Very interesting answer from Leica on 35mm 1.4

[ho_co]

Advertisement (gone after registration)

@Howard--stupid airport network ate my long reply!

 

Anyway--no hackles raised--and no worries; I know we're all trying to help Tim!

Jamie--

Isn't that the pits? All the logic, calculations, personal experiences and generosity ready to go--and the network coughs up a fur ball.

 

Actually, I have to admit I don't even know what hackles are.

 

Seriously, thanks. I know sometimes I come across all wrong, despite my efforts.

 

--HC

[ho_co]

That may be beautiful Carl, but you have no idea how irritating it is...

Yeah, Carl! Get out of here!

 

We want to see Tim's shots of rulers in Barcelona hotel rooms!

 

And you Mark! "Something silly like $20 each way + taxes." Do you have any idea what an 8-hour flight does to my knees? To say nothing about what it does to my pocketbook with today's dollar!

 

--HC

[ho_co]

I quoted above what Mr Puts said about flatness of field when comparing the Zeiss Biogon 35/2 and the Summicron M ASPH 35/2. (http://www.imx.nl/photosite/comments/c016.html)

 

Previously, when he compared the pre-ASPH 35/2, the 35/2 ASPH and the 35/1.4 ASPH, he said of the Summicron: "Most spectacular is the flatness of field of the ASPH. Here we have a lens which at full aperture gives a completely flat image!" (http://www.imx.nl/photosite/leica/mseries/testm/m2-35.html)

 

In that same article, he compares the three lenses at distances around one meter: "Close up performance ... is good for all three lenses. The f/1,4 aperture of Summilux-M ASPH 35mm however shows curvature of field and vignetting. Here I would not go into a detailed comparison. All three are very competent in this area, with Summicron-M ASPH 35mm slightly ahead."

 

I don't find these citations inconsistent. He is comparing at different stages. I think Jamie's citation in regard to the Summilux is one that he wrote for Leica. He also (earlier? later?) while not on Leica's payroll compared two new Leica lenses with an earlier model--a different kettle of fish. Later still, he compares the Summicron with a Zeiss lens. The Summicron was amazingly good when it first came out and outperformed the others. Almost ten years later Zeiss comes with a new design that outperforms the Leica lens in the same aspect where the Summicron had earlier been the winner, but falls below it in every other category.

 

Times change.

 

--HC

[tashley]

Settle in with a glass of wine and check me out on the following :

 

 

 

 

That is: Tim focused on the birdhouse at, say, 10 m. He then rotated the camera by about 17°. He was still 10 m from the birdhouse, but because he had turned a bit, the sensor has changed angle. So we're looking at this situation:

 

Distance from center of sensor to birdhouse when focusing: 10 m

Displacement of birdhouse from center to edge of sensor: ca 10 mm

Rotation of sensor: ca 17°

Displacement of sensor area now receiving image of birdhouse: ca 3 mm further from birdhouse

 

Since I don’t have focus formulas handy, I did a quick calculation of DoF.

 

Assuming:

CoC = 0,002 mm

distance = 10.000 mm

focal length = 35 mm

f-stop = 4

 

Then total depth of field is 1311,72 mm.

 

Same assumptions, except the distance to the birdhouse is now 3 mm further after rotating by 17°:

CoC = 0,002 mm

distance = 10.003 mm

focal length = 35 mm

f-stop = 4

 

Then total DoF is 1312,51 mm, an increase of 0,79 mm.

 

Please correct any of my math. It’s been a while since I played with geometry. The basic assumptions I made are these:

 

If my math is correct (there’s at most an outside chance that it is ), then with these assumptions the difference in DoF implies to my mind that there would be little chance of seeing any focus shift due to rotating the camera through this short angle.

 

 

--HC

 

HC, you leave me breathless!

 

I never did the math but I intuitively felt this to be the case, whereas if I am focussing on a flat-fronted building twenty feet away and the edges are sharper, I know that it is because they are futher away and therefore further into the backfocus zone. That's when Mr pythag would be useful.

 

Wonder if he ever saw a birdbox? The Mystery Of the Birdbox does have a classical ring to it....

 

;-)

 

Tim

[tashley]

It was suggested yesterday that I do shots of all corners as well as the centre. Lovely weather today and gorgeous light, plus, yes, there's a bird living there (but I didn't get him in shot - flighty little bugger).

