Home
SearchSearch MemberlistMemberlist RegisterRegister ProfileProfile Log in to check your private messagesLog in to check your private messages Log inLog in

Fast 135mm lenses (and two zooms at f=135mm)
View previous topic :: View next topic  


PostPosted: Tue Jul 21, 2020 12:56 pm    Post subject: Reply with quote

I believe that Pentax added an additional optical element in the second version of SMC Takumar F2.5 just to improve performance in the corners. So it is very strange that this lens performed so badly in the test. Possibly the poor performance of Stephan's SMC Takumar was due to a tilted adapter.

The photos below show the performance of my copy of SMC Takumar F2.5 for wide open and F8. It seems to me that the wide open performance is respectable and only slightly lower than for F8.

The images were captured with a 24 MP Sony A99 camera. The image samples are 100% crops.




PostPosted: Tue Jul 21, 2020 11:57 pm    Post subject: Reply with quote

memetph wrote:
iangreenhalgh1 wrote:
I really don't see the point in looking at the extreme corners, especially with a 135mm lens. Unless they are really bad, no-one will ever notice. A centre crop would be much more informative.

There are no bad 135mm.....Unless you watch the corners and check the CA.


Exactly. That is why I don't see the point in looking at corners. No-one has ever walked up to one of my large prints and looked at the corners, not once. One print I sold years ago was taken with a Minolta MD 2/45 that was absolutely full of fungus, I was just trying it to see if it worked before cleaning it. So the minutiae of performance is irrelevant really.


PostPosted: Wed Jul 22, 2020 1:52 am    Post subject: Reply with quote

To look at corner performance focus the corner not the center...

To look at rendering of fast outer portions of glass farthest from the center, place a mask in the center of the front element.


PostPosted: Wed Jul 22, 2020 2:15 pm    Post subject: Reply with quote

Gerald wrote:
I believe that Pentax added an additional optical element in the second version of SMC Takumar F2.5 just to improve performance in the corners. So it is very strange that this lens performed so badly in the test.


There's a simple explanation. The S-M-C Takumar 2.5/135mm (2nd computation, 6 lenses) has much more field curvature than the other 135mm lenses tested here. Since I'm using my lenses primarily for landscape work, I'm testing them by focusing the center of the image.

Of course i can focus the corners: Now the corners are quite sharp, but large parts of the image including the center are now blurred. Not what I want ... and not a good testing method if you want to shoot landscape images.






Gerald wrote:

Possibly the poor performance of Stephan's SMC Takumar was due to a tilted adapter.

No. When I'm focusing to the center, all four corners of the Tak 2.5/135mm (at f2.5) are blurred.

Stephan


PostPosted: Wed Jul 22, 2020 2:19 pm    Post subject: Reply with quote

Happy Dog Like 1 Thank you!

Thank you Stephan!


PostPosted: Wed Jul 22, 2020 2:24 pm    Post subject: Reply with quote

visualopsins wrote:
To look at corner performance focus the corner not the center...


No.

I'm testing the lenses for real world use; if I focus the corners and have a blurred center of the image, the lens is of very limited value for my landscape & architecture photography.

Of course, shooting reportage is another thing. And some lenses (such as the Novoflex T-Noflexar 5.6/400mm) are specifically designed with this purpose in mind: The German "Schnellschuss-Objektiv" was made mainly for animal photography; field curvature is irrelevant there, but the excellent CA correction of the newest T-Noflexar 5.6/400mm is a huge plus!

S


PostPosted: Wed Jul 22, 2020 7:32 pm    Post subject: Reply with quote

stevemark wrote:
visualopsins wrote:
To look at corner performance focus the corner not the center...


No.

I'm testing the lenses for real world use; if I focus the corners and have a blurred center of the image, the lens is of very limited value for my landscape & architecture photography.

Of course, shooting reportage is another thing. And some lenses (such as the Novoflex T-Noflexar 5.6/400mm) are specifically designed with this purpose in mind: The German "Schnellschuss-Objektiv" was made mainly for animal photography; field curvature is irrelevant there, but the excellent CA correction of the newest T-Noflexar 5.6/400mm is a huge plus!

S


Real world use is wide open? Just kidding. Sorry I missed the purpose of the comparisons before.

So we're down to curvature of field differences. How to measure curvature of field? How far away to focus before oof corners go away?


PostPosted: Wed Jul 22, 2020 10:36 pm    Post subject: Reply with quote

visualopsins wrote:

Real world use is wide open? Just kidding. Sorry I missed the purpose of the comparisons before.


