28mm Summicron on micro four thirds? corner softness?

[martinb]

Advertisement (gone after registration)

I have a chrome Summicron-M 28mm not in use at the moment, because I currently don't have a M body.

I've been thinking of selling it, but I also think that now might not be the best time do so. Bad economy etc.

Then I've started to think about using it on a micro four thirds body, probably the Olympus EP-1. I know there's the crop factor of 2, but I like 50mm or equivalent lenses anyway and think that the added DOF might be benititial in some cases.

 

However, I've read reports about corner softness with this lens on M4/3. What's your experience? is it a possible adapter issue? or is it just the way it is and no cure for the problem?

I would be happy for any input! Thanks!

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[RonSmith]

I have been using the 28 Summicron on my G1 (with a Voigtlander adapter) for just over three months. Most often I am shooting at large apertures, so the DOF usually renders the outer parts of the frame out of focus. I've been pretty satisfied with the images I've been getting though.

 

Here are a couple samples:

 

 

 

[martinb]

Thanks Ron!

Looks good, but hard to judge corner softness from.

[denniswksit]

I have been using EP1 since it has been sold in the market. I usually put in my 28mm Summicron on the body for snap shot. During the weeks, I am satisfied with the quality of the images. The very edges might be a little bit soften, but it does not really affect the image quality as a whole.

[martinb]

I don't think I'll go any further with this If it's not possible to get sharpness in the corners. I won't bother getting a 150 USD adapter to put my 4000 USD lens on to a 700 USD body. Doesn't make sense for me if it doesn't perform like it should.

[marknorton]

TBH, I think you're right.

 

I was very enthusiastic about the ability to focus through the lens but I think it was Sean Reid who first showed that the IQ with the Leica wideangles, normal lenses in MFT, is nothing special. As it happens, I think the Panasonic kit lens is quite reasonable and my MSO is enjoying using the G1 as delivered.

 

All this goes to show Leica were not telling porkies when they said supporting wide angles is difficult. They did well with the M8 as it turned out and we expect them to have, somehow, cracked the more difficult problem with the M9.

 

Doesn't disguise the fact the G1 is a perfectly reasonable P&S but as Tim Ashley pointed out, if you have the G1 and M8/<choose your lens> to choose from, why would you ever go for the G1?

[jankap]

Advertisement (gone after registration)

What could be the reason of this corner softness?

It is hard to believe that this can happen.

The image is cropped radically with the best part remaining.

Jan

[ho_co]

This was covered with surprise but in depth when the G1 first appeared: There was a very long thread on the topic, whose consensus IIRC was that shorter M lenses, being of generally less retrofocal design than similar focal lengths for reflex cameras, send light rays to the sensor at a steeper angle than it's designed for.

 

There's nothing wrong with the lens, and nothing wrong with the Micro FourThirds sensor; it's just that they weren't intended to work together. Future sensor designs could change that.

 

In addition, the problem can't lie with the adapter, since it's nothing but a lens mount extension.

[lct]

Are there microlenses on the G1 sensor?

[jankap]

In addition, the problem can't lie with the adapter, since it's nothing but a lens mount extension.

An extension ring is an element in the optical path too. It has to kill the light, that is cropped off. At least part of it.

Jan

[leicanut2]

Ron I like the pictures, I just got the panasonic r to micro 4/3 adapter but I am torn between the E-P1 and the G1.

Also both companies are also coming out with a new version by the end of the year. And I gotta see what Leica does on 9/9.

I just sold my R8.

 

Jan

[ho_co]

No, Jan. The lens is designed for a particular distance, flange to sensor. The adapter simply adds the missing extension to allow the lens to focus to infinity. The adapter isn't an "extension ring."

 

Extension tubes are used to focus closer than the lens' helicoid allows; they are not "in the optical path," but locate the lens further from the sensor. The increased distance means the light travels further, and its intensity is decreased. Same thing happens when you focus a macro from infinity to 1:2: There is a reduction of the amount of light transmitted.

 

That's completely different from the use of the adapter in question, because the adapter is a tube that returns the lens to its infinity focus capability.

[SJP]

I do not see how a lens that is sharp corner to corner on a FF sensor (i.e film) or a M8 sensor could be less sharp on a micro 4/3 camera which covers a even smaller area of the focal plane. If anything a 4/3 sensor should be even sharper as you remain closer to the optical axis.

 

The absence of microlenses might concievably cause vignetting (intensity not color) but I just do not see how the sharpness could be influenced.

[lct]

...The absence of microlenses might concievably cause vignetting (intensity not color) but I just do not see how the sharpness could be influenced.