 

35 Cron f4, fixed exposure and focussed on the box then twisted my angle for each corner whilst staying in the same spot - and as HC has Pythogorized, this should be an accurate-ish method.

 

Centre:

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

Bottom Left:

 

Bottom Right:

 

Top Right:

 

Top Left:

 

 

Tim

[marknorton]

It's clear, same results from each corner and the centre is off...

 

This is at f4; as you change aperture, how does it change?

[tashley]

Advertisement (gone after registration)

It's clear, same results from each corner and the centre is off...

 

This is at f4; as you change aperture, how does it change?

 

 

Hi Mark,

 

At F2 it's pretty sharp across the frame, between f2.8 and f5.6 you see the same sort of thing as F4 but to slightly verying degrees. By F8 you're almost sharp everywhere, F11 is great and by F16 it's the same sharpness everywhere but a little soft in general.

 

I use f4 because the effect is most pronounced there, though this appears to vary slightly with distance to subject and so on.

 

Can you run the exact same Heath Robinson test with yours? At f4 and distance to subject of twenty to twenty five feet?

 

Best

 

TIm

[egibaud]

:-)))

 

Today I spent 2 hours at Casanova Foto in Barcelona, Spain.

 

I brought back my Lux 35 and before getting back home with anything I wanted to do some testing.

 

Results were:

 

35cron: sharp center, softer corners at 2.0. Everything sharp at 2.8

50cron: sharp center, softer corners at 2.0. Everything sharp at 2.8

90 Elmarti: everything sharp at 2.8 althoug maybe corners were slightly worse, perfect at 4.0

 

This time all lenses, included the 35cron were perfectly sharp at all aperture in the CENTER. My 35Lux and Tim's were soft in the center!

 

So I left home with 50cron and 90 Elmarit and I'am on waiting list for a 28 Elmarit.

So I have changed my plans from 15-35-75 to 15-28-50-90 overall cost is the same although I lose speed as I had 1.4 with the 35.

 

I am glad I did not touch the RF.

 

Thanks to all,

 

Eric

[tashley]

Are Casanova good? I'm going to Barcelona tomorrow!

 

Just for reference, my first two returns were 35 lux, the one I have now is a 35 cron. I also tested in-store and it seemed ok, but more rigorous tests at home disagreed.

 

Keep well

 

Tim

[wparsonsgisnet]

Bill, thanks for running this. What do you take from it? To me, the 5.6 shot shows focus shift backwards. If you did it at f4 I suspect it would look worse?

 

Best

 

Tim

 

Tim, I agree that the 5.6 shows less sharpness than the 1.4.

 

Unfortunately, these were hand held. I will do this again, including the 24, the 50, and the 75 as well.

 

I don't usually use the near-focus distance of my lenses. Usually, I am most interested in what's 7-20 feet away.

 

I'll work on this tonite.

 

Regards,

[marknorton]

OK, I've done some more testing, with my 35/1.4. Two tea cartons, one at the centre, one at the edge of the frame, and here are the results (100% crops)

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

f1.4 Centre:

 

 

f1.4 Edge:

 

 

f4 Centre:

 

 

f4 Edge:

 

 

f11 Centre:

 

 

f11 Edge:

 

 

These are shot at a distance of about 1.2m and certainly show some centre softness at f4; more to me, it shows how great the lens is at f1.4...

 

Tea anyone?

[AlanG]

HC, you leave me breathless!

 

I never did the math but I intuitively felt this to be the case, whereas if I am focussing on a flat-fronted building twenty feet away and the edges are sharper, I know that it is because they are futher away and therefore further into the backfocus zone. That's when Mr pythag would be useful.

 

Wonder if he ever saw a birdbox? The Mystery Of the Birdbox does have a classical ring to it....

 

;-)

 

Tim

 

Actually Tim, if the lens has a relatively flat field, it is supposed to focus on a plane - center to edge. Thus the hypotenuse distance is not important. Reframing the center subject to the corner defeats the purpose. The focusing distance is calculated on axis. The corner focus distance is adjusted by lens design. You'd be better off focusing on a flat wall. But you know the result of that already.

 

If your lens is focusing differently in the corners than at the center, then it is not functioning as a flat field lens, regardless of the design claims. I think your many tests have clearly proven this is the case with your 35 1.4. But even if you accept that it is not a flat field lens, the focusing mount of the lens should probably be adjusted so that it focuses accurately at the center. By the way, did you ever try measuring the focus distance with a tape measure from the sensor plane to compare with a specific marking on the barrel such as 10 feet? Then you could see if the calibration of the mount is off. (You said the rangefinder was accurate with your other lenses so you don't want to alter that setting.)

