50mm APO-Summicron - reported aperture smaller than 50mm Summilux-ASPH?

[jaapv]

Advertisement (gone after registration)

9 hours ago, Chazphoto said:

Jaap, did you also see an F stop to T stop comparison on DXO for the 50mm APO?  Would be curious to see if my observations are borne out.

For those writing about the real-world significance or getting exercised about it, I'm not on the same page as Jean-Michel.  Useful to know, but once known you deal with it - and it is mostly not relevant.

Thanks

Chazphoto

DXO does not list any other Leica lenses unfortunately 

[jaapv]

11 hours ago, Warton said:

half stop?! You gonna be kidding me. If that were true, you would have noticed the aperture size difference just by looking at it. You don't even need to do any test to figure out a 1/2 stop difference in aperture. Even sloppiest lens maker in this world would not make this kind of mistake.

it's a 50mm lens, meaning at f/2 the diameter is 25mm; at f/2.8 the diameter is 17.9mm; half stop between is around f/2.4 which is almost 21mm.

You cannot tell the difference between 25mm and 21mm side by side? Sorry, I cannot believe this.

Exactly what don’t you believe? That a mathematically calculated aperture value does not necessarily represent the amount of light transmission of the glass in the lens?  

[Greenhilltony]

Think about the nature of F-stops, you can understand that they offer a way to step-wisely control your exposure on the same lens, but hard to say transferable to every other lens. This could lead to error when relying on an external meter or other non-TTL metering approaches, obviously, and all you can do is compensate it by learning from the results. 

[Jared]

3 hours ago, jaapv said:

Exactly what don’t you believe? That a mathematically calculated aperture value does not necessarily represent the amount of light transmission of the glass in the lens?  

I think what he meant was the diameter of the lens at full aperture couldn’t be so small as 21mm rather than the expected 25mm or it would be easily visible to the casual observer. Thus, the explanation must lie in either vignetting or transmission, not in Leica making a mistake on the physical diameter of the aperture.

[Jared]

On 7/22/2023 at 4:40 PM, Chazphoto said:

Hi,

Having kicked this off, I thought I should record further observations.  In short, I think that the APO-Summicron 50mm requires half a stop more exposure at f2 than the 50mm Summilux ASPH, v1.  Here are my findings.

I popped into Leica Store Mayfair today with my M10-P and 50mm Summilux ASPH (BC version 1) and borrowed their 50mm APO-Summicron.  It has been an overcast, drizzly day in London; good even, unchanging light for a test.  I took two photographs out of the window, focused on a traffic sign on the other side of the road, from exactly the same position in the shop.  First at f2 and second at f2.8 with the Summilux and then the same with the APO-Summicron.  The shutter speed on the 'Lux shot was 1/500th and on the APO shot was 1/360.  The shutter speed on the f2.8 shots were both 1/250th.  I took a further exposure at f5.6 on both lenses and the shutter speed was 1/60th.  (If relevant, no hood deployed on either lens.)

I talked with one of the staff and while talking tested the same f2 and f2.8 shots on an inside wall, which was evenly lit for one of the large black & white photos on display.  Same half stop difference in shutter speeds was observed at f2 between the two lenses and same shutter speed shown at f2.8.  The lady I was talking with suggested that this may be something to do with the firmware not recognising the new lens (which doesn't make sense given that the APO has been out for approx 10 years) and provided an M11 to test with.  That was interesting, because it defaults to multi-zone metering (I assume) and the shutter speeds were different, but the 50 'Lux required 1/640 and the 50 APO required 1/500.  

So, two cameras tested with the same lenses in two scenes and same half stop.  The accidental observation I had made a week and a bit ago and mistaken for odd reporting of the aperture in software, is harder to measure because I wasn't trying to keep the frame and exposures the same.  However, assuming there was some difference from the lenses, that was testing another 50mm APO against a regular shape 50mm Summilux, both secondhand.

I would speculate whether the vignetting observed on the APO at f2 could be causing this?  Seems too near the edges to be the cause.  The lady in the shop suggested the number of elements in the 50mm APO absorbed more light than the 50mm Summilux, which I think doesn't make sense.  My guess is that the lens is just shy of f2 fully open - a phenomenon seen with some f1.4 lenses I've had in the past, where there is not a full stop less in shutter speed when you open up.  (Not something I've tested on my 50mm Lux BC.)

Cheers

Chazphoto

I think you may need to run another test to narrow this down a bit more. Try shooting an evenly illuminated surface with both lenses set to the exact same aperture, ISO, and shutter speeds. Make sure you have any lens profikes disabled in LR or Capture One that might try to correct for distortion and vignetting. Then measure the resulting brightnesses in the center and as you work your way towards the corners.

Just because the light meter recommended a different shutter speed with the APO doesn’t tell you much by itself. I suspect you will find the APO, because it is being shot wide open at f/2, has a bit more vignetting, so brightnesses for the two exposures will be quite close in the middle of the frame, but the APO has more falloff. By the time you reach f/2.8 or slower, both lenses will be stopped down a bit from maximum, and illumination will be more even.