Yes, wide open ... for landscapes! I've explained it before, but most of my best landscape / townscape / architecture photos are taken in twilight, sometimes with tripod, sometimes without. I'll give you two examples:

1) Without tripod: Rome, after ten days of waiting and preparing finally the perfectly clear air and beautiful weather come together. I've taken photos all day long, but i know the time before and after sunset will be magnificent. I know quite a few "spots" in town where I'd like to take images during these two hours. That means a lot of running, from one location to the next one, and sometimes using my Zeiss 2.8/16-35mm at f2.8. The Sony 2.8/70-200 G SSM most of the time is at f2.8: Wide open.

2) With tripod: Firenze, a most beautiful spot on top of a hill close to the town center (no, not Piazzale Michelangelo ...). Forty minutes after sunset, clear air and a breathtaking light. I'm using the Sony 2.8/300mm G SSM to shoot the cupola of the Duomo. The light is changing rapidly, at f2.8 or f4 and ISO 100 I get something between 1 s and maybe 5 s. At f11 this would result in exposure times of 15 s to 2 min. Not good for the sensor (hot pixels), and not good for quick adjustments of my camera.

Yes, i do use wide open for landscapes. Not always of course, but often in tricky situations when the light is most beautiful.

S


EDIT: I've even been shooting landscapes wide open with the Zeiss APO Distagon 1.4/28mm (to be precise: townscapes ...). It's quite incredible what you can do without tripod using these lenses. They are, however, too heavy for me - the Otus mentioned before is nearly as heavy as my Canon FD 2.8/300mm Fluorite (1.4kg vs 1.9kg).
S


PostPosted: Fri Jul 24, 2020 2:41 pm    Post subject: Reply with quote

Using these lenses wide open is simply bad technique and totally unnecessary with modern cameras that produce clean files even at high ISO. It makes no sense to use a lens at it's weakest, much better to play to it's strengths i.e. stopping it down, to say nothing of the dof issues.


PostPosted: Sat Jul 25, 2020 4:58 am    Post subject: Reply with quote

Here's a little comparison I did earlier. Although the SMC Takumar 135/2.5 (II) shows quite some CA, it was pretty sharp in the corners wide open. Mind you, this is a 42+ mp camera.

BTW 200% is a typo, should be 100%.


PostPosted: Sat Jul 25, 2020 8:45 am    Post subject: Reply with quote

There is something seriously wrong with the Yashinon I'd say. Takumar looks best and Mamiya has more corner problems.


PostPosted: Sat Jul 25, 2020 9:42 am    Post subject: Reply with quote

D1N0 wrote:
There is something seriously wrong with the Yashinon I'd say. Takumar looks best and Mamiya has more corner problems.


Yeah, the Yashinon is either a really bad lens, or there is something wrong.
This was not a very reliable test though, because the subject is not completely at infinity and I think the corners and center are not at the same focal plane. The 'issue' could be with either the Takumar only or also with the Mamiya, yet to be determined. But it shows the Takumar has good resolution in the corners.


PostPosted: Sat Jul 25, 2020 12:43 pm    Post subject: Reply with quote

caspert79 wrote:

This was not a very reliable test though, because the subject is not completely at infinity and I think the corners and center are not at the same focal plane.


That's a bit misleading ... some lenses (probably the Takumar with its field curvature) might profit from this setup, others (with no field curvature, or with a field curvature in opposite direction to the Takumar) might look worse than with a neutral test (all parts of the image at infinity).

Stephan


PostPosted: Sat Jul 25, 2020 12:46 pm    Post subject: Reply with quote

so a bit of field curvature can be useful when the corners are a bit closer than the center :p


PostPosted: Sat Jul 25, 2020 1:10 pm    Post subject: Reply with quote

D1N0 wrote:
so a bit of field curvature can be useful when the corners are a bit closer than the center :p


Yes, of course ... the Contax 2.8/25mm is said to be computed deliberately with some field curvature. This gave some degree of freedom to reduce other aberrations which were more disturbing in practical life. However, I'm not sure whether this is true.

The Novoflex 400mm and 600mm lenses for sure have lots of field curvature - but it's irrelevant when shooting animals ...

S


PostPosted: Sat Jul 25, 2020 1:42 pm    Post subject: Reply with quote

stevemark wrote:
caspert79 wrote:

This was not a very reliable test though, because the subject is not completely at infinity and I think the corners and center are not at the same focal plane.


That's a bit misleading ... some lenses (probably the Takumar with its field curvature) might profit from this setup, others (with no field curvature, or with a field curvature in opposite direction to the Takumar) might look worse than with a neutral test (all parts of the image at infinity).

Stephan


Yes, in a way it is misleading. It depends on your purpose if lack of curvature is better.


PostPosted: Sat Jul 25, 2020 8:01 pm    Post subject: Reply with quote

stevemark wrote:
Gerald wrote:

Possibly the poor performance of Stephan's SMC Takumar was due to a tilted adapter.