Happens with wides on FF bodies or APS with less than perfect microlenses yet. Not only light falls off in the corners but there is a loss of sharpness in some cases as well. Examples: early Elmarit-R 28/2.8 on Canon 5D or CV 21/4 P on Epson R-D1.

[sandymc]

What could be the reason of this corner softness

 

It's actually spherical aberration due to a combination of the angle of light in the corners and the thickness of the IR filter/coating on the sensor. Hasselblad explain this better than I can, complete with diagrams - the same thing happens with the Biogon 38mm on some digital backs. You can find the article here:

Luminous Landscape Forum (Powered by Invision Power Board)

 

Sandy

[SJP]

combination of the angle of light in the corners and the thickness of the IR filter/coating on the sensor

the M8 sensor also has a certain thickness of coatings etc. and I sincerely doubt that Kodak sensors are completely different from the rest of the pack in that respect. The angle of light of a Leica lens with an adapter ring (larger distance that the regular micro 4/3 lenses) should be roughly the same as the normal optics especially if you include the larger crop factor/smaller sensor. I still do not see it.

 

Spherical abberation is important if a lens does not manage to reach the corners of the sensor in in focus, but the Leica glass is intended to be able to do just that over a much larger surface area than micro4/3. So the point is not applicable here and or only holds for inferior Haseleblad optics

[ho_co]

Stephen--

Your logic is good so far as it goes. With a smaller sensor, the edges of the image are cut off. And since the edges of the lens field are less well corrected than the center, only the better-corrected image rays remain.

 

But that's not the problem here. Consider a room full of identical buckets, all standing face up on the floor, each in contact with the next orthogonally, both North-South and East-West. You're standing on the same floor at one end of the room with a box of tennis balls. Your task is to toss one ball and only one ball into each bucket without moving from where you're standing. The box of tennis balls contains as many balls as there are buckets, each labeled with the number of the bucket it's to enter.

 

Because some buckets are further away, they present only a narrow ellipse for you to aim at. Other buckets are right in front of you, and you have a larger circular opening to drop your balls into. The narrower ellipses present the problem that your ball must hit exactly, else it will bounce off and either out of the bucket area or into a wrong bucket.

 

The buckets obviously represent the wells into which you're trying to drop the photons (balls).

 

Microlenses on the M8 are aligned to help direct the balls. If the Micro FourThirds cameras use microlenses, they aren't aligned for light coming from steep angles. That's one of the breakthroughs of the Kodak/M8 sensor design.

 

Remember as well that the buckets are set up in a Bayer array. That means four buckets must all get the same information to register a datum accurately. A sensor design that didn't require four pixels for a data point would theoretically have better results at the edges, since fewer photons would have to hit the buckets properly to count.

 

Or just look at facts: The Panasonic lenses all (or the shorter ones at least) require firmware correction to produce good results. Since Leica doesn't want its name used on lenses that aren't optically sufficient, but require tweaking to be brought up to snuff, the shorter lenses from Panasonic are all labeled Lumix, not Leica.

 

This was covered in painful detail in the nearly 600 posts of: http://www.l-camera-forum.com/leica-forum/leica-m8-forum/73949-dreaded-comparison-thread-g1-m8-2-a.html. (I think that's the relevant thread. )

 

Wide angle R lenses work fine on Micro FourThirds. Wide angle M lenses don't.

[sandymc]

the M8 sensor also has a certain thickness of coatings etc. and I sincerely doubt that Kodak sensors are completely different from the rest of the pack in that respect.

 

No, actually, the thickness of coating is exactly what makes the M8 sensor different from the rest of pack. In order not to get the spherical aberration, Leica/Kodak had to make the AA filter and IR filter thinner than other sensors. That's why we have moire, and why we need external IR filters.

 

Sandy

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[lct]

...If the Micro FourThirds cameras use microlenses, they aren't aligned for light coming from steep angles...

So microlenses or lack thereof would be the culprits in your opinion. This could explain why the 'cron 28/2 shows some vignetting with the R-D1 as well but losses of data are negligible despite its larger sensor.

[ho_co]

Yes, just about, lct.

 

My guess is that it's not so much the presence or absence of micro lenses, but their special arrangement in the M8.

 

Similarly: An acquaintance says his Zeiss Hologon 15/8 produces a decent image on his M8. But when a friend tried it on an R-D1, he was very disappointed with the light fall-off and detail loss toward the edges. (Yes, I know it's a forbidden usage, but you know us Leica tinkerers. )

 

And as Sandy said, there are a lot of factors involved, for which the M8 sensor was designed to compensate as fully as possible. To avoid the problem described in the Hasselblad download, Kodak/Leica designed the IR-absorbing cover glass much thinner than usual. They compensated for the light fall-off with the concentrically displaced microlenses, but in thinning the cover glass they allowed a lot more IR through.