It is perfectly credible that the APO has lower light transmission in general. This could be a combination of glass type, thickness of elements, coating differences, etc. It is even possible that the APO has less light because it is controlling scatter at f/2 better than the Summilux. If I were to guess, though, my money would be more vignetting with the APO when set to f/2. That’s full aperture for the APO but one stop down for the Summilux. Plus, the exposure difference you mentioned goes away at smaller apertures. This suggests vignetting since a mistake in the physical size of the opening is extremely unlikely.

[wizard]

vor 15 Minuten schrieb Jared:

If I were to guess, though, my money would be more vignetting with the APO when set to f/2. That’s full aperture for the APO but one stop down for the Summilux. Plus, the exposure difference you mentioned goes away at smaller apertures. This suggests vignetting since a mistake in the physical size of the opening is extremely unlikely.

Exactly my thoughts, too.

[jaapv]

Advertisement (gone after registration)

Nobody has suggested a mistake in the aperture size   here is  vignetting diagrams of the

 

 

apo-Summicron

 

0034043446833456

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!

 

 

And here the Summilux (by Ken Rockwell)

 

 

 

Not much difference...

 

 

 

[Chazphoto]

Quick snip of the Summilux vignetting graph from Leica, to compare with the APO-Summicron graph that Jaap posted above.  Doesn't show the position at f2, contrary to the Rockwell results, but the differences at f2.8 are interesting.

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!

[Chazphoto]

Quick response to Jared.

The test I ran was similar, in terms of even lighting (albeit of a scene not an evenly lit blank wall), same ISO and same aperture (of course).  I only checked the shutter speeds in camera because I was in the shop and I haven't viewed the files in C1 Pro or LR.  Not sure how I would perform the measurement you mention.

Vignetting was my initial thought also for very much the reasons you state.  I wasn't sure, however, if that made enough sense given the centre-weighting pattern of the M10's metering.  In other words, could the lower amount of light at the edges of the frame, where the weighting given by the meter is lower, cause a half stop increase in exposure?  The fact that the difference disappears from f2.8 onwards, where we know vignetting performance is closer, does suggest that vignetting is the cause.  As I say, it is interesting, but not necessarily real-world relevant.

Just a comment about f2 being 25mm (and I may also have misunderstood Warton's post #18); my understanding is that it isn't the width of the front element or the diameter of the aperture that is the numerator on the fraction, but rather the diameter of the widest part of the cone of light that can be admitted through the lens.  With the position of the aperture blades on the M lenses being relatively closer to the front element than on an SLR lens, you might be able to reason that the aperture size is visible, but I doubt you could actually measure it by eye.  Again, this is only my understanding and if anyone knows about lens construction more professionally, please say!

Good thread.

Thanks

 

Chazphoto

 

 

[lct]

I suspect the culprit is vignetting here. In AE mode with same apertures (f/2 on lenses, f/2.8 in RawDigger) and same ISO values (64), the camera chooses a faster shutter speed for the 50/1.4 asph v1 (1/800s) than for the 50/2 apo (1/640s) and OoC jpegs look darker with the latter. 

M11 + 50/1.4 asph v1:
https://photos.smugmug.com/photos/i-kVBmzzs/0/260e3666/X3/i-kVBmzzs-X3.jpg

M11 + 50/2 apo:
https://photos.smugmug.com/photos/i-pDfx3k4/0/615c0335/X3/i-pDfx3k4-X3.jpg

[Jared]

10 hours ago, Chazphoto said:

Quick response to Jared.

The test I ran was similar, in terms of even lighting (albeit of a scene not an evenly lit blank wall), same ISO and same aperture (of course).  I only checked the shutter speeds in camera because I was in the shop and I haven't viewed the files in C1 Pro or LR.  Not sure how I would perform the measurement you mention.

Vignetting was my initial thought also for very much the reasons you state.  I wasn't sure, however, if that made enough sense given the centre-weighting pattern of the M10's metering.  In other words, could the lower amount of light at the edges of the frame, where the weighting given by the meter is lower, cause a half stop increase in exposure?  The fact that the difference disappears from f2.8 onwards, where we know vignetting performance is closer, does suggest that vignetting is the cause.  As I say, it is interesting, but not necessarily real-world relevant.

Just a comment about f2 being 25mm (and I may also have misunderstood Warton's post #18); my understanding is that it isn't the width of the front element or the diameter of the aperture that is the numerator on the fraction, but rather the diameter of the widest part of the cone of light that can be admitted through the lens.  With the position of the aperture blades on the M lenses being relatively closer to the front element than on an SLR lens, you might be able to reason that the aperture size is visible, but I doubt you could actually measure it by eye.  Again, this is only my understanding and if anyone knows about lens construction more professionally, please say!

Good thread.

Thanks

 

Chazphoto

 

 

Keep in mind the M10’s “center weighted” metering is a little different from film M’s. It’s based on the full length of a shutter blade rather than just a center spot, so it may well give a bit more emphasis to the edges of the frame. No it is not   It is limited toward the edges of the frame by the angle of view of the  measuring cell. Film Ms measure from the film amen the shutter is open, again within the FOV of the cell. There is a measured metering pattern in the FAQ. 

[IkarusJohn]

I must say, I find the explanations implausible.  Leica lenses come from a film background, which means that you take a light reading with your hand held meter (which knows not which lens you are using), and you set the shutter and aperture, relying on what is on your camera and lens.  Shooting a whole role of film using a new APO lens, only to find that the f/stop is not consistent with your other lenses makes no sense to me - t-stop or f-stop, we work on what is printed on the lens barrel.

So, having both lenses, I thought the proof of a pudding is in the eating.  Out came my tripod, and I realised that my 50 Summilux ASPH is in Wetzlar, and I don't have a digital M camera (they're in Wetzlar too).

So, SL with adapter, and Noctilux 0.95 with my black chrome APO.  If the APO transmits light differently at f/2 from the ASPH, it will also do so with the Noctilux.  Each shot at ISO 100, 1/60s and f/2, at infinity, pointed at a white wall (defocused).

First, the Noctilux:

 

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!

[IkarusJohn]

Second, the black chrome APO Summicron-M 50 at f/2:

 

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!

Tiny bit of vignetting and lighting uneven (slightly darker on the right), but otherwise, I don't see anything unexpected.

Edited July 25, 2023 by IkarusJohn

[Jared]

12 hours ago, jaapv said:

Nobody has suggested a mistake in the aperture size   

 

 

Actually, that’s exactly what Chazphoto suggeted in Post 12 when he said, “My guess is that the lens is just shy of f/2 when fully open.”

[jaapv]

Yes I know. I ignored that one as being fantasy. In a way it is understandable as many lenses will open up a bit more beyond the click without any effect, however. 

[Jared]

17 minutes ago, IkarusJohn said:

Second, the black chrome APO Summicron-M 50 at f/2:

 

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!

Tiny bit of vignetting and lighting uneven (slightly darker on the right), but otherwise, I don't see anything unexpected.

Are you sure there is no profile being applied for the lenses? ‘Cause Leica’s own spec sheet shows about a stop and a half of falloff for the APO at f/2, and your image shows almost no vignetting at all.

As to your general premise that hand held light meters wouldn’t have worked in the film days if f/stops weren’t accurate… Well, there is a reason for T-stops to exist. F/stops aren’t and never were all that accurate, and certainly illumination could vary by a half stop or more from one lens design to another despite the same indicated aperture, and vignetting has always darkened corners significantly for the vast majority of lenses when shot wide open, and handheld meters never corrected for that. They can’t. 

[jaapv]

BTW this is the metering pattern of digital M cameras it is limited towards the edges by the FOV of the metering cell. This cannot be influenced by vignetting. 

I would add that Lindolfi is head of a university physics laboratory. 

Edited July 25, 2023 by jaapv

[IkarusJohn]

No profiles that I am aware of; but if there were, the problem is what exactly?  I certainly did nothing in LR to either image.

The discussion of t-stops is all very interesting, but we are not shooting video, and these lenses are not designed for it.  We use f/stops.  The two images, taken on the SL with the same ISO, same f/stop and same shutter speed, taken from a tripod (no lens hoods) using the Fotos App as a remote, then downloaded to LR and exported should show a real life difference.  If there was any.

At least, for me, using these lenses, I see no cause for concern.  When using my digital M cameras and my SL, I do use the in camera meter; with my M-A, I rely on what is on the lens barrel, so such variation is important, particularly when using slide film.

[jaapv]

Yes but my post  above of the metering pattern, the same spot as an M6 makes the whole idea of vignetting influencing exposure more than unlikely. 

[Advertisement]

Hallo,

Look here: Leica M

[IkarusJohn]

38 minutes ago, jaapv said:

Yes but my post  above of the metering pattern, the same spot as an M6 makes the whole idea of vignetting influencing exposure more than unlikely. 

Of course.  I’m not sure about multi-field metering (it would be unusual), but a vignette will only affect metering if you put the spot there.  I had understood that the original post suggested that the APO was 1/2 stop under exposing at f/2 compared to the ASPH.  That is a different issue to vignetting.  Or have I missed something?

All lenses vignette to varying degrees when used wide open - if it’s natural, it’s always something I have liked.

While I concede that my experiment was with a black chrome APO and a Noctilux, the black chrome has the same formula as the standard APO (just the red numbers, scalloped focus ring and vented clip on hood) and the Noctilux is the same vintage as the ASPH, and very similar at f/2, if the APO was under exposing, surely it would be apparent?