New 50mm APO article/review + Interview with Peter Karbe at overgaard.dk

[biglou]

Advertisement (gone after registration)

Camera - M240

Lens detection - ON

 

The obvious next test could be to do it again without lens detection, at 2 and then 4 or 5.6 for instance.

One of your lenses could also have a diaphragm calibration problem, who knows.

[freitz]

The obvious next test could be to do it again without lens detection, at 2 and then 4 or 5.6 for instance.

One of your lenses could also have a diaphragm calibration problem, who knows.

 

Yes. Unfortunately I do not own these lenses and had them for a very short time. I cannot redo the test at this time.

[Advertisement]

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

[freitz]

I can try an make the question clearer. I would think the APO has better light transmittance over the Summilux which would be the T stop difference? Thus the F/2 APO might actually be at f/2 and the summilux at 1.4 is actually at f/2 for the amount of light that hits the sensor. Does this make sense?

 

"Three apochromatic glass elements provide a common focus plane for 3 light wavelengths, effectively eliminating chromatic aberration, with two high refractive index elements further improving light transmission."

 

From this website - https://www.lensrentals.com/rent/leica/lenses/normal/leica-50mm-f2-apo-summicron-asph

[biglou]

I can try an make the question clearer. I would think the APO has better light transmittance over the Summilux which would be the T stop difference? Thus the F/2 APO might actually be at f/2 and the summilux at 1.4 is actually at f/2 for the amount of light that hits the sensor. Does this make sense?

 

"Three apochromatic glass elements provide a common focus plane for 3 light wavelengths, effectively eliminating chromatic aberration, with two high refractive index elements further improving light transmission."

 

From this website - https://www.lensrentals.com/rent/leica/lenses/normal/leica-50mm-f2-apo-summicron-asph

 

 

If you reason in terms of T stops there may be a difference but only a small one, 0.1 or maybe at most 0.2 EV

The number of lentils inside are about the same as the number of glass/air separations.

In modern lenses the coatings is very efficient to remove the greatest part of reflexions, transmission is almost perfect.

From your experiment i would not draw conclusions about the lenses.

 

More evident factors are the firmware, the part of the image light measured by the camera cell, the vignette etc...

And the experiment itself also, why not ?

[Lenshacker]

Camera - M240

Lens detection - ON

 

Try it with the lens detection off, that will take firmware corrections out of the equation and make it easier to judge what is going on. If there is a "non-uniformity correction" taking place based on lens coding, the firmware may be dragging the center down to equal the edges. This could be built into the exposure. Easy to factor out in this experiment, just turn off lens detection.

 

Multicoated 50mm lenses are better than 95% transmission. My 1977 Nikkor 55/1.2 is a T1.25. You are losing much more light due to the angle of inclination of light hitting the sensor rather than transmission loss through the lens.

Edited April 6, 2015 by Lenshacker

[BerndReini]

I have used both lenses on my M9 and there is no T stop difference (T stops btw. are mostly of concern with zoom lenses not primes). If anything, the 50 Apo appears darker at f2 than the 50lux at f2 due to stronger vignetting.

[Bateleur]

Advertisement (gone after registration)

Lovely insights and an engaging read. Thank you!

 

Regards

Charles

[lct]

[...] the 50 Apo appears darker at f2 than the 50lux at f2 due to stronger vignetting.

+1

[Erik Gunst Lund]

Thanks, nice with some insights to the thourghts behind the design.

 

You state:

 

When you anyways can see an f-stop in Lightroom it is because it is calculated based on what the small eye on the top front of the Leica M body perceive to be the light condition, compared to how much light hits the sensor.

 

It's 'the small eye' yes, but it's compared to how much light is measured reflecting off the aperture blades before the exposure, not by the sensor. Leica has a patent for this since they came out with the M8 and 6-bit coded lenses.

That's also why lenses that go deep into the m-mount will give less accurate readings, also covering the small eye with a finger will give wrong f-stop 'readings'...

Edited April 7, 2015 by Erik Gunst Lund

[Peter Branch]

When I recently re-calibrated my workflow one unexpected result was the discovery that in the centre of the image my 28mm f/2 Summicron-M ASPH was consistently about 1/3 stop "faster" at all apertures than the other 4 lenses which were all within about 1/8 stop at all apertures.

 

One of the lenses tested was a 50mm f/1.4 Summilux-M ASPH

 

The tests were done under carefully controlled conditions and the determination of "exposure" was made by measuring the densities of the grey scale images of a Spydercheckr24 test card.

 

Some years ago I did a similar test using film with a 50mm f/1 Noctilux-M and a 50mm f/2 Summicron-M. The results were interesting in that in the centre of the image the Noctilux was a true f/1 lens with an image brightness 4 times that of the f/2 lens. However the total light transmission was nothing like 4 times due, one must assume, to the inherent high vignetting of the Noctilux.

[01af]

In testing, the same shot on a tripod in the same light; Summilux at f/1.4, 1/500 s, ISO 200/24° produced a darker image than the Apo-Summicron at f/2, 1/500 s, ISO 200/24°.

Impossible. Something went wrong with your test.

 

Everything else being equal, a shot taken with the Summilux-M 50 mm Asph at full aperture of course will be brighter that the same shot taken with the Apo-Summicron-M 50 mm Asph at full aperture—just as expected. The transmission losses are small and virtually the same for both lenses; no significant difference here.

 

Even with both lenses at f/2, the Summilux picture will appear brighter than the Apo-Summicron picture due to vignetting. A 1:1.4 lens stopped down to f/2 will always vignette less than a 1:2 lens at full aperture. While the brightness will be the same at the frame's center, the pictures' overall appearance won't.

[Overgaard]

Thanks, nice with some insights to the thourghts behind the design.

 

You state:

 

When you anyways can see an f-stop in Lightroom it is because it is calculated based on what the small eye on the top front of the Leica M body perceive to be the light condition, compared to how much light hits the sensor.

 

It's 'the small eye' yes, but it's compared to how much light is measured reflecting off the aperture blades before the exposure, not by the sensor. Leica has a patent for this since they came out with the M8 and 6-bit coded lenses.

That's also why lenses that go deep into the m-mount will give less accurate readings, also covering the small eye with a finger will give wrong f-stop 'readings'...

 

I actually don't know if LR is guessing based on information in the file (as to the difference between what the "small eye" and the lightmeter inside the camera - that information could be recorded in the file) or if LR compare the luminance used in the image to the luminance "outside" the camera based on the small eye. The result is the same.

[Overgaard]

The F-stop and the T-stop is the difference between what the lens is supposed to capture of light based on the F-stop and the actual amount of light coming through the lens.

 

In primes lenses like the 50mm lenses it should be almost no difference. There are talk about that different coatings reduce the light from 0 to 4%.

 

As an f-stop is reduction of light with 50% (going from f/2 to f/2.8 reduces the light through the lens to half) or 100% (going from f/2.0 to f/.1.4 double the light intake), the 0-4% is somewhat a 1/12th f-stop in most extreme cases.

 

The t-stop mostly applies to zoom lenses, tele lenses and complicated constructions. The 85/1.2 from Canon is supposedly a T 1.4 and te 70-200 f/2.8 is supposedly T 3.8 (Don't flame me, I haven't measured it myself).

 

Those lenses, using an external lightmeter ... it's turning on a headache :-)

 

TMI:mad: (Too Much Information).

 

I would say it is 98% unlikely that the t-stop is very far from the f-stop on the Summilux and 50mm APO. I will see if I can find out.

[01af]

From Thorsten's article:

However, my experience with the Apo-Summicron-M 50 mm 1:2 Asph is that the "guessed" f-stop is often wrong. It will go to extremes as f/11 or f/16 when the lens is in fact f/2.