No. When I'm focusing to the center, all four corners of the Tak 2.5/135mm (at f2.5) are blurred.

Stephan


Well, if all four corners are blurred, this doesn't mean you don't have a tilted adapter. A tilted adapter can easily produce four corners blurred. Maybe you have a situation where field curvature is combined with tilted adapter and/or lens decentering.



stevemark wrote:
Gerald wrote:
I believe that Pentax added an additional optical element in the second version of SMC Takumar F2.5 just to improve performance in the corners. So it is very strange that this lens performed so badly in the test.


There's a simple explanation. The S-M-C Takumar 2.5/135mm (2nd computation, 6 lenses) has much more field curvature than the other 135mm lenses tested here.
Stephan


You are right when you say that the SMC Takumar 135mm F2.5 suffers from some field curvature. Indeed, I measured the field curvature of my Takumar and found about -0.13mm.

A negative field curvature means that the Petzval surface is convex toward the lens, what is typical of telephoto lenses:


A field curvature of -0.13mm for a 135mm lens also means that when the center is focused to infinity, the corners are focused to about 140 m.



In my opinion, the main point in this discussion is: how importante is in practice a field curvature of about 0.1mm (100 μm)?

It is important to note that almost all photographic lenses, even the most prestigious ones, suffer from field curvature by that amount. See, for example, the field curvature of the Zeiss Otus 55mm F1.4, as measured by Roger Cicala:


https://www.lensrentals.com/blog/2016/09/fun-with-field-of-focus-part-1/

NOTE: The vertical scale is wrong. Divide the numbers by two to get the correct field curvature.



Another example, a bunch of modern 50mm lenses:

https://www.lensrentals.com/blog/2020/01/practical-use-of-field-curvature-wide-and-telephoto-primes/

NOTE: The vertical scales were corrected.


The effect of field curvature on the image quality depends on the relative aperture (F-number). Theoretically, when there is some field curvature, the circle of confusion is given by (field curvature)/(F-number). So, for a F2.5 lens a field curvature of, say, 0.1mm will give a circle of confusion of 40 μm, which corresponds to resolution of only 25 lines per millimeter.

When field curvature is a concern, the photographer should focus to an in-between distance, so the the circle of confusion is divided by two. So, for a curvature field equal to 0.1mm, the circle of confusion for a F2.5 lens would be now 20 μm, with a corresponding resolution of 50 lpm, a very respectable value. Of course, if the lens is stopped down to F4 or F5.6 the resolution increases to outstanding 70 lpm or 100 lpm, respectively.

In conclusion, a field curvature of 0.1 mm, or so, shouldn't be a problem for a photographer when using a 135mm F2.5 or F2.8 lens wide open. Most probably, the photographer would be in greater trouble if he wanted to shoot landscapes with a Zeiss Milvus 50mm F2 wide open!


PostPosted: Sat Jul 25, 2020 8:53 pm    Post subject: Reply with quote

Interesting post Gerald.
Would a macro lens give a flat Petzval surface, or not at infinity?


PostPosted: Sat Jul 25, 2020 9:17 pm    Post subject: Reply with quote

Gerald wrote:
stevemark wrote:
Gerald wrote:

Possibly the poor performance of Stephan's SMC Takumar was due to a tilted adapter.

No. When I'm focusing to the center, all four corners of the Tak 2.5/135mm (at f2.5) are blurred.

Stephan


Well, if all four corners are blurred, this doesn't mean you don't have a tilted adapter. A tilted adapter can easily produce four corners blurred. Maybe you have a situation where field curvature is combined with tilted adapter and/or lens decentering.

If the adapter was tilted, it would result in some corners having frontfocus and others having backfocus. I will test that tomorrow.
As with any real world lens, de-centering may be an issue, especially since the Tak is about 50 years old, and quality control was not as good around 1970 as it was ten years later.



Gerald wrote:

You are right when you say that the SMC Takumar 135mm F2.5 suffers from some field curvature. Indeed, I measured the field curvature of my Takumar and found about -0.13mm.


How did you measure field curvature? It would be good to check the actual amount of field curvature also on my lens ... !

Gerald wrote:

In conclusion, a field curvature of 0.1 mm, or so, shouldn't be a problem for a photographer when using a 135mm F2.5 or F2.8 lens wide open. ...


All very well written ... BUT if I have to choose between these two 135mm lenses (both 100% corner crops, images taken wide open) it's VERY clear which on I chose!!



One is the Zeiss CY 2.8/135mm, the other one is the S-M-C Takumar 2.5/135mm (II).