 

Now you see why I do most of my critical work shooting tethered. I can zoom in and check focus and depth of field to make sure it is what I want.

[tashley]

OK, I've done some more testing, with my 35/1.4. Two tea cartons, one at the centre, one at the edge of the frame, and here are the results (100% crops)

 

[ATTACH]30445[/ATTACH]

 

f1.4 Centre:

 

[ATTACH]30442[/ATTACH]

 

f1.4 Edge:

 

[ATTACH]30443[/ATTACH]

 

f4 Centre:

 

[ATTACH]30438[/ATTACH]

 

f4 Edge:

 

[ATTACH]30439[/ATTACH]

 

f11 Centre:

 

[ATTACH]30440[/ATTACH]

 

f11 Edge:

 

[ATTACH]30441[/ATTACH]

 

These are shot at a distance of about 1.2m and certainly show some centre softness at f4; more to me, it shows how great the lens is at f1.4...

 

Tea anyone?

 

Verrry eentersting herr doktor.. seems that you have what I have... so let's both have a cup of tea! And yes, it is FAB at 1.4...

 

Tim

[tashley]

Actually Tim, if the lens has a relatively flat field, it is supposed to focus on a plane - center to edge. Thus the hypotenuse distance is not important. Reframing the center subject to the corner defeats the purpose. The focusing distance is calculated on axis. The corner focus distance is adjusted by lens design. You'd be better off focusing on a flat wall. But you know the result of that already.

 

If your lens is focusing differently in the corners than at the center, then it is not functioning as a flat field lens, regardless of the design claims. I think your many tests have clearly proven this is the case with your 35 1.4. But even if you accept that it is not a flat field lens, the focusing mount of the lens should probably be adjusted so that it focuses accurately at the center. By the way, did you ever try measuring the focus distance with a tape measure from the sensor plane to compare with a specific marking on the barrel such as 10 feet? Then you could see if the calibration of the mount is off. (You said the rangefinder was accurate with your other lenses so you don't want to alter that setting.)

 

Now you see why I do most of my critical work shooting tethered. I can zoom in and check focus and depth of field to make sure it is what I want.

 

Just to calrify Alan, teh more recent tests are on a 35 cron, I sent two luxes back. You are right about the above of course, but I was merely trying to show with that test that the lens mount or sensor alignment aren't off and to explain why so many people (like Mark above) have this problem without knowing it. Rules of thirds and so on mean that we tend to focus and then re-frame, reducing the evidence!

 

Best

 

Tim

[tashley]

Tim, I agree that the 5.6 shows less sharpness than the 1.4.

 

Unfortunately, these were hand held. I will do this again, including the 24, the 50, and the 75 as well.

 

I don't usually use the near-focus distance of my lenses. Usually, I am most interested in what's 7-20 feet away.

 

I'll work on this tonite.

 

Regards,

 

 

Be interested to see the results. I think most people will have this issue when they really test for it rigorously. I just want to get my hands on Jamie's lux, torture test it and then steal it!

 

;0

 

Tim

[sergiolov]

Settle in with a glass of wine and check me out on the following :

 

 

Sergio--

I agree completely, but that is beside the point that Mark and I were making. Obviously I didn't explain clearly, so I’ll try to be less oblique here.

 

First:

Whether we consider the geometry or not, we should see the same difference in focus when shifting from center to left as we see when shifting from center to right.

 

If we find a different shift of focus when swinging 17° left as compared to swinging 17° right, then we have a camera or lens problem.

 

In other words, whatever the shift of focus is, it should be symmetrical about the lens axis.

 

Second:

Indeed, the way the 'reframe procedure' is usually described, it is recommended that one "focus on another object at the same distance as the subject." If one does this, there are no geometric repercussions because both objects are the same distance from the camera: Focus on one object 10 m away, then turn and take a picture of another object also 10 m away.

 

Third—and here is where our friend Mr Pythagoras and your calculations enter the fray:

As I understand it, Tim wasn't trying to reframe by focusing on one object and then changing the subject. He focused on the birdhouse, then shifted the camera to put the birdhouse to the side of the frame. The edge of the sensor would have moved rearward a couple millimeters, changing the actual distance by that amount.