I cannot find any reasonable (or actually any at all) explanation for this and have asked Leica Camera AG. I haven't gotten an answer yet why this could be.

As far as I can tell, there are two factors that seem to affect the accuracy of the aperture "guessimations": the lens speed and the quality of the light. The guessed aperture tends to be more accurate with faster lenses (don't ask me why) and in soft, diffuse light. It usually is less accurate (i. e. guessed aperture smaller than actual aperture) for slower lenses and in harsh backlight conditions.

 

In general, I don't find the guessed apertures in the Apo-Summicron-M 50 mm Asph files signficantly less accurate than with other Leica M lenses. By the way, when the light is too low for the metering system to give any meaningful f-stop estimation then the camera will enter "f/4" into the file's metadata by default.

Edited April 7, 2015 by 01af

[Peter Branch]

If I recall correctly "T" stops first appeared on cine lenses.

 

The differences between the geometric aperture and the light transmitted aperture were never that great but, in an industry which needs to minimise changes in perceived illumination when editing, even tiny differences really matter.

 

Differences between one print and another caused by different light transmission through lenses are effectively irrelevant.

 

I agree with O1af any Leica 50mm f/1.4 lens at full aperture will always give a significantly brighter image than any 50mm f/2 lens at full aperture.

[freitz]

The F-stop and the T-stop is the difference between what the lens is supposed to capture of light based on the F-stop and the actual amount of light coming through the lens.

 

In primes lenses like the 50mm lenses it should be almost no difference. There are talk about that different coatings reduce the light from 0 to 4%.

 

As an f-stop is reduction of light with 50% (going from f/2 to f/2.8 reduces the light through the lens to half) or 100% (going from f/2.0 to f/.1.4 double the light intake), the 0-4% is somewhat a 1/12th f-stop in most extreme cases.

 

The t-stop mostly applies to zoom lenses, tele lenses and complicated constructions. The 85/1.2 from Canon is supposedly a T 1.4 and te 70-200 f/2.8 is supposedly T 3.8 (Don't flame me, I haven't measured it myself).

 

Those lenses, using an external lightmeter ... it's turning on a headache :-)

 

TMI:mad: (Too Much Information).

 

I would say it is 98% unlikely that the t-stop is very far from the f-stop on the Summilux and 50mm APO. I will see if I can find out.

 

Post your results I would like to see. I have a friend that has both I will try to have him re-create my original findings from my test.

[freitz]

Impossible. Something went wrong with your test.

 

Everything else being equal, a shot taken with the Summilux-M 50 mm Asph at full aperture of course will be brighter that the same shot taken with the Apo-Summicron-M 50 mm Asph at full aperture—just as expected. The transmission losses are small and virtually the same for both lenses; no significant difference here.

 

Even with both lenses at f/2, the Summilux picture will appear brighter than the Apo-Summicron picture due to vignetting. A 1:1.4 lens stopped down to f/2 will always vignette less than a 1:2 lens at full aperture. While the brightness will be the same at the frame's center, the pictures' overall appearance won't.

 

I can only post my results from my original test. I saw a clear difference. I am traveling for work right now and will try to pull them images off my server and post them here.

[CheshireCat]

I have a friend that has both I will try to have him re-create my original findings from my test.

 

Be sure to remind him to remove the ND filter from the Lux this time

[freitz]

Be sure to remind him to remove the ND filter from the Lux this time

 

Lol

[Advertisement]

Hallo,

Look here: Leica M

[Erik Gunst Lund]

I actually don't know if LR is guessing based on information in the file (as to the difference between what the "small eye" and the lightmeter inside the camera - that information could be recorded in the file) or if LR compare the luminance used in the image to the luminance "outside" the camera based on the small eye. The result is the same.

 

I know...

 

As I stated, it's from the camera exif;

It's 'the small eye' yes, but it's compared to how much light is measured reflecting off the aperture blades before the exposure, not by the sensor. Leica has a patent for this since they came out with the M8 and 6-bit coded lenses.