S


PostPosted: Sat Jul 25, 2020 9:40 pm    Post subject: Reply with quote

Thanks Gerald for taking the time to post the info from lensrentals! This one also https://www.telescope-optics.net/curvature.htm


PostPosted: Sun Jul 26, 2020 6:23 pm    Post subject: Reply with quote

stevemark wrote:

How did you measure field curvature? It would be good to check the actual amount of field curvature also on my lens ... !



Field curvature is usually measured in terms of radius of curvature of the Petzvall surface, but we can also define field curvature as the distance a lens has to move to focus to an off-axis object which was previously on the optical axis of the lens.

Field curvature in is measured with the aid of a micrometer, but in practice you can use the lens focusing helicoid and the respective distance marking as a substitute for the micrometer. Nonetheless, it is important that the focusing system isn't plagued with excessive play, otherwise the results could be confusing.

The procedure I used to measure the field curvature of the SMC Takumar 135mm F2.5 was the following:

1. Focus to a distant object and measure the physical distance along the focusing scale that the focusing ring must be rotated to focus the object on and off axis. Let's call this distance X. In general X is very small, I found X as about 1mm for my Takumar copy.

2. Measure the distance, along the focus scale, between ∞ and the next marking(s). Call this distance Y. I measure Y as about 4mm for the 35m marking (You could also use the markings for 15m or 10m, which should give the same results).

3. Use the following formula to determine the lens travel along the axis when the lens is focused from ∞ to the auxiliary marking:

1/f = 1/d1 + 1/d2

where:

f = focal length
d1 = distance from object to optical center of the lens
d2 = distance from image to optical center of the lens

In my case, f = 135 mm d1 = 35,000 mm, so d2 = 135.52 mm

4. Subtract f from d2 and call the difference Z. Note that Z = 0.52mm is the distance along the optical axis that the lens should move
to focus an object at the auxiliary distance 35m.

5. Finally, use the following formula to calculate the field curvature:

field curvature = XZ/Y

or,

field curvature = 1mm x 0.52mm / 4mm = 0.13mm

NOTE: For auxiliary distances of 15m and 10m, I found Y = 10mm and 16mm, with the resulting calculated field curvature of 0.12mm. The slight difference in relation to the auxiliary distance of 35m is probably due to the imprecision of the markings.






stevemark wrote:


All very well written ... BUT if I have to choose between these two 135mm lenses (both 100% corner crops, images taken wide open) it's VERY clear which on I chose!!



One is the Zeiss CY 2.8/135mm, the other one is the S-M-C Takumar 2.5/135mm (II).

S


I agree with you. If I could choose, I would also use the Zeiss C/Y Sonnar, which is really an excellent lens. You are a lucky man for having so many good lenses in your arsenal!

Even though the SMC Takumar is faster (F2.5 x F2.8 ), the SMC Takumar is lighter (444g x 585g), thinner (67.5mm vs 68.5mm) and shorter (85mm x 93mm) than the Zeiss Sonnar. Certainly, the Takumar has a smaller telephoto ratio. This is an indication that the back group of the Takumar has more optical power (more negative) than the corresponding Sonnar's. Normally, the more negative the back group, the greater the field curvature.


PostPosted: Wed Jul 29, 2020 7:38 pm    Post subject: Reply with quote

Is the depth of field less at the edges than at the center, or an even thickness center to edges?

Are we looking at dof projection same 3-d shape as front element?


PostPosted: Fri Jul 31, 2020 4:20 pm    Post subject: Reply with quote

visualopsins wrote:
Is the depth of field less at the edges than at the center, or an even thickness center to edges?


Interesting question, I have never seen anyone discussing possible variations of the depth of field throughout the image field. Nonetheless, the depth of field depends on the effective aperture, so, in principle, vignetting would produce a slight increase in depth of field at the edges and corners. In practice, the increase in depth of field at the edges and corners is not significant enough to be taken into account.


PostPosted: Sat Aug 01, 2020 6:48 am    Post subject: Reply with quote

Waiting for the zoom lenses part... twiddling my thumb...


PostPosted: Sat Aug 01, 2020 2:15 pm    Post subject: Reply with quote

hasenbein wrote:
Waiting for the zoom lenses part... twiddling my thumb...


Uh-Oh ...!! You are right! I had taken all the images, but forgot to publish them ... i think i'll be able to publish them this evening, just after celebrating a bit our Swiss national holiday Wink

S

EDIT: I have re-tested my S-M-C Takumar 2.5/135mm (II):
1) with a slightly tilted (visible with naked eye if you look carefully!) cheap M42 => Sony E mount adapter: Now three corners are blurred, and one corner is nearly perfect (at f2.5, of course)
2) with a M42 => Sony A mount adapter: all four corners are equally blurred