optimized for central 3/4 of frame, and thus still not strong enough at the edges and corners, interacting with mild field curvature. Central 3/4 of the frame should perform really well on 36MP sensor. Outstanding color correction might nonetheless show a trace of weakness on 36MP, but it’s the best of the best 24/25mm nonetheless. "
" Nikon 24mm ƒ/1.4G — assuming correction of its moderately strong color fringing, sharpness over the central 3/4 of the frame should be excellent, and stopping down should deliver very good results over the whole frame."
" Zeiss 21mm f/2.8 Distagon —should carry forward nicely, but will likely show some limits that are not apparent at 21 or 24MP, including some traces of chromatic errors."
alundeb wrote:
Thank you. Your article on MTF measurements implicitly say so, but it is not as easy to grasp how dispersion varies with magnification while not with aperture. Do you know how Zeiss measure CA at infinity? That facility is not mentioned in the datasheet for the K8 for instance.
It should not be too difficult to collect some more samples at very little magnification (I hope you agree that maginfication in the order of 0.002 to 0.001 can be regarded as representative of infinity). The ones from Wayne and Thorsten are just single samples and not verified to be representative of the whole frame, as I understand that assymetry may be present. ...Show more →
(removed nested quotes)
CA at infinity can be measured by switching the slit illumination filter on the K8, with the slit in the meridonal axis. You can either manually move one of the guide measurement lines in the microscope, or save the image data and do the calculations later.
In a normal case, you'd have a D50 (well, if you're serious anyway) spectrum light shining through the slit, and by switching out the recording module to an RGB module with defined RGB primaries you can do a three-point CA measurement "in one go" on white light too. This module lowers the absolute resolution of the recording though, and does not really fit the CA measurement very well (you want to know the magenta wavelength combination displacement, not the warm red/deep blue combination that most RGB modules use). I much prefer the filter-switch method, since you can then decide on spectrum endpoints yourself, and hence choose how much purple fringe you want to include in your measurements and so on...
Unfortunately there's not space enough to fit a filter holder on the recording / microscope end of things with some solutions, otherwise this would have been very easy. Certainly easier to do than to change the illumination source filter.
OK, I now have picked up my ZE 25. Very early pics indicate that Wayne does indeed know what he is doing. There is definite far corner softness, even at f:4.0. It affects roughly one half of the LCD screen when viewed on camera with maximum magnification. Meaning, really not a lot for some of us, and much too much for others.
There is also a strange phenomenon, that looks like contrast amplification. It is not huge, but it is there.
If I have time, I will proceed with comparing ZE 21, 25 and 35 f:1.4. That should shed some light on what I am seeing.
philber wrote:
OK, I now have picked up my ZE 25. Very early pics indicate that Wayne does indeed know what he is doing. There is definite far corner softness, even at f:4.0. It affects roughly one half of the LCD screen when viewed on camera with maximum magnification. Meaning, really not a lot for some of us, and much too much for others.
There is also a strange phenomenon, that looks like contrast amplification. It is not huge, but it is there.
If I have time, I will proceed with comparing ZE 21, 25 and 35 f:1.4. That should shed some light on what I am seeing....Show more →
Thanks philber. I don't think anybody is disputing corner softness at f/4. The main question here appears to be whether the corners are acceptable at f/8. There we have i) different opinions on what is acceptable and ii) different test results posted by the various testers.
I went out last week along the Thames embankment with the Zeiss 25mm F2 and Canon 24mm TSE II to try and do some comparative tests between the two lenses. A few observations:
a) I found it almost impossible to frame a shot 'normally' and have detail in the corners that I thought should be 'sharp'. In all cases I found that when I framed to make a correctly composed shot it would have corner information that was much closer to the camera and DOF would have a huge influence on the apparent sharpness of the corners. To get detail in the corners I had to artificially tilt up until e.g. the far bank of the Thames was placed at the bottom of the frame. To my mind it becomes very difficult to have elements in a plane at 'infinity' (at least 400+ft).
b) Canon TSE is sharper in the far corners than the Zeiss. The Zeiss has a zone of 3mm in the corners that are affected by a slight softness. This doesn't look as extreme as Waynes but my version doesn't look as good as Jorges either. Something to do with the difference between the sensor sizes they were shot on (I'm using a 1DSIII)? Something to do with the composition (Wayne's look as though DOF might be a problem). Flipping back and forth between the two in the corners you have to stop down the Zeiss to F8 to match the Canon at F4. The edges of the Zeiss look best at F11 but by then the centre sharpness is slightly affected by diffraction. Outside the 3mm soft zone the Zeiss seems microscopically sharper.
c) The Canon has less CA (none to be honest) while the Zeiss has slight CA.
d) The Zeiss has the rather odd effect that the bokeh at F2 looks exactly like the bokeh at F2.8 so you can shoot at F2.8 to get the same 'blur' but with much better control of vignetting.
Looking at the two I would say that they are created for very different markets. If you were Bernd and Hilla Becher looking for a lens that gives you the truth and nothing but then I would choose the Canon, it is very clinical in its reproduction and technically it minimizes all the usual problems with a lens of this type (negligible distortion, CA, vignetting).
However if I was looking for a lens that is more poetic and lyrical in its reproduction of a scene then the Zeiss is the one to choose. For the type of pictures I produce when I am out and about (and not shooting for work) I always found the Canon a bit boring, the Zeiss on the other hand is much more in tune with how I want an image presented.
thanks Phillipe and Simon for volunteering more testing. Looking forward to comparisons with the other "best of Zeiss" - FLs and many more shots, especially those lyrical ones
simonw wrote:
I went out last week along the Thames embankment with the Zeiss 25mm F2 and Canon 24mm TSE II to try and do some comparative tests between the two lenses. A few observations:
a) I found it almost impossible to frame a shot 'normally' and have detail in the corners that I thought should be 'sharp'. In all cases I found that when I framed to make a correctly composed shot it would have corner information that was much closer to the camera and DOF would have a huge influence on the apparent sharpness of the corners. To get detail in the corners I had to artificially tilt up until e.g. the far bank of the Thames was placed at the bottom of the frame. To my mind it becomes very difficult to have elements in a plane at 'infinity' (at least 400+ft).
b) Canon TSE is sharper in the far corners than the Zeiss. The Zeiss has a zone of 3mm in the corners that are affected by a slight softness. This doesn't look as extreme as Waynes but my version doesn't look as good as Jorges either. Something to do with the difference between the sensor sizes they were shot on (I'm using a 1DSIII)? Something to do with the composition (Wayne's look as though DOF might be a problem). Flipping back and forth between the two in the corners you have to stop down the Zeiss to F8 to match the Canon at F4. The edges of the Zeiss look best at F11 but by then the centre sharpness is slightly affected by diffraction. Outside the 3mm soft zone the Zeiss seems microscopically sharper.
c) The Canon has less CA (none to be honest) while the Zeiss has slight CA.
d) The Zeiss has the rather odd effect that the bokeh at F2 looks exactly like the bokeh at F2.8 so you can shoot at F2.8 to get the same 'blur' but with much better control of vignetting.
Looking at the two I would say that they are created for very different markets. If you were Bernd and Hilla Becher looking for a lens that gives you the truth and nothing but then I would choose the Canon, it is very clinical in its reproduction and technically it minimizes all the usual problems with a lens of this type (negligible distortion, CA, vignetting).
However if I was looking for a lens that is more poetic and lyrical in its reproduction of a scene then the Zeiss is the one to choose. For the type of pictures I produce when I am out and about (and not shooting for work) I always found the Canon a bit boring, the Zeiss on the other hand is much more in tune with how I want an image presented.
+1 - lyrical is a good description, like the 28/2 but better 3D
simonw wrote:
Looking at the two I would say that they are created for very different markets. If you were Bernd and Hilla Becher looking for a lens that gives you the truth and nothing but then I would choose the Canon, it is very clinical in its reproduction and technically it minimizes all the usual problems with a lens of this type (negligible distortion, CA, vignetting).
However if I was looking for a lens that is more poetic and lyrical in its reproduction of a scene then the Zeiss is the one to choose. For the type of pictures I produce when I am out and about (and not shooting for work) I always found the Canon a bit boring, the Zeiss on the other hand is much more in tune with how I want an image presented.
That's a very good analogy. Some might argue for a more clinically perfect lens and add the poetry in development/ processing because with the "poetic and lyrical" lens, it's impossible to get the clinical look when needed but with the clinical lens, one might achieve the second in PP. But it really comes down to personal preference and shooting style in the end.
Tariq Gibran wrote:
That's a very good analogy. Some might argue for a more clinically perfect lens and add the poetry in development/ processing because with the "poetic and lyrical" lens, it's impossible to get the clinical look when needed but with the clinical lens, one might achieve the second in PP. But it really comes down to personal preference and shooting style in the end.
Canon, I think, are very good at designing these kind of 'technical' lenses at longer focal lengths (such as the 85mm F1.2 and 135mm F2) so it doesn't suprise me that their new wideangle TSE's follow suit.
However despite this I still find myself drawn to lenses like the ZE 85mm F1.4 that are 'characterful' (soft glow wide open, biting sharp at F8) because they allow some artistic interpretation in how they are used.
I think that the Zeiss 25mm F2, while not as extreme as some of the other Zeiss offerings, has an interesting visual presentation.