 

That is: Tim focused on the birdhouse at, say, 10 m. He then rotated the camera by about 17°. He was still 10 m from the birdhouse, but because he had turned a bit, the sensor has changed angle. So we're looking at this situation:

 

Distance from center of sensor to birdhouse when focusing: 10 m

Displacement of birdhouse from center to edge of sensor: ca 10 mm

Rotation of sensor: ca 17°

Displacement of sensor area now receiving image of birdhouse: ca 3 mm further from birdhouse

 

Since I don’t have focus formulas handy, I did a quick calculation of DoF.

 

Assuming:

CoC = 0,002 mm

distance = 10.000 mm

focal length = 35 mm

f-stop = 4

 

Then total depth of field is 1311,72 mm.

 

Same assumptions, except the distance to the birdhouse is now 3 mm further after rotating by 17°:

CoC = 0,002 mm

distance = 10.003 mm

focal length = 35 mm

f-stop = 4

 

Then total DoF is 1312,51 mm, an increase of 0,79 mm.

 

Please correct any of my math. It’s been a while since I played with geometry. The basic assumptions I made are these:

 

If my math is correct (there’s at most an outside chance that it is ), then with these assumptions the difference in DoF implies to my mind that there would be little chance of seeing any focus shift due to rotating the camera through this short angle.

 

 

 

I’m glad you brought up the issue, Sergio, because I had never considered it and it is clearly something that must be recognized.

 

BTW, I really like your phrase “the mystery of the bird box.” There's a charm and playfulness in it. Your analysis of the lens adjustment also seems spot on.

 

There is one place where I believe Dale and you may be in error in regard to the lens assembly, but you may have put your finger on the source of Tim’s problem.

 

As I understand it, the focusing helicoid has a thread on both sides, and acts as a differential between the focusing ring and the lens head. Thus the outside of he helicoid may be cut the same for all lenses of a given design, but the inner thread is cut specifically for focal length variations. Leica previously did exactly this for lenses 50mm and longer, and never used shims. For that reason I doubt that they shim the wideangles.

 

Today I believe most if not all Leica lenses are labeled as having their design focal length. E Puts says that Leica's tight design standards make variances from plan virtually impossible. That is, the lens assembly is a go/no-go procedure: The assembly itself is part of the quality control.

 

Now if some tolerance is allowed in the optical units of wideangles but they are all mounted in the same helicoid, that could explain the differences between Jamie’s, Mark’s, Eric’s and Tim’s experiences. (And I think that's what we're searching for at least until Tim returns from Barcelona.)

 

Just an idea, following up on your suggestion that Leica use shims. Only a possibility, and one I tend to discount.

 

BTW, I believe that instead of using shims to adjust lenses, Leica simply adjusts the optical unit forward or back in the focusing mount, as Dale also mentioned.

 

--HC

 

Howard,

first off all, thanks for taking the time to read my not so clear post. I have real difficulty in finding the right words, so I must try to expose my thoughts with the words I know.

Starting from the end, I completely agree about the shims. I don't know which method Leica uses to calibrate the lenses, so I just followed Dale in this, but the concept is the same.

 

Regarding the focus and recompose. The plane that you must consider is not only that of the sensor, but essentially that of the subject. A lens focuses on a distant plane, and when you rotate the camera, you rotate that plane. Take a small triangular piece of paper,

with one vertex simulating the camera and the opposite base simulating the plane that contains in focus subjects. Rotate it, and see how the subjects go out of focus....

 

If your triangle has a vertex angle of approx. 45 degrees, it reproduces the viewing angle of the 35mm lens on the m8. ( in case of a focus distance of 10 meters, rotating the camera of 17,5 degrees will put out of focus the original subject of 460mm, or D minus Dcos.17,5deg.) The mistery is solved... but I am not so sure that I did not expose it in such a way that I am the only one to understand it, LOL.

 

Let me know what you think.

 

Sergio

 

 

PS. I read now the post from Alan.G. I think he probably explained the thing better than I did

[wlaidlaw]

The tape measure test should be interesting. Before I retired, my office was in the Spitalfields area of the City of London, which was often used for filming period dramas. I was always amazed to see the so-called focus pullers measuring the camera to actor distances with a tape measure, when the actor was standing on his chalk mark during the run throughs. They then wrote all these distances down in a notebook. When they were actually filming, the focus puller was adjusting the focus ring by hand, referring to his notebook - no new-fangled rangefinders and as for auto-focus....

 

I was talking to a cameraman when a man came up, took the lens off the camera and went away with it. I asked what he was doing and the cameraman told me he was the Zeiss technician, who was taking the lens away for a field service and clean. I asked if this did not upset the focus and image. It was explained to me that these Arri-Zeiss lenses were designed for field disassembly and would go back together perfectly. That was why they cost upwards of £30,000. And we think Leica lenses are expensive.

 

Wilson

[tashley]

 

I was talking to a cameraman when a man came up, took the lens off the camera and went away with it. I asked what he was doing and the cameraman told me he was the Zeiss technician, who was taking the lens away for a field service and clean. I asked if this did not upset the focus and image. It was explained to me that these Arri-Zeiss lenses were designed for field disassembly and would go back together perfectly. That was why they cost upwards of £30,000. And we think Leica lenses are expensive.

 

Wilson

Hmmm, I want me one of them. Do you think they make an adaptor?

 

;-)

 

Tim

[wlaidlaw]

Hmmm, I want me one of them. Do you think they make an adaptor?

 

;-)

 

Tim

 

The answer is yes Tim. They were showing an adapter on the Zeiss stand at Photokina last year. They had a lens the size of a large vegetable marrow on a Zeiss Ikon i.e. M mount. I assume if you can afford the lens and adapter, you can afford to employ a focus puller to measure your distances and focus the lens for you.

 

I once watched Lord Snowdon at work. He had two assistants reloading backs and bodies and passing them to him (he was using Hassleblad and Nikon). He passed the camera back to an assistant if he wanted a lens changed.

 

"Mop my brow serf, put the 55mm Distagon on and look snappy about it" - that's me saying that to my assistant not Snowdon who was unfailingly polite to everyone - even the ones who asked him REALLY silly questions.

 

Wilson

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[ho_co]

Let me know what you think.

...

PS. I read now the post from Alan.G. I think he probably explained the thing better than I did

Sergio--

First off, I think your English is quite good, but I do know the strain one goes through to get technical in a foreign language.

 

I agree completely with your analysis, in both your first and your second posts, and I agree with Alan's as well.

 

The reasons I approached the matter as I did were:

1) Tim's birdhouse is not in my mind a 'focus and recompose' situation;

2) You raised the geometry question quite succinctly;

3) In Tim's non-'focus-and-recompose' situation, the only place I see applicability of the focus plane vs subject plane relationship is at the sensor.

 

My math skills are quite challenged, and your post may have just completely misproved what I thought, but kindly and subtly! I will go back and evaluate my slow and tortuous grasp of the situation in the light of your immediate one.

 

Summary: As I see it, your input is directly related to why photographing a plane--say, a building wall--might show different focus at center and edge, simply because the edges are further than the center; and also to why if you focus on the end of the wall, the center is likely not to be sharp. Rotating the camera will move the new focus plane to intersect with the first plane, and will change the distance to objects in the first plane.

 

But also as I see it, in Tim's case of shooting the birdbox, the angle will have changed so little that the change in distance to the birdbox is negligible.

 

Back to you, guys! I know I may be missing something important here, and I'd love for you to point it out. After reading your second post a couple times and looking again at Alan's, I am becoming fuzzily aware that there is something that is at the moment just outside my grasp, Could I be more confused than I think?

 

 

In the matter of Tim's various lenses, both Jamie and I quoted Erwin Puts above. One of his assessments of the 35/1.4 implied that it was completely flat field. Later he compared it to the 35/2 and said the 35/2 had flatter field than the Summilux (which is to say, the Summilux' field isn't all *that* flat). And later still, he tested the Summicron (the winner against the Summilux, remember) against the Zeiss Biogon, and found that although he prefers the Leica lens overall to the ZM design, the Zeiss lens did have a flatter field than the Summicron.

 

So with the willingness of so many people to look at this issue, I think we've got the solution: The 35 Summilux is a great lens and has a very flat field for a lens of this speed, but particularly with a digital sensor (we never saw these complaints before the M8, did we?), it isn't as flat-field as we might want. The 35 Summicron is better in that regard but still not perfect. The 35 Biogon uses a far simpler design and is able to achieve the flattest field of the three, but at the expense of almost every other optical characteristic.

 

My respects to all. I hope I haven't misunderstood something, and I hope I've been helpful. I think Tim's set of 4-cornered birdhouses with the Summicron and Mark's of tea-bag packages with the Summilux support this assessment.

 

--